Analysing English Sentence Structure

Andrew Radford has acquired an unrivalled reputation over the past forty
years for writing syntax textbooks in which difficult concepts are clearly
explained without excessive use of technical jargon. Analysing English
Sentence Structure continues in this tradition, offering a well-structured
intermediate course in English syntax and contemporary syntactic
theory. Chapters are split into core modules, each focusing on a specific
topic, and the reader is supported throughout with learning aids such as
summaries, lists of key hypotheses and principles, extensive references,
exercises with handy hints, and a glossary of terminology. Both teachers
and instructors will benefit from the book’s free online resources, which
comprise an open-access Students’ Answerbook, and a password-
protected Teachers’ Answerbook, each containing comprehensive answers
to exercises, with detailed tree diagrams. The book and accompanying
resources are designed to serve both as a coursebook for use in class, and
as a self-study resource for use at home.

Andrew Radford is Emeritus Professor at the University of Essex. He

has written a number of popular textbooks, including Transformational
Grammar (1988) and Minimalist Syntax (2004), and research monographs
on Colloquial English (2018) and Relative Clauses (2019).
Analysing English
Sentence Structure
An Intermediate
Course in Syntax

Andrew Radford
University of Essex
Shaftesbury Road, Cambridge CB2 8EA, United Kingdom

One Liberty Plaza, 20th Floor, New York, NY 10006, USA

477 Williamstown Road, Port Melbourne, VIC 3207, Australia

314–321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre,

New Delhi – 110025, India

103 Penang Road, #05–06/07, Visioncrest Commercial, Singapore 238467

Cambridge University Press is part of Cambridge University Press & Assessment,

a department of the University of Cambridge.

We share the University’s mission to contribute to society through the pursuit of

education, learning and research at the highest international levels of excellence.

[Link]
Information on this title: [Link]/highereducation/isbn/9781009322966
DOI: 10.1017/9781009322980

© Andrew Radford 2024

This publication is in copyright. Subject to statutory exception and to the provisions

of relevant collective licensing agreements, no reproduction of any part may take
place without the written permission of Cambridge University Press & Assessment.

First published 2024

A catalogue record for this publication is available from the British Library.

A Cataloging-in-Publication data record for this book is available from the Library of
Congress

ISBN 978-1-009-32296-6 Hardback

ISBN 978-1-009-32293-5 Paperback

Cambridge University Press & Assessment has no responsibility for the persistence
or accuracy of URLs for external or third-party internet websites referred to in this
publication and does not guarantee that any content on such websites is, or
will remain, accurate or appropriate.
 This book is dedicated to a very special person – my dear
wife Khadija. Thank you for the unstinting love, friendship,
happiness and laughter you have brought me over the past four
decades, for sharing your ever-welcoming family with me, and
for broadening my cultural horizons to include couscous and
(through your two amazing sons, Karim and Mourad)
Olympique Marseille: Allez l’OM!
 Contents

Preface xiii
Acknowledgements xvii
1 Background 1
1.0 Overview 1
1.1 Basic concepts 1
1.1.1 Prescriptive and descriptive grammar 1
1.1.2 Syntactic data 3
1.1.3 Categories and features 8
1.2 Generating syntactic structures 13
1.2.1 Merge 13
1.2.2 X-bar syntax 17
1.2.3 Adjunction 18
1.3 Invisible structure and structural relations 23
1.3.1 Null constituents 23
1.3.2 Case-marking and c‑command 28
1.4 Movement operations 32
1.4.1 A‑bar Movement 32
1.4.2 A‑Movement 36
1.4.3 Head Movement 38
1.5 Blocking mechanisms 44
1.5.1 Constraints 44
1.5.2 Filters 49
1.6 Summary 52
1.7 Bibliographical notes 58
1.8 Workbook 62
Exercise 1.1 62
Exercise 1.2 64
Exercise 1.3 65
Exercise 1.4 66
Exercise 1.5 67
viii Contents

2 A‑Movement 69
2.0 Overview 69
2.1 VP-Internal Subject Hypothesis 69
2.1.1 Subjects in Belfast and Standard English 69
2.1.2 Further evidence for VP‑internal subjects 73
2.2 Argument structure and theta roles 77
2.2.1 Predicates and arguments 77
2.2.2 Theta roles 79
2.2.3 Theta-marking 82
2.3 Unaccusative structures 84
2.3.1 Unaccusative predicates 85
2.3.2 Unaccusatives compared with other predicates 87
2.3.3 The syntax of unaccusative subjects 90
2.4 Passive structures 95
2.4.1 Simple passives 95
2.4.2 Cross-clausal passives 99
2.4.3 Constraints on Passivisation 103
2.5 Raising and Control structures 106
2.5.1 Raising structures 107
2.5.2 Control structures 110
2.5.3 Differences between Raising and Control 115
2.6 Mixed structures 119
2.6.1 A mixed Passive, Control and Raising sentence 119
2.6.2 A mixed Unaccusative and Raising sentence 122
2.7 Summary 123
2.8 Bibliographical notes 125
2.9 Workbook 127
Exercise 2.1 128
Exercise 2.2 129
Exercise 2.3 130
Exercise 2.4 132
Exercise 2.5 133
Exercise 2.6 135

3 Agreement 138
3.0 Overview 138
3.1 Agreement, A‑Movement and Case-marking 138
3.1.1 Agreement and A‑Movement 138
3.1.2 Case-marking 144
3.1.3 Valued and unvalued features 149
3.2 Agreement and case in expletive clauses 152
3.2.1 Expletive it clauses 152
 Contents ix

3.2.2 Expletive there clauses 158

3.2.3 Conditions on the use of expletives 165
3.3 Agreement, A‑Movement and case in infinitives 168
3.3.1 Raising infinitives 169
3.3.2 ECM infinitives 174
3.3.3 Passive infinitives 177
3.3.4 Infinitival CPs 179
3.4 Cross-clausal agreement 184
3.4.1 Agreement across a finite TP 184
3.4.2 Agreement across a finite CP 192
3.4.3 Copy Raising with non-expletive subjects 198
3.5 Summary 203
3.6 Bibliographical notes 206
3.7 Workbook 210
Exercise 3.1 210
Exercise 3.2 211
Exercise 3.3 213
Exercise 3.4 214

4 The clause periphery 217

4.0 Overview 217
4.1 The Cartographic approach 217
4.1.1 Force and topic projections 219
4.1.2 In situ and ex situ topics 222
4.1.3 Multiple topic structures 225
4.2 Focus projections 228
4.2.1 Peripheral focused constituents 228
4.2.2 Comparing topic and focus 231
4.3 Modifier and finiteness projections 233
4.3.1 Modifier projections 233
4.3.2 Finiteness projections 235
4.4 Complete and truncated clauses 238
4.4.1 Complete clauses 238
4.4.2 Truncated clauses 241
4.5 Summary 243
4.6 Bibliographical notes 246
4.7 Workbook 249
Exercise 4.1 249
Exercise 4.2 251
Exercise 4.3 253
Exercise 4.4 254
x Contents

5 More peripheral constituents 257

5.0 Overview 257
5.1 Negative and Interrogative Inversion 258
5.1.1 Negative Inversion 258
5.1.2 Interrogative Inversion 262
5.2 Wh‑questions without Interrogative Inversion 267
5.2.1 Embedded wh‑questions 267
5.2.2 How come questions 270
5.3 Yes–no questions 277
5.3.1 Root yes–no questions 277
5.3.2 Embedded yes–no questions 279
5.4 Exclamative and relative clauses 281
5.4.1 Exclamative clauses 281
5.4.2 Relative clauses 283
5.5 Summary 288
5.6 Bibliographical notes 290
5.7 Workbook 292
Exercise 5.1 293
Exercise 5.2 294
Exercise 5.3 296
Exercise 5.4 297

6 The subperiphery 300

6.0 Overview 300
6.1 Auxiliary and subject projections 300
6.1.1 Auxiliary projections 300
6.1.2 Subject projection 305
6.2 Subperipheral adverbs 310
6.2.1 Analysing ADVPs as specifiers 310
6.2.2 Relative ordering of multiple ADVPs 315
6.3 Word order variation 317
6.3.1 ADVPs and auxiliaries 318
6.3.2 ADVPs and subjects 319
6.3.3 Postmodifying ADVPs 321
6.4 Other subperipheral constituents 324
6.4.1 PPs and subordinate clauses 324
6.4.2 Floating QPs: Stranded under movement 327
6.4.3 Floating QPs: In a dedicated projection 332
6.5 Summary 337
6.6 Bibliographical notes 339
 Contents xi

6.7 Workbook 340

Exercise 6.1 341
Exercise 6.2 342
Exercise 6.3 343
Exercise 6.4 344

7 Abbreviated registers 347

7.0 Overview 347
7.1 Subject Drop 347
7.1.1 The nature of Subject Drop 347
7.1.2 A Truncation account of Subject Drop 352
7.1.3 Subject+Auxiliary Drop 354
7.2 Auxiliary Drop 358
7.2.1 A Truncation account of Auxiliary Drop 358
7.2.2 A Left Edge Ellipsis account 362
7.2.3 A prosodic account 364
7.3 Article Drop 367
7.3.1 A Truncation analysis of Article Drop 368
7.3.2 Null article and article ellipsis accounts 371
7.4 Be Drop 375
7.4.1 Characteristics of Be Drop 376
7.4.2 Truncation accounts of Be Drop 378
7.4.3 Be Drop and Article Drop 383
7.5 Object Drop 390
7.5.1 A Topic Drop analysis of null objects 392
7.5.2 An Article Drop analysis of null objects 395
7.5.3 Null objects as instances of a null pronoun/pro 398
7.6 Summary 402
7.7 Bibliographical notes 405
7.8 Workbook 407
Exercise 7.1 407
Exercise 7.2 411
Exercise 7.3 413
Exercise 7.4 415
Exercise 7.5 417

Glossary and Abbreviations 420

References 462
Index 541
 Preface

Aims
This book is a response to the question: ‘What kind of course materials
can we use for an intermediate course intended as a follow-up to an intro-
duction to English syntax?’ Some teachers use handbooks, and others use
sets of original research articles, but both of these can be too demanding
for many students, and do not contain essential teaching materials such as
exercises with helpful hints and model answers, or a glossary of terminol-
ogy. This book aims to fill this gap in the market, and to serve as an inter-
mediate coursebook in syntactic analysis and English syntax, designed as a
follow-up to introductory syntax books. It aims to give readers:

• a thorough grounding in syntactic analysis and argumentation, and in

how the syntactic component of a grammar works
• an appreciation of how data can be used to evaluate the strengths and
weaknesses of competing accounts of syntactic phenomena
• familiarity with a wide range of phenomena in English syntax, and
how they can be analysed using concepts and constructs discussed in
the book
• an understanding of Minimalist and Cartographic approaches to syntax,
and how they handle specific phenomena
• hands-on experience of analysing specific phenomena through the ex-
tensive exercise material in the ‘Workbook’ module of each chapter
• detailed feedback on exercise answers for students and teachers in the
(free) online Answerbooks (one for students, another for teachers)

It is a dual-purpose book, designed to function both:

• as a coursebook, providing extensive materials for lectures, seminars,

workshops, classes, and assignments; and
• as a self-study aid for students working from home
xiv Preface

Level of the book

The book is designed as a follow-up course for students who have already
taken an introductory course in syntax (such as Radford 2016 or Radford
2020) and presupposes familiarity with:

• approaches to grammar; the goals of linguistic theory; the nature of

categories and features; words, phrases, clauses and sentences; univer-
sals; variation
• syntactic structure: X-bar syntax: heads, complements, specifiers and
adjuncts; the classic CP+TP+VP model of clause structure; tests for con-
stituent structure
• null constituents: null subjects; null T and C heads; null D and P heads
• Head Movement operations; Auxiliary Inversion; Auxiliary Raising;
Verb Raising
• Movement to spec‑CP; Wh‑Movement in interrogatives, exclamatives
and relatives; Topicalisation; Focusing; Constraints on movement

For readers who may need reminding about some of these topics, Chapter 1
provides a Background summarising the key ideas that are taken as as-
sumed knowledge in the rest of the book.

Key features

Key features of the book include the following:

• It is written in an accessible style and adopts a hands-on approach in

which students get to do syntax for themselves, with extensive practice
exercise material in the Workbook modules of the book, and model an-
swers and detailed solutions provided in the accompanying free online
Answerbooks for students and teachers
• It makes extensive use of authentic materials from a wide variety of
written and spoken sources, and covers a range of unusual structures
which tend to be neglected in most textbooks, thereby offering an inter-
esting perspective on microvariation
• It takes an in-depth look at a range of topics not touched on in any de-
tail in earlier books of mine, with two chapters devoted to the syntax
of the clause periphery, another to the syntax of the subperiphery, and
another to the syntax of abbreviated registers of English (as found e.g.
in recipe books, social media, newspaper headlines, etc.)
• It has a modular structure, with each chapter including a number of
core modules, each tackling a distinct topic and having its own set of
 Preface xv

dedicated exercise materials (with answers in the Answerbooks): for ex-

ample, Exercise 2.1 is linked to Module 2.1, Exercise 2.2 to Module 2.2,
Exercise 2.3 to Module 2.3, and so on. This modular structure makes it
easier to regulate the pace of learning (e.g. for teachers/learners to de-
cide whether to cover one, two, or more modules per week), and to skip
specific modules or chapters if so desired (e.g. modules on non-standard
structures, or the chapter on Agreement, etc.)
• All modules within chapters are subdivided into sections, so breaking
up the material into shorter (and more readable) units
• The book begins with a Background chapter reminding students of key
ideas that they will have covered in an earlier introductory course, some
of which they may – Heaven forbid! – have forgotten
• The core chapters in the book contain extensive materials designed to
help students, including an overview and summary of the main points
and key constructs in each chapter, and a set of bibliographical notes
that form a useful starting point for a literature review for research pro-
jects on a wide range of topics
• Each chapter also includes a ‘Workbook’ component containing exten-
sive exercise material (one exercise for each core module in the chapter),
and helpful hints for students on how to tackle exercises, plus a list of
key constructs for students to utilise in tackling exercise examples
• Some Workbook exercise examples have an [M] after them, and this means
that a model answer is provided in the Students’ Answerbook (see below).
Other exercise examples have an [S] after them, indicating that they are
for self-study (so that after reading the relevant module, helpful hints and
model answer associated with the exercise, students can analyse the self-
study examples on their own, and then compare their answers with those
provided in the Students’ Answerbook). The remaining exercise examples
are intended for teachers to use as the basis for hands-on problem-solving
work in workshops, seminars, classes, or assignments: answers to all of
these non-self-study examples are given in the Teachers’ Answerbook
(along with model answers, and answers to self-study exercises).
• There is a detailed Glossary of technical terms and abbreviations at the
end of this book

Free online resources

This book is published in conjunction with the following free online

Answerbooks which provide comprehensive written answers to exercise
examples, with detailed argumentation and evaluation of alternative
xvi Preface

analyses along with discussion of the theoretical implications of par-

ticular analyses:

• A Students’ Answerbook providing a model answer for every exercise,

together with a set of comprehensive written answers to all the self-
study material in the Workbook modules of each chapter
• A Teachers’ Answerbook providing comprehensive written answers to
every single exercise example in the Workbook modules

The Students’ Answerbook is available for readers to download for free

at [Link]/analysingenglish. The Teachers’ Answerbook can be
downloaded for free by teachers, and can be accessed from a password-pro-
tected website at [Link]/analysingenglish.
 Acknowledgements

I am grateful to the following for providing the help set out below:
• Five anonymous reviewers at Cambridge University Press for extensive
comments, suggestions, and criticisms of an earlier draft of the book
• Helen Barton (Commissioning Editor) and her team at Cambridge Uni-
versity Press for helping steer me through the protracted publication
process ever since I submitted a first draft of the manuscript in October
2021
• Gordon Lee (copy-editor) and Preethika Ramalingam (typesetter) for in-
valuable help in preparing the manuscripts and proofs for publication
• Mike Jones and Andrew Weir for comments on a draft of Chapter 7
• Giuliano Armenante and Ingo Reich, for sharing their ideas with me on
tense drop in Module 7.4
• The following for kindly providing me with example sentences: Alison
Henry for the Belfast English examples (2, 3) in Chapter 2; Johannes
Gisli Jónsson for the Icelandic example (40) in Chapter 2; Andrew Spen-
cer for the Russian example (16) in Chapter 3.

And a footnote from the Rolling Stones

In 1965, when I was an undergraduate at Trinity College Cambridge, the

Rolling Stones released their hit single The Last Time. The chorus line goes
like this:

Well this could be the last time

This could be the last time
Maybe the last time
I don’t know, oh no, oh no
I’ve been publishing syntax books with Cambridge University Press since
my Italian Syntax book way back in 1977. I’ve just celebrated my 77th
birthday (in Morocco, with my wife’s family) and there may simply not
be enough time left for me to publish another (since writing another book
xviii Acknowledgements

and the accompanying answerbooks would take three years, and the pub-
lication process would take at least another year). So, this could indeed
be my last ever book. If so, let me just say: it’s been challenging for me
writing them, and I hope it has been equally challenging for you reading
them!
1 Background

1.0 Overview

Since this is an intermediate book intended as a follow-up to an introduc-

tion to syntax, it presupposes that (before tackling it) readers will already
have taken an introductory course in syntax (let’s call it Syntax 101), and
acquired a grounding in the basics of syntax. But just in case Syntax 101
wasn’t quite as memorable as you had hoped, I thought it might be useful
to start with a chapter which serves as a reminder of key ideas that will no
doubt have been covered in Syntax 101 (and which are taken as assumed
background knowledge in Chapters 2–7). To signpost key terms in this
Background chapter, I print them in bold.

1.1 Basic concepts

This module provides a brief introduction to approaches to syntax, to the

nature of data in syntax, and to the role played by categories and features
in syntactic description.

1.1.1 Prescriptive and descriptive grammar

Linguists analysing the syntax of a particular language attempt to devise
a grammar of the language that describes the range of structures found in
the language. However, in order to attain a deeper understanding of the
nature of natural (i.e. human) language, they seek to investigate the extent
to which the grammar of any given language reflects universal properties
(i.e. properties shared by the grammars of all languages), and the extent
to which there is variation between different languages. Consequently, the
ultimate goal of the study of grammar is to develop a theory of Universal
Grammar/UG which identifies universal properties shared by all languages
on the one hand, and parameters of variation between different languages
(or language varieties) on the other.
The goal of a linguist describing the syntax of a specific language (e.g.
English, the focus of this book) is to seek to understand the ways in which
words are combined together to form phrases, clauses and sentences.
2 Background

However, a fundamental methodological question that needs to be resolved

at the outset concerns what kind of approach to adopt in studying gram-
mar. One traditional view sees the role of grammar as being essentially
prescriptive (i.e. prescribing norms for grammatical correctness, linguistic
purity and literary excellence). However, a more modern view sees the role
of grammar as being inherently descriptive (i.e. describing the way peo-
ple speak or write their native language). We can illustrate the differences
between these two approaches in relation to the following TV dialogue
between the fictional Oxford detective Morse and his assistant Lewis, as
they are examining a dead body (where italics mark items of grammatical
interest):

(1) M orse : I think he was murdered, Lewis

L ewis : Who by, sir?
M orse : By whom, Lewis, by whom. Didn’t they teach grammar at that
comprehensive school of yours?

Morse was educated at a grammar school – that is, an elitist school which
sought to give pupils a ‘proper education’ and taught them grammar,
so that they could learn to speak and write ‘properly’ (i.e. in a prestig-
ious form of standard English). Lewis, by contrast, was educated at a
comprehensive school – that is, a more socially inclusive type of school
which admitted pupils from a much broader social spectrum and didn’t
force-feed them with grammar. The linguistic skirmish between Lewis and
Morse in (1) revolves round the grammar of an italicised phrase which
comprises the preposition by and the pronoun who(m). The differences
between what the two men say relates to (i) the form of the pronoun (who
or whom?), and (ii) the position of the pronoun (before or after the prepo-
sition by?). Lewis uses the pronoun form who, and positions the pronoun
before the preposition when he asks Who by? Morse corrects Lewis and
instead uses the pronoun form whom and positions the pronoun after the
preposition when he says By whom? But why does Morse correct Lewis?
The answer is that Morse was taught traditional prescriptive grammatical
rules at his grammar school, including two which can be outlined infor-
mally as follows:

(2) (i) The form who is used as the subject of a finite verb, and whom as the
object of a verb or preposition
(ii) Never end a phrase, clause or sentence with a preposition

When Lewis asks Who by? he violates both prescriptive rules. This is
because the pronoun who(m) is the object of the preposition by and rule (2i)
stipulates that whom must therefore be used, and rule (2ii) specifies that the
preposition should not be positioned at the end of a phrase. The corrected
 1.1 Basic concepts 3

form By whom? produced by Morse obeys both rules, in that whom is

used in conformity with rule (2i), and by is positioned in front of its object
whom, thereby avoiding violation of (2ii).
The more general question raised by our discussion here is the fol-
lowing. When studying syntax, should we adopt a descriptive approach
and describe what ordinary people like Lewis actually say, or should we
adopt a prescriptive approach and prescribe what people like Morse think
they ought to say? There are several reasons for rejecting the prescriptive
approach. For one thing, it is elitist and socially divisive in that a privi-
leged elite attempts to lay down grammatical norms and impose them on
everyone else in society. Secondly, the grammatical norms which prescrip-
tivists seek to impose are often derived from structures found in ‘dead’
languages like Latin, which is somehow regarded as a model of grammat-
ical precision and linguistic purity: and so, because Latin made a distinc-
tion between subject and object forms of pronouns, English must do so
as well; and because Latin (generally) positioned prepositions before their
objects, English must do so as well. Such an approach fails to recognise
typological diversity in languages – that is, that there are many different
types of structure found in the world’s 8,000 or so known languages.
Thirdly, the prescriptive approach fails to recognise sociolinguistic varia-
tion – that is, that different types of structure are found in different styles
and varieties of English (e.g. By whom? is used in formal styles of English,
and Who by? or By who in non-formal styles). Fourthly, the prescriptive
approach also fails to recognise that languages are constantly evolving,
and that structures used centuries ago may no longer be in use today (e.g.
whom is an archaic form which has largely dropped out of use and is no
longer part of the grammar of teenagers today). And fifthly, prescriptive
rules are very often oversimplistic, in the sense that they paint an over-
simplified picture of what is in fact a more complex linguistic reality (as
our discussion of by who/m and who by illustrates). For reasons such as
these, the approach taken to grammar in work over the past sixty years or
so has been descriptive.
What this means is that in attempting to devise a grammar of (for exam-
ple) English, contemporary linguists aim to describe the range of grammati-
cal structures found in present-day English. But how do we determine what
is or isn’t grammatical in present-day English? What kinds of data can we
use? This is the question addressed in the next section.

1.1.2 Syntactic data

One way of collecting syntactic data is to study usage (i.e. the range of struc-
tures used by people when they speak or write). Contemporary linguists who
adopt this kind of approach rely on data from a corpus (e.g. a computerised
4 Background

database such as the British National Corpus) containing authentic examples

of spoken or written English. Such corpora offer the advantage that they
contain millions of sentences, and the sentences have usually been codified/
tagged by a team of researchers, so simplifying the task of searching for
examples of a particular construction. Some linguists treat the web as a form
of corpus, and use a search engine to find examples from the internet of the
kind of structures they are interested in.
However, although usage data (from corpora or the web) provide a useful
source of information about what people say or write, there are some down-
sides associated with a usage-based approach. For one thing, a corpus may
contain relatively few examples of low-frequency structures. Secondly, it
is generally not possible to ask the speakers who produced the sentences in
the corpus questions (e.g. ‘How would you negate this sentence?’). Thirdly,
a corpus may contain examples of production errors (slips of the tongue, or
pen, or keyboard) which would probably be judged as unacceptable even
by the people who produced them. And (in the case of internet examples),
it is sometimes unclear whether someone producing a given sentence (who
may use an identity-concealing pseudonym like CutiePie or MasterBlast-
er as their name) is a native speaker of English or not (i.e. someone who
has acquired and used English as a first language in an English-speaking
environment from birth or early childhood, and who speaks the language
fluently), and if so what variety/dialect of English they speak.
A very different approach to grammaticality is to rely on introspective
judgements by native speakers (i.e. their ‘gut feelings’ or ‘intuitions’ about
whether a particular sentence is or isn’t grammatical in their native lan-
guage). For example, any native speaker of English would readily accept I
don’t like syntax as a grammatical sentence of English, but not *I no like
syntax (where the asterisk marks ungrammaticality). Consequently, an ap-
proach widely used by linguists over the past seven decades (particularly by
Noam Chomsky and his followers) has been to devise grammars on the ba-
sis of native-speaker intuitions about grammaticality. Where linguists are
describing aspects of their own native language, they often rely primarily
on their own introspective grammaticality judgements.
However, although extensively used, this approach of relying on intro-
spective judgements about the grammaticality of sentences has been crit-
icised by some as being unscientific (hence yielding potentially unreliable
results). One problem is that people are sometimes (whether consciously or
subconsciously) influenced by prescriptive rules inculcated at school, and
hence may give a prescriptive judgement about the grammaticality of a
particular sentence which reflects the ‘proper English’ they were taught to
use at school, rather than a descriptive judgement about the ‘real English’
which they actually use when talking to their friends.
 1.1 Basic concepts 5

A second problem which arises from asking native speakers whether

such-and-such a sentence is grammatical or ungrammatical in their variety
of English arises in relation to so-called marginal sentences – that is, sen-
tences of doubtful grammaticality, such as the following:

(3) a. He ought to apologise, oughtn’t he?

b. He ought to apologise, shouldn’t he?

Such sentences are referred to as tag questions (with the italicised part of
the sentence following the comma being the tag). Normally in tag ques-
tions, the auxiliary in the tag is a (contracted negative) copy of that used in
main clause (i.e. the part of the sentence preceding the comma). However,
use of the oughtn’t tag in (3a) results in a relatively degraded sentence for
speakers like me, simply because the contracted negative form oughn’t is
obsolete in my variety of English. Instead, I’d prefer to use should in the tag
(which can freely have the contracted negative form shouldn’t), as in (3b):
but since this results in a structure with mismatching auxiliaries (ought
in the main clause, shouldn’t in the tag), it still feels ungainly. Thus the
problem posed by asking people to make a judgement on whether a given
sentence is grammatical or ungrammatical is that such a binary judge-
ment is problematic for marginal sentences like (3a, 3b) which are neither
clearly grammatical nor clearly ungrammatical. More generally, marginal
sentences pose a problem for the use of introspective judgements about
sentence acceptability for a number of reasons. Firstly, different individuals
may disagree in their judgements of particular sentences (and may have
different tolerance thresholds): this means that relying on the intuitions of
one person alone could give misleading results. Secondly, the same individ-
ual may sometimes give conflicting judgements about the same sentence on
different occasions. Thirdly, it can sometimes be very difficult to judge the
grammaticality of a sentence in isolation (without an appropriate context).
Fourthly, grammaticality is sometimes a matter of degree rather than an
absolute property (e.g. a given sentence may be more acceptable than some
sentences but less acceptable than others). Fifthly, native speakers who are
not linguists very often have no clear idea what it means for a sentence
to be ‘grammatical’ or not (since grammaticality is a technical term which
non-linguists may have little conception of): rather, all that non-experts
can do is say how acceptable they find a sentence, and this may depend on
a range of factors which have little to do with grammaticality, including
how frequent a given structure is, whether it contains taboo language or
concepts, and so on. And sixthly, linguists who rely on their own grammat-
icality judgements tend to give different judgements from non-linguists,
and are vulnerable to the accusation that (however unwittingly) they may
tailor their grammaticality judgements to fit their analysis (e.g. they may
6 Background

judge a given sentence to be grammatical because their analysis predicts

that it should be).
Because of the potential unreliability of informal intuitions, some lin-
guists prefer to adopt an experimental approach to eliciting native-speak-
er judgements, particularly when dealing with marginal structures whose
grammaticality status is not clearcut. One type of experiment involves ask-
ing a group of native speakers to judge the grammaticality of a large set
of test sentences which are flashed up on a computer screen one at a time,
with subjects being asked to rate the acceptability of each sentence on a
seven-point scale on which 7 means ‘completely acceptable’ and 1 means
‘completely unacceptable’. Using this scale, a marginal sentence might be
rated as 4, for example.
However, it should be acknowledged that there are a number of draw-
backs to experimental studies. For one thing, they require considerable time
and money to set up: it can take months to design an experiment, collect
the data, and process the results; and a design flaw (or problematic results)
may require the whole experiment to be re-designed and subsequently re-
run. Moreover, it is in the nature of experiments that (in order to meet
stringent methodological requirements on experimental design) they can
only be used to collect data relating to a specific (and narrow) set of phe-
nomena. Furthermore, experiments can sometimes produce results which
are skewed by the design of the experiment. In addition, how acceptable
(or otherwise) people perceive a sentence to be may depend on a whole
range of extraneous factors other than its grammaticality: these extraneous
factors include, for example, how interesting it is, how long it is, how plau-
sible it is, how frequent the relevant type of structure is, how easy it is to
imagine a context where it could be used, and whether or not the sentence
expresses ideas which offend cultural or religious sensibilities or contains
taboo words. Furthermore, the results which experiments yield can be far
from straightforward to interpret: for example, they sometimes produce
results which represent acceptability in terms of many different shades of
grey, rather than as a black-and-white issue. Moreover, in order to achieve
statistical significance in results, it may be necessary to discard outliers (i.e.
atypical results).
The approach that I will adopt in this book is a hybrid one, combining my
own introspective judgements with usage data I have collected from live
unscripted radio and TV broadcasts, supplemented with internet-sourced
examples. I should add two caveats to this. The first is that I tend to accept
a wider range of structures than many other people do: for example, one
reviewer noted that he found some of the examples of complex structures
in Chapter 4 unacceptable in his English. And the second is that my cor-
pus data reveal many types of structure which are widely considered to be
 1.1 Basic concepts 7

ungrammatical in standard English. For example, below are a few types of

non-standard structures reported in earlier books of mine:

(4) a. What a mine of useless information that I am (Sir Terry Wogan, BBC
Radio 2; Radford 1988: 501)
b. That’s the guy who I think’s sister is the lead singer in a new band
(Radio presenter, Top Shop, Oxford Street, London; Radford 1988:
526)
c. What is thought has happened to him? (Reporter, BBC Radio 5;
Radford 2004a: 429)
d. To which of these groups do you consider that you belong to? (Form
issued by the Council in the town where I live; Radford 2009a: 233)
e. This is a team [which Fabio Capello has seen them play] (Gabriel
Marcotti, BBC Radio 5; Radford 2019: 90)

These sentences are unusual from the perspective of standard English in

several respects. For instance, the use of that in (4a) is odd because that
is neither used in main clauses nor in wh‑clauses in standard varieties.
In (4b), the affix’s is separated from its potential host who, resulting in
(non-standard) affix stranding (whereas standard English requires ‘That’s
the guy whose sister I think is the lead singer in a new band’). In (4c),
the pronoun what has undergone a Passivisation operation which moves
it from being the subject of has to a position at the front of the main
clause, in spite of the widespread claim that subjects can’t be extracted
out of finite clauses in standard English. In (4d), we find (non-stand-
ard) preposition doubling, resulting in one copy of the preposition to
being placed at the beginning of the sentence, and another at the end. In
(4e), we find use of the resumptive pronoun them to reprise the relative
pronoun which, so giving rise to a (nonstandard) resumptive relative
structure.
Sentences like those in (4) raise important questions about whether such
sentences are grammatical in English. One perspective is that they are un-
grammatical, and are perhaps the result of accidental processing errors.
On this view, we might conjecture that people who produce preposition
doubling structures like (4d) forget that they already fronted the preposition
to along with which, and so spell out the preposition again at the end of
the sentence – and indeed, a processing account of this kind is outlined in
Radford et al. (2012). Still, it is much less obvious what kind of process-
ing errors could give rise to the other types of sentences in (4). Moreover,
some of the structures in (4) are relatively frequent (in that I have collected
hundreds of examples of them): for example, one linguist told me (after
finishing a PhD on relative clauses!) that his initial reaction to resumptive
relatives was to treat them as ungrammatical, but that after reading my
8 Background

(2019) book on relative clauses in everyday English, he began to notice

just how frequent they are in everyday conversation. Of course, it might
be that some of the structures in (4) originated as processing errors, and
then became grammaticalised (i.e. treated as part of the grammar) in some
varieties.
An alternative perspective (which I defend in a book-length study in
Radford 2018) is to suppose that what we are dealing with in sentences
like (4) is microvariation between different varieties of English: for ex-
ample, there are varieties of English (including mine) which allow use of
that in exclamatives in main and subordinate clauses alike; there are va-
rieties (including mine) which allow resumptive relatives; there are vari-
eties (including mine) which allow subjects to be extracted out of certain
finite clauses; there are varieties (not including mine) which allow affix-
stranding in possessive structures, and so on. In this book, I will adopt the
microvariation approach, and accordingly some of the sentences discussed
in the core of the book (or set as exercises in the Workbook modules) in-
volve analysing interesting non-standard structures in English.

1.1.3 Categories and features

As noted at the beginning of §1.1.1, the goal of a linguist describing the
syntax of a given language is to seek to understand the ways in which
words are combined together to form larger structures (phrases, clauses and
sentences). Properties of individual words determine the range of structures
they can appear in, and these properties for centuries have been described
by grouping words into categories on the basis of grammatical proper-
ties they share in common. Traditional categories include nouns like dog
(denoting an object), verbs like sing (denoting an action), adjectives like
happy (denoting a state), adverbs like cleverly (denoting the manner in
which something is done), and prepositions like under (denoting a loca-
tion): these are generally termed lexical categories because most lexical
items (= dictionary items = words) belong to categories like these. Lexical
categories typically have a very large membership: for example, there are
dozens of prepositions in English, and thousands of nouns. Words belong-
ing to lexical categories are traditionally called content words because
they have descriptive content (e.g. dog is a noun, and it’s easy enough to
draw a picture of a dog).
Work since the 1950s, however, has placed increasing emphasis on
so-called functional categories – that is, categories whose members are
words with a grammatical function, serving to mark properties such as
definiteness, tense, mood, aspect, clause type etc. These include determiners
like the/this/that/these/those, quantifiers like all/both/each/several/many,
auxiliaries/auxiliary verbs such as will/would/can/could/may/might, and
 1.1 Basic concepts 9

complementisers (i.e. clause-introducing particles) like the italicised words

at the beginning of the bracketed clause below:

(5) I didn’t know [that/if/whether she would be there]

Functional categories tend to have a very small membership: for example,

there are only around a dozen auxiliaries in English. It has become stand-
ard practice to abbreviate the names of categories using capital letters, and
employ (for example) N for noun, V for verb, P for preposition, A for adjec-
tive, ADVP for adverb, D for determiner, Q for quantifier, AUX for auxiliary,
and C for complementiser.
The categorial status of a word determines its distribution – that is, what
range of positions it can occupy in sentences. For example, consider what
kind of single word can occur in the gap ( — ) position in a sentence like
the following:

(6) He — go home early

The answer is: an auxiliary like will/would/can/could/may/might/must/did,

but not, for example, a determiner like the, or a complementiser like if, or
an adjective like happy (and so on).
However, categories alone aren’t sufficient to describe the grammatical
properties of words. This is because many words have different forms, depend-
ing on the types of structure in which they occur. By way of illustration, con-
sider the words this and that. These both belong to the category D/determiner,
and yet they have different forms in different uses – as can be illustrated by
the examples below (where a prefixed asterisk indicates ungrammaticality):

(7) a. Do you like this/that/*these/*those dress?

b. Do you like these/those/*this/*that dresses?

In (7a), we can fill the italicised position with this/that but not these/those,
whereas conversely in (7b) we can fill the italicised position with these/those
but not this/that. Why should this be? The answer does not lie in the catego-
rial status of the words (both are determiners), but rather in a finer-grained
property which is generally described in terms of the grammatical features
carried by words. More specifically, the forms this/that are singular forms
which are used to modify a singular noun like dress, whereas these/those are
plural forms used to modify a plural noun like dresses. One way of handling
this is to suppose that nouns and determiners inflect for number, and carry
the feature [singular-number] in sentences like (7a), and the feature [plu-
ral-number] in sentences like (7b). Determiners and the nouns they modify
agree in number (i.e. they have the same number value – for example, a plu-
ral determiner modifies a plural noun), and this type of agreement between a
noun and a determiner modifying it is traditionally termed concord. Features
10 Background

are enclosed in square brackets, and are often abbreviated to save space: they
typically comprise an attribute/property and a value – as in [Sg-Num] ‘singu-
lar number’, where [Num] is the attribute and [Sg] the value.
Another class of words whose subcategorial properties can be described
in terms of features are pronouns like I/we/you/he/she/it/they. Although
these are traditionally categorised as personal pronouns, they differ from
each other in a number of respects. For example, I is a singular pronoun
and we a plural pronoun, and this difference can be captured by positing
that I carries the number feature [Sg-Num], and we [Pl-Num]. Likewise, he
is a masculine pronoun, she a feminine one, and it a neuter/inanimate pro-
noun; these differences can be captured by treating them as carrying gen-
der features, with he being [Masc-Gen], she [Fem-Gen] and it [Inan-Gen].
A further difference between the various types of pronoun relates to their
person properties: the pronouns I/we are first person pronouns denoting
the speaker/s, you is a second person pronoun denoting the addressee/s (i.e.
the person or persons being spoken to), and he/she/it/they are third person
pronouns denoting one or more entities that are neither speaking nor being
addressed. These differences can be captured by supposing that I/we carry
the person feature [1-Pers], you [2-Pers], and he/she/it/they [3‑Pers].
An additional property of pronouns which can be captured in terms of
features is that they inflect for case, as can be illustrated in terms of the
italicised items below:

(8) He says that his house has bankrupted him

The pronoun h e is traditionally said to have three distinct case forms: the
nominative form he, the accusative form him and the genitive form his.
These differences can be captured by supposing that he carries the feature
[Nom-Case], him [Acc-Case], and his [Gen-Case]. It can also be argued that
nouns too carry case – for example, a noun like J o h n has the common
nominative/accusative form John, and the genitive form John’s: see case in
the Glossary at the end of the book for a table of the different case forms
of nouns and pronouns in English.
Another class of words which have a complex range of different forms
that can be captured in terms of features are verbs. These are traditionally
divided into two distinct types – namely lexical verbs and auxiliary verbs.
Lexical verbs are verbs which have inherent descriptive content (e.g. pour
describes an action, die an event, and think a cognitive state), whereas
auxiliary verbs have no descriptive content but rather mark grammatical
properties such as tense, mood and aspect. Auxiliaries also differ from lex-
ical verbs in their syntactic properties. For example, an auxiliary like will
can undergo Inversion (and move to a position in front of its bold-printed
subject) in a question like (9a) below, but a lexical verb like want cannot (as
 1.1 Basic concepts 11

we see from the ungrammaticality of 9b) and instead requires the support
of the auxiliary do (as in 9c):

(9) a. Will he visit Paris?

b. *Wants he to visit Paris?
c. Does he want to visit Paris?

For obvious reasons, this phenomenon is referred to as Do-support.

Some auxiliaries are used to mark aspect (i.e. whether an action is in
progress or completed) – as illustrated by the contrast between the auxilia-
ries italicised below:

(10) a. He is writing a novel

b. He has written a novel

In (10a), the novel-writing activity is still in progress, whereas in (10b) it is

completed. A traditional way of describing this difference is to suppose that
the auxiliary be in (10a) marks progressive aspect, whereas the auxiliary
have in (10b) marks perfect aspect (i.e. marks the perfection/completion of
the relevant act). This difference can be captured in terms of an aspectual
feature if we posit that (in the relevant uses) b e carries the feature [Prog-
Asp] and have [Perf-Asp].
Features can also be used as a way of characterising the properties of
verb forms which are marked for mood. In this respect, consider the itali-
cised verb forms below:

(11) a. I am surprised that he was arrested

b. The judge ordered that he be released
c. ‘Be quiet!’, he said

The italicised verb b e in these examples is traditionally said to be indicative

in mood in declarative (i.e. statement-making) clauses like those in (11a),
subjunctive in mood in (11b) where it denotes a hypothetical event, and
imperative in mood in a sentence like (11c) where it conveys an order. In
much the same way, the complementiser that can be treated as signalling
that the clause it introduces is indicative in mood in (11a), but subjunc-
tive in mood in (11b): the two complementisers differ in that (for many
speakers) indicative that can be omitted, but subjunctive that cannot. These
mood differences can be characterised in terms of features such as [Ind-
Mood], [Subj-Mood] and [Imp-Mood].
Indicative verb forms have the property that they inflect for tense and
agreement – as can be illustrated by the italicised forms below:

(12) a. He is, and always was, passionate about politics

b. I am, you are, and he is passionate about politics
12 Background

In (12a), is marks present tense and was past tense, and this difference can
be characterised in terms of features such as [Pres-Tns] and [Past-Tns].
In (12b), am is used with a first person singular subject like I, are with a
second person subject like you (or with a plural subject like we/they), and
is with a third person singular subject like he. In traditional terms, an
indicative verb is said to agree in person and number with its subject: on
one implementation of this view, the features [1-Pers, Sg-Num] of the first
person singular pronoun I are copied onto the verb b e , so that b e carries the
same features in this use and hence is spelled out in its first person singular
present tense form am in a sentence like I am hungry.
Verb forms that are marked for tense/mood are traditionally termed
finite, whereas those not so marked are termed nonfinite. In this connec-
tion, consider the contrast between the italicised clauses below:

(13) a. Mary was glad that he apologised

b. Mary demanded that he apologise
c. I can’t imagine him apologising
d. It would be sensible for him to apologise
e. It’s important to know when to apologise

The italicised clauses in (13a, 13b) are finite, since they contain a bold-
printed verb marked for tense/mood: in (13a) the verb apologised is finite
by virtue of marking past tense and indicative mood; and in (13b), the
verb apologise is finite by virtue of marking subjunctive mood. A clause
containing a verb in the indicative mood (like that italicised in (13a) can
be used to denote a real (or realis, to use the relevant grammatical term)
event or state occurring at a specific point in time; by contrast, a subjunc-
tive clause (like that italicised in 13b) denotes a hypothetical or unreal (=
irrealis) event or state which has not yet occurred and which may never
occur. By contrast, the bold-printed verbs italicised in (13c–13e) are non-
finite, in that they are not marked for tense or mood. For example, the
verb apologising in (13c) is nonfinite because it is a tenseless and mood-
less gerund form. Likewise, the verb apologise in (13d, 13e) is a tenseless
and moodless infinitive form (as we see from the fact that it follows the
infinitive particle to).
A second property which differentiates finite from nonfinite verbs is that
finite verbs permit a nominative subject (like I, we, he, she or they), where-
as nonfinite verbs permit either an accusative subject (like me, us, him, or
them) or a silent (i.e. unpronounced) subject. On the basis of the form of
their subjects, we can therefore categorise the verbs apologised/apologise in
(13a, 13b) as finite, since they have a nominative subject (he). By contrast,
the italicised clauses in (13c, 13d) are nonfinite, since they have an accusa-
tive subject (him). Similarly, the italicised clause in (13e) is also nonfinite,
 1.2 Generating syntactic structures 13

because it has a silent subject – that is, an ‘understood’ but ‘unpronounced’

subject which is a silent counterpart of the overt subject pronoun you in
‘It’s important to know when you have to apologise’.
To summarise: our discussion in Module 1.1 has served to illustrate the
different types of data that linguists use in syntactic analysis, and has
shown how categories and features play an important role in syntactic
description.
You should now be able to tackle Exercise 1.1.

1.2 Generating syntactic structures

Syntax explores the ways in which words can be combined together to

generate (= form) larger structures (phrases, clauses and sentences). In the
conception of the organisation/architecture of a grammar assumed in much
work inspired by Noam Chomsky, grammars contain three components.
One of these is a syntactic component that generates/forms syntactic struc-
tures. These syntactic structures are in turn inputted into two other com-
ponents. One is the PF/phonological component, and this serves to map/
convert syntactic structures into PF representations (i.e. representations
of their phonetic form, telling us how they are pronounced). The second
component which syntactic structures are inputted into is the LF/semantic
component and this maps/converts syntactic structures into LF representa-
tions (representations of their logical form – i.e. of linguistic aspects of
their meaning). Thus, the three components of a grammar are intercon-
nected in the manner shown in schematic form below:

(14) LF/semantic LF/semantic

component representation
syntactic syntactic
component structure
PF/phonological PF/phonetic
component representation

On this view, for each sentence, the syntactic component generates a syn-
tactic structure which feeds into the LF component where it is assigned
an LF representation, and also feeds into the PF component where it is
assigned a PF representation. In this module, we’ll start by looking at key
syntactic operations used to build syntactic structures.

1.2.1 Merge
A core operation which the syntactic component uses to generate syntactic
structures is merge, which combines pairs of categories together to form a
14 Background

larger unit. To illustrate this, consider the italicised sentence produced by

speaker b in the dialogue below:

(15) speaker a : What did the manager say about the customer?
speaker b : That he had complained about something

Let’s first reflect on the category that each of the italicised words in (15b )
belongs to. That is a C/complementiser, in that it is a clause-introducing
word that tells us that the clause it introduces is indicative in mood and
declarative in force/type, so that the clause expresses a statement. Had is an
auxiliary marking past tense, and so let’s take it to belong to the category
of T (= auxiliary inflected for tense). Complained is a V/verb; and about is
a P/preposition.
More problematic is the status of he and something. These are tradition-
ally termed pronouns, and since pro- is a prefix meaning ‘in place of’, this
might suggest that pronouns are used in place of nouns. However, senten
ces such as the following suggest that this is not the case:

(16) a. He had complained about something

b. He had complained about some trivial matter
c. He had complained about some mistake with his bill
d. He had complained about some meal he had eaten

Thus, in place of the pronoun something in (16a) we can have a quantifier

phrase/QP like some trivial matter in (16b), some mistake with his bill in
(16c), or some meal he had eaten in (16d). Sentences like (16) suggest that
something has the same distribution (i.e. can occupy the same range of
positions) as a QP/quantifier phrase, and thus something functions as a
pronominal QP (i.e. a pronoun which occupies a position that could alter-
natively be occupied by a QP like some trivial matter).
Similar considerations hold in the case of he (traditionally called a per-
sonal pronoun, as noted in §1.1.3). Sentences like (17) below illustrate the
types of expression which he can stand in place of:

(17) a. He had complained about something

b. The guy had complained about something
c. The owner of the car had complained about something
d. The man who was in room 13 had complained about something

In place of the pronoun he in (17a) we can have a determiner phrase/DP

like the guy in (17b), the owner of the car in (17c), or the man who was in
room 13 in (17d). This suggests that he has the same distribution as a DP/
determiner phrase (like the guy etc.), and thus functions as a pronominal
DP (i.e. a pronoun which occupies a position that could alternatively be
occupied by a DP like the guy).
 1.2 Generating syntactic structures 15

Given these background assumptions, let’s return to the structure of the

italicised sentence produced by speaker b in (15) above. On the assumption
that merge is an operation which combines one constituent (i.e. syntactic
unit) with another to form a larger constituent, we can suppose that speaker
b ’s sentence is formed as follows.
The P/preposition about merges with the pronominal QP something to
form the PP/prepositional phrase about something, with the structure below:

(18) PP

P QP
about something

The V/verb complained then merges with the PP in (18) to form the VP/verb
phrase complained about something, with the structure below:

(19) VP

V PP
complained
P QP
about something

Next, the T auxiliary (i.e. past tense auxiliary) had merges with the VP in (19)
to form the larger constituent had complained about something. But what is
the categorial status of this larger constituent? A plausible assumption would
be that it is a TP/tense phrase. But full phrases like VP, PP etc. can serve as
sentence fragments (i.e. self-contained utterances used e.g. as short replies to
questions) – as illustrated by the replies produced by speaker b below:

(20) s p ea k e r a : What did the customer do?

s p ea k e r b : Complain about the bill (VP)

(21) s p ea k e r a : What did the customer complain about?

s p ea k e r b : About the bill (PP)

By contrast, the phrase had complained about something can’t be used as

a sentence fragment – as we see from the ungrammaticality of speaker b ’s
reply in (22) below:

(22) s p ea k e r a : What had the customer done to upset the manager?

s p ea k e r b : *Had complained about something

On the other hand, the fuller reply given by speaker b in (23) below is fully
grammatical:

(23) s p ea k e r a : What had the customer done to upset the manager?

s p ea k e r b : He had complained about something
16 Background

Why should it be that the shorter reply *Had complained about something
can’t be used as a sentence fragment in (22b ), but the longer reply He had
complained about something in (23b ) can?
The answer given to this question within work dating back to the mid-
1970s (following pioneering work by Jackendoff 1974, 1977a) is the fol-
lowing. Let’s suppose that merging a T/tensed auxiliary with a VP/verb
phrase forms an intermediate constituent which is larger than a word but
smaller than a complete phrase: let’s call this a T′ (T‑bar) constituent. Let’s
further assume that merging this intermediate T‑bar constituent had com-
plained about something with its subject (here, the pronominal DP he) forms
a full phrase, namely the TP/tense phrase he had complained about some-
thing, with the structure shown below:

(24) TP

DP T′
he
T VP
had
V PP
complained
P QP
about something

Merging the C/complementiser that with the TP/tense phrase in (24) above
will in turn form the CP/complementiser phrase in (25) below:

(25) CP

C TP
that
DP T′
he
T VP
had
V PP
complained
P QP
about something

If we suppose that only full phrases can function as sentence fragments, the
analysis in (25) can account for the CP that he had complained about some-
thing being able to be used as a sentence fragment in (15b ), the VP com-
plain about the bill in (20), the PP about the bill in (21), and the TP he had
complained about something in (23) – but not the T‑bar had complained
about something in (22), because this is an incomplete phrase (lacking the
subject that it requires).
 1.2 Generating syntactic structures 17

1.2.2 X-bar syntax

The tree diagram in (25) represents a classic analysis of clause structure
(widely adopted in work since the 1980s) which sees clauses as CP+TP+VP
structures. Example (25) illustrates a number of key properties of syntactic
structures assumed in this approach to syntax. Firstly, syntactic structures
are binary-branching in nature: that is to say, each mother has two daugh-
ters, in the sense that each non-terminal constituent (i.e. each constituent
which is not at the foot/bottom of the tree) branches down into two other
constituents immediately beneath it: for example, in (25), CP branches
down into C and TP; TP into DP and T‑bar; T‑bar into T and VP; VP into V
and PP; and PP into P and QP.
A second assumption embodied in (25) is that every phrase has a match-
ing head: for example, PP about something is headed by P about; VP com-
plained about something is headed by V complained; TP he had complained
about something is headed by T‑had; and CP that he had complained about
something is headed by C‑that. An alternative (but equivalent) terminology
is to say that every phrase is a projection of its head, in the sense that the
properties of a phrase are determined by the nature of its head (e.g. com-
plained about something is a verb phrase precisely because its head is the
verb complained). In these terms, PP is a projection of P, VP a projection of
V, TP a projection of T, and CP a projection of C. By extension, a constit-
uent like TP/CP etc. is sometimes referred to as tense projection/comple-
mentiser projection (as an alternative to the equivalent term tense phrase/
complementiser phrase).
A third assumption made in (25) is that there are three different types of
projection: a minimal projection (i.e. a head or word-level category X); an
intermediate projection (X′/X-bar) which is larger than a head but smaller
than a complete phrase; and a maximal projection (XP) which represents
a full phrase. Since what is novel in this approach (compared to earlier
models of syntax) is the postulation of intermediate X-bar projections, this
model has become known as X-bar Syntax.
A fourth assumption in (25) is that syntactic structures are built up by
a series of merge operations. For example, by merging P about with its
QP complement something we form the PP about something; by merging
V complained with its PP complement about something we form the VP
complained about something; and by merging C that with its TP comple-
ment he had complained about something we form the CP that he had
complained about something. However, the picture becomes more compli-
cated when we consider what the T/tense auxiliary had is merged with,
because this is merged with two different constituents. To introduce some
new terminology, T‑had first-merges with its VP complement complained
about something to form T‑bar had complained about something, and then
18 Background

T‑had second-merges with the pronominal DP he to form the TP he had

complained about something. In structures in which a head merges with two
separate constituents, the constituent which the head first-merges with is
called its complement, and the constituent which the head second-merges
with is called its specifier: thus, the VP complained about something is
the complement of T‑had in (25), and the DP he is the specifier of T‑had
(or, equivalently, VP occupies comp‑TP/the complement position in TP,
and DP occupies spec‑TP/the specifier position in TP). Complements differ
from specifiers in that complements follow the head that they are (first-)
merged with, whereas specifiers precede the head that they are (second-)
merged with.
The structure in (25) also illustrates a fifth property of syntactic
structures – namely that each merge operation involves merging a minimal
projection with a maximal projection (i.e. merging a head H with an XP).
So, for example, the set of merge operations which give rise to (25) are the
following: P merges with QP; V merges with PP; T first-merges with VP and
second-merges with DP; and C merges with TP.
A sixth property of syntactic structures illustrated by (25) is that there
are different types of maximal projection/XP. Some (like TP in 25), are
specifier + head + co p e ent structures; others (like CP, VP and PP in 25)
are head + co p e ent structures; and yet others (like QP and DP in 25)
have no complement or specifier and contain just a head (recall that he is
analysed as a DP and not just as a D because it occupies a position which
could alternatively be occupied by a DP like the guy, but not by a D like the).
A tacit assumption made in our discussion above is that the derivation
of structures (i.e. the set of operations used to form them) proceeds in a
bottom-up fashion, with lower constituents in a tree like (25) being formed
before higher ones. Thus, the initial merge operation that applies in (25) is
that P‑about merges with QP something to form PP about something; next,
V‑complained merges with this PP to form VP complained about something;
then T‑had first-merges with this VP to form T‑bar had complained about
something, and T‑had subsequently second-merges with DP he to form TP
he had complained about something; finally, C‑that merges with this TP
to form CP that he had complained about something. In each case, smaller
constituents are formed before larger ones.

1.2.3 Adjunction
Although our discussion so far has assumed that syntactic structures are
built up by merge operations, traditional work dating back to the 1960s
posits that there is a second type of structure-building operation by which
structures can be formed – namely adjunction. Two types of adjunction
are assumed in such work. One involves a head adjunction operation by
 1.2 Generating syntactic structures 19

which one head adjoins to another to form a complex head. To illustrate

this, compare the following sentences:

(26) a. He will run up the stairs

b. Up the stairs he will run
c. He will pick up the bag
d. *Up the bag he will pick

At first sight, the verb phrases run up the stairs and pick up the bag might
look as if they have the same kind of structure, and comprise a V (run/pick)
which is merged with a PP complement headed by P-up (= up the stairs/
up the bag) to form a VP (= run up the stairs/pick up the bag). And yet, the
run-VP and the pick-VP behave very differently in certain respects.
For one thing, up the stairs can be fronted/preposed in a sentence like
(26a) in order to highlight it as in (26b), but up the bag in (26c) cannot be
fronted for highlighting purposes – as we see from the ungrammaticality
of (26d). What this suggests is that up the stairs is a prepositional phrase
in (26a) since it can be preposed for emphasis in (26b): hence we can say
that run is a prepositional verb here (i.e. a verb with a PP complement).
Conversely, the fact that fronting up the bag in (26c) results in the ungram-
matical sentence (26d) suggests that it is not a PP. Rather, pick up seems to
be a complex verb comprising the verb pick and the prepositional particle
up, and thus can be termed a particle verb. One way of capturing these dif-
ferences is to suppose that up is the head of a separate PP in the VP run up
the stairs but is adjoined to the verb pick in the VP pick up the bag so that
the two verb phrases have the respective structures shown below:

(27) a. VP b. VP

V PP V DP
run
P DP V P D NP
up pick up the bag
D NP
the stairs
What the analyses in (27) claim is that up is a preposition which projects
into the prepositional phrase up the stairs in (27a), but up is a particle
adjoined to the verb pick in (27b) to form the complex verb pick up. This
correctly predicts that the string up the stairs can be fronted in (26a/26b)
because it is a PP, but not in the verb+particle structure in (26c/26d)
because the string up the bag is not a constituent there. It also predicts
that pick up (being a V, albeit one with a complex internal structure) can
be coordinated with another V (like carry in ‘John will pick up and carry
the bag’).
20 Background

A second type of adjunction operation assumed in earlier work is phrasal

adjunction – that is, an operation by which a phrase/XP is adjoined to a
constituent larger than a head. To illustrate this, consider the role played by
the italicised prepositional phrases in the clause bracketed below:

(28) I feel [that in all likelihood the president before long may face a
backlash]

Here, the PP before long modifies the T‑bar may face a backlash, and the PP
in all likelihood modifies the TP the president before long may face a back-
lash. If modifiers are adjuncts, this means that the PP before long (as used
here) adjoins to the T‑bar may face a backlash to form the even larger T‑bar
before long may face a backlash; and that the PP in all likelihood adjoins
to the TP the president before long may face a backlash to form the even
larger TP in all likelihood the president before long may face a backlash.
Given these assumptions, the that-clause bracketed in (28) above will have
the structure in (29) below:

(29) I feel CP

C TP
that
PP TP
in all likelihood
DP T′
the president
PP T′
before long
T VP
may face a backlash
On this view, phrasal adjunction involves adjoining a maximal projection
to a projection larger than a head (i.e. to an intermediate or maximal pro-
jection). Thus, the PP before long (a maximal projection) adjoins to the
T‑bar may face a backlash (an intermediate projection) to form an even
larger T‑bar; and the PP in all likelihood (a maximal projection) adjoins to
the TP the president before long may face a backlash (a maximal projection)
to form an even larger TP. Introducing some technical terminology, we can
say that the PP before long in (29) is a T‑bar adjunct, whereas the PP in all
likelihood is a TP adjunct.
An interesting property of adjunction is that adjuncts can typically be
positioned to the left or right of constituents they adjoin to. This can be
illustrated by the following set of examples (where 30a, 30b, and 30c are
alternative replies that could be given by speaker b to the question that
speaker a asks):
 1.2 Generating syntactic structures 21

(30) s p ea k e r a : What has the crash done to the car door?

s p ea k e r b : (a) Dented it

(b) Slightly dented it

(c) Dented it slightly

In (30a), the complement of the T auxiliary has is the italicised verb phrase/
VP dented it. In (30b), this VP is enlarged by adjoining slightly to the left of
it, forming the larger VP slightly dented it. And in (30c), slightly is instead
adjoined to the right of the VP dented it, forming the larger VP dented it
slightly. We can take slightly to be an ADVP/adverbial phrase here (and
not an ADV/adverb) because it can be substituted by a phrase like ever so
slightly which is clearly an ADVP. Given the assumption that slightly in (30)
is an ADVP which adjoins to VP to form a larger VP, (30b/30c) will have
the respective structures shown in (31a/31b) below:

(31) a. VP b. VP

ADVP VP VP ADVP
slightly slightly
V DP V DP
dented it dented it
In both structures, slightly is a VP adjunct – that is, a constituent that
adjoins to a VP to form an even larger VP.
Adjunction differs from merge in three main respects. One is that whereas
merge gives rise to a fixed ordering of constituents (e.g. a head first-merges
with a complement that follows it, and second-merges with a specifier that
precedes it), phrasal adjuncts can typically either precede or follow the con-
stituent they adjoin to (so that the two PPs in 29 above can alternatively be
positioned to the right of the constituents they adjoin to). The second is that
whereas merge gives rise to a constituent different in kind from the head
involved in the operation (e.g. first-merging T with its complement forms a
T‑bar; and second-merging T with its specifier forms a TP), adjunction gives
rise to a larger constituent of the same kind (e.g. adjoining slightly to a VP
forms another VP). The third is that adjunction is a potentially recursive op-
eration, in the sense that (for example) you can go on adding more and more
adjuncts to a given VP (forming successively larger and larger VPs), whereas
by contrast, a head can only have (i.e. second-merge with) a single specifier.
The recursive nature of adjunction can be illustrated by what speaker b
says in the dialogue below:

(32) s p ea k e r a : What has Sherlock Holmes decided to do with the sample?

s p ea k e r b : Examine it carefully, under a microscope, in the lab,

tomorrow
22 Background

Here, the bold-printed VP examine it has three italicised XPs adjoined to

it. The ADVP carefully adjoins to the VP examine it to form the larger VP
examine it carefully; then the PP under a microscope adjoins to the result-
ing VP to form the even larger VP examine it carefully, under a micro-
scope; next, the PP in the lab adjoins to this VP to form the even larger VP
examine it carefully, under a microscope, in the lab; and finally, the ADVP
tomorrow adjoins to the structure thereby formed to derive the structure
shown below (simplified by not showing the internal structure of the two
PPs, for lack of space):

(33) VP

VP ADVP
tomorrow
VP PP
in the lab
VP PP
under a microscope
VP ADVP
carefully
V DP
examine it

A further property of phrasal adjuncts is that they can be recursively

stacked on top of the constituent they modify in any order. Thus, a predic-
tion made by the analysis in (33) is that the adjuncts can be freely ordered
with respect to each other. And, as the sentences below show, this is indeed
the case:

(34) a. He has decided to examine it carefully, under a microscope, in the

lab, tomorrow
b. He has decided to examine it carefully, in the lab, tomorrow, under a
microscope
c. He has decided to examine it tomorrow, carefully, under a
microscope, in the lab
d. He has decided to examine it under a microscope, carefully, in the
lab, tomorrow

The free ordering of the italicised constituents following the VP exam-

ine it makes it plausible to suppose that they can all function as VP
adjuncts.
To summarise: in Module 1.2 we have seen how syntactic structures can
be formed by successive merge and adjunction operations which build up
structures (one layer at a time) in a bottom-up fashion.
You should now be able to tackle Exercise 1.2.
 1.3 Invisible structure and structural relations 23

1.3 Invisible structure and structural relations

In this module we will see that part of the structure of sentences can some-
times be invisible/inaudible, because sentence structures can contain null
constituents (also referred to as empty categories) – that is, constituents
which have morphological, syntactic and semantic properties, but have
no overt/audible phonetic form. We will then go on to look at a structural
relation (c‑command) which plays a key role in governing how syntactic
operations apply.

1.3.1 Null constituents

A key assumption underlying much contemporary work in syntax is that lan-
guages make widespread use of null constituents. One way such constituents
can arise is through ellipsis (i.e. omission of redundant material) – as in the
sentence below (where strikethrough marks material undergoing ellipsis):

(35) a. I know I should see the dentist, but I just don’t want to see the dentist
b. There are two cans of paint on the shelf, and three cans of paint on
the floor
c. If he is fond of her (and I’m sure he is fond of her), he should tell her

The sentence in (35a) contains two occurrences of the VP/verb phrase see the
dentist. To avoid repetition, the second occurrence of this VP can undergo
ellipsis and thereby be omitted, as indicated by strikethrough. Likewise,
the second occurrence of the noun phrase/NP cans of paint can undergo
ellipsis in (35b), as can the second occurrence of the adjectival phrase/
AP fond of her in (35c). I shall assume here that constituents undergoing
ellipsis are present in the syntax, but are given a silent spellout (and so are
unpronounced/silent) in the phonology. A general assumption in relevant
research is that ellipsis is subject to a Recoverability Condition specifying
that constituents can only undergo ellipsis if their contents are recoverable
(i.e. if we can recover what has been omitted). The recoverability require-
ment is satisfied in (35) because the omitted VP/NP/AP is mentioned earlier
in the sentence (the earlier occurrences being italicised).
In addition to null constituents which arise from ellipsis, there are other
constituents which are inherently null (and hence have no phonetic form).
In this respect, consider the expressions highlighted in the sentence below:

(36) He bought roses and several orchids

Here, several orchids is a QP/quantifier phrase which has been coordi-

nated (i.e. joined together by and) with the expression roses. On the tra-
ditional assumption that only likes (i.e. constituents that are alike/of the
24 Background

same type) can be coordinated, it follows that roses must have the same
status as the QP several orchids and hence roses must also be a QP. But
if so, what is the head quantifier of this QP? The answer given in much
contemporary research is that the relevant QP is headed by a null quanti-
fier with a meaning paraphrasable as ‘an unspecified quantity of’. On this
view, roses in (36) will be a QP with the structure below (where ø denotes
an inherently null item):

(37) QP

Q NP
ø roses

On the assumption that merge involves a head combining with an XP/

maximal projection, it follows that roses will be an NP/noun phrase, and
not just an N. This predicts that in place of roses we can have a larger NP
like bunches of roses. The assumption that indefinite nominal arguments are
QPs headed by an overt or null quantifier is known as the QP Hypothesis.
Now consider the sentence below:

(38) I saw John and the waiter arguing

Here, the waiter is a DP/determiner phrase (denoting a definite/specific

individual) which has been coordinated with John. Assuming that only
likes can be coordinated, it follows that John must have the same DP status
as the waiter and hence must likewise be a DP; and since the DP containing
John has no overt determiner, it must be headed by a null determiner and
have the structure below:

(39) DP

D NP
ø John
Since DPs mark definiteness/specificity, the DP in (39) will have a mean-
ing paraphrasable as ‘the specific individual John’. The assumption that all
definite expressions are DPs is known as the DP Hypothesis. Of course, it
follows from the DP Hypothesis that a definite pronoun like he will also be
a DP – as is the case in (25) above.
A further type of constituent which can be argued to contain a null head
is nominals (i.e. expressions containing or comprising a noun) like those
italicised below which are said in traditional grammar to have an adverbial
function:

(40) a. I’ll see you Sunday

b. I’ll do it my own way
c. I’m staying home today
 1.3 Invisible structure and structural relations 25

An interesting property of the italicised nominals in (40) is that they can

be substituted by PPs like those highlighted below in which the italicised
nominal is the complement of a (bold-printed) preposition:

(41) a. I’ll see you on Sunday

b. I’ll do it in my own way
c. I’m staying at home today

It is therefore plausible to take adverbial nominals like those italicised in

(40) to be PPs headed by a null counterpart of the preposition on/in/at.
Some evidence for analysing adverbial nominals as the complement of
silent prepositions comes from examples such as the following:

(42) a. Do you want to go home or to the supermarket?

b. I’m going to be busy tomorrow and at the weekend
c. Shall we stay a week, or for a couple of weeks?

The fact that in each case the italicised adverbial nominal is coordinated
with a bold-printed PP headed by an overt preposition (to/at/for) suggests
that adverbial nominals are PPs headed by a null preposition (given the
assumption that only like constituents can be coordinated).
The null P analysis of adverbial nominals can be extended to adverbial
pronominals like here/there/where/then/when/why. These are tradition-
ally analysed as adverbs, but can be taken to be pronouns function-
ing as the object of a null preposition, with here meaning ‘at/to/in this
place’, there meaning ‘at/to/in that place’, then meaning ‘at that time’,
where meaning ‘at/to/in what place’, when meaning ‘at what time’, and
why meaning ‘for what reason’. Some support for a PP analysis of such
adverbs comes from the observation that the relevant preposition can
sometimes be spelled out overtly, as shown by the bold-printed preposi-
tions below:

(43) a. Where is he going to?

b. Where is he staying at?
c. I’ve had a change of heart in the last week and I’ll tell you for why
([Link])

Such considerations suggest the hypothesis that adverbial (pro)nominals

are PPs headed by a null preposition: let’s call this the PP Hypothesis.
Given this, adverbs like there (as in He went there) or where (as in Where
did he go?) will be PPs with the respective structures shown below:

(44) a. PP b. PP

P DP P QP
ø there ø where
26 Background

In (44a), there will be a pronominal DP meaning ‘that place’, whereas in

(44b) where will be a pronominal QP meaning ‘what place’ (if we treat
what as an interrogative quantifier); and the null preposition in both
sentences (He went there/Where did he go?) can be taken to be a null
counterpart of to.
A further type of null constituent can be found in the following sentence:

(45) She said [she was feeling tired] and [that she was going to bed]

Here, that she was going to bed is a CP/complementiser phrase headed by

a C/complementiser that which serves to mark the clause it introduces as
declarative (i.e. as a statement). But on the assumption that only likes can
be coordinated, it follows that the first bracketed clause (she was feel-
ing tired) must also be a CP; and given that it contains no overt comple-
mentiser, it must be headed by a null complementiser, and thus have the
structure below:

(46) CP

C TP
ø
DP T′
she
T VP
was
V AP
feeling tired
Since this is an indicative clause, we can suppose that the null comple-
mentiser in (46) serves to mark the clause as indicative in mood, and
declarative in force (i.e. having the force of a statement).
An interesting property of CP+TP+VP structures like (46) above is
that each separate type of projection/layer of structure serves a differ-
ent function. Thus, the VP layer identifies the type of action or event
involved; the TP layer marks tense, and hence situates the action/event
in time; and the CP layer marks clause type (i.e. whether the clause
is declarative, interrogative, exclamative or imperative in force). From
this perspective, it could be argued that it is a matter of conceptual
necessity for complete/non-defective clauses to contain all three types
of projection. This means that main clauses too (even though they are
not introduced by overt complementisers like that/if/whether) must be
CP+TP+VP structures, and hence a sentence like It is raining must have
the CP+TP+VP structure below:
 1.3 Invisible structure and structural relations 27

(47) CP

C TP
ø
DP T′
it
T VP
is raining
The VP raining identifies the event taking place, the TP headed by is indi-
cates the (present) time of the event, and the CP layer serves to mark the
sentence as declarative in force, hence as making a statement. (Note that
raining is a VP rather than just a V in 47 because a head can only merge
with a maximal projection/XP, hence the T head is can only merge with a
VP, not with a V. Moreover, raining can be substituted by a larger expres-
sion which is clearly a VP, like raining cats and dogs.)
The assumption that clauses always contain a TP headed by a T which
marks (present or past) tense proves potentially problematic for a clause
like that produced by speaker b below:

(48) s p ea k e r a : What are you trying to establish?

s p ea k e r b : Whether they lied to me

The italicised clause produced by speaker b contains a CP layer headed by the

interrogative complementiser whether, and a VP layer headed by the verb lie,
but doesn’t (at first sight) seem to contain any TP layer headed by a present
or past tense T auxiliary. However, on the assumption that subjects occupy
the specifier position in TP (e.g. it is in spec‑TP in 47 above, and she is in
spec‑TP in 46), it follows that the subject they in (48 b ) must also occupy the
specifier position in a TP. But if so, what is the head of the relevant TP?
An answer given in work dating back in spirit to Chomsky (1955, 1957)
is that tensed clauses which contain no present or past tense auxiliary are
TPs whose head T constituent contains a (present or past) tense affix/Af. On
this view, the clause produced by speaker b in (48) above has the syntactic
structure shown below, where Af denotes a past tense affix:

(49) CP

C TP
whether
DP T′
they
T VP
Af
V PP
lie to me
28 Background

(The structure is simplified by not showing the internal structure of the PP

to me, because this is not relevant to our discussion here.) At the end of
the syntactic derivation (i.e. when all syntactic operations have finished
applying), the resulting structure is sent to the LF/semantic component (to
be assigned a logical form representing linguistic aspects of its meaning),
and to the PF/phonological component (to be assigned a phonetic form
determining how it is pronounced). In the PF component, the past tense
affix in T is lowered onto the head V of VP by an operation traditionally
termed Affix Hopping – as shown by the arrow below:

(50) CP

C TP
whether
DP T′
they
T VP
Af
V PP
lie+Af to me

Spellout rules in the phonology specify that a past tense affix is spelled
out on a regular verb in the default form +d (where the default form is the
form used with verbs that don’t form their past tense in some other way).
As a result, the verb is ultimately spelled out as lie+d (written as lied).
The head T position of TP (once it has been vacated by the affix) is silent/
empty/null.

1.3.2 Case-marking and c‑command

Thus far, we have concentrated on the structure of finite clauses (i.e. clauses
which contain an auxiliary or non-auxiliary verb marked for mood/tense).
But what about nonfinite clauses (i.e. clauses not containing any AUX/V
marked for mood/tense)? As a case in point, consider the structure of the
clause produced by speaker b below:

(51) speaker a : What was his suggestion?

speaker b : For us to meet up

What speaker b says in (51) can be accommodated within the CP+TP+VP

model of clause structure if we treat for as an infinitival complementiser,
and to as an infinitival T/tense particle (here with future time reference), as
shown below (simplified by not showing the internal structure of the VP
meet up):
 1.3 Invisible structure and structural relations 29

(52) CP

C TP
for
DP T′
us
T VP
to meet up
Here, the complementiser for serves not only to mark clause type (in that
it tells us that the clause it introduces denotes an irrealis event – that is,
a hypothetical one which hasn’t yet taken place and may never do so),
but also to determine the case of the clause subject, in the sense that for
requires the subject to have accusative case (hence the use of the accusative
subject us rather than its nominative counterpart we). This suggests that
complementisers case-mark the subject of their clause. But how exactly
does case assignment work?
An assumption made in work in syntax since the 1960s is that the struc-
tural relation c‑command (an abbreviation of constituent-command) plays
a key role in syntactic operations – including case-marking. This assump-
tion can be embodied in the principle below (very loosely based on Chomsky
1986a):

(53) Structure Dependence Principle

Syntactic operations and conditions are structure dependent in the
sense that they are sensitive to hierarchical structure, and canonically
involve a c‑command relation between a given constituent and
another constituent that it c‑commands. For example:
(i) movement involves a head attracting (i.e. triggering movement of)
another constituent that it c‑commands
(ii) case-marking involves a head assigning case to a constituent that it
c‑commands
(iii) anaphora involves an anaphor (like himself) being bound by (i.e.
taking its reference from) a constituent that c‑commands it
(iv) polarity items are subject to a condition that they be c‑commanded
by an appropriate licenser (e.g. any is licensed by being
c‑commanded by not in ‘I did not ask for any help’)

We can define the relation c‑command as follows:

(54) C‑command
One constituent X c‑commands another Y if the two are independent
(i.e. if neither contains the other), and if the mother of X contains Y
30 Background

In the light of (53) and (54), consider the role that c‑command plays in
case-marking in a structure like (52) above. On the assumption that the com-
plementiser for is an accusative case assigner (i.e. in traditional terms, for
is transitive), it follows that the subject DP us in (52) is assigned accusative
case by the complementiser for, because C for c‑commands DP us (in that
the mother of C for is CP, and CP contains DP us as one of its constituents).
The c‑command analysis of case-marking can be extended to finite clauses
like the CP whose structure is shown below (where ø denotes a null C):

(55) I’m sure CP

C TP
that/ø
DP T′
she
T VP
will
V DP
help him

If we follow Chomsky (2001) in supposing that finite complementisers are

nominative case assigners, it follows that the subject DP she will be assigned
nominative case by the finite complementiser that/ø which c‑commands it
in (55); and likewise, they/she will be assigned nominative case by the finite
complementisers whether/ø in (50/46) above.
Still, an issue which arises with the c‑command account of case-marking
outlined here is the following. In a structure like (55) above, the finite com-
plementiser that c‑commands two different pronominal DPs – namely she
on the one hand and him on the other. This raises the question of why that
doesn’t (wrongly) assign nominative case to him as well as (rightly) assign-
ing nominative case to she (leading to the ungrammatical outcome *I’m
sure that she will help he). The answer given in work dating back to the
1980s is to suppose that UG/Universal Grammar contains a set of principles
which regulate how grammatical operations apply. One of these (given the
fuller name the Relativised Minimality Condition by Rizzi 1990) can be
characterised informally as follows (where above means ‘c‑commanding’
and below means ‘c‑commanded by’):

(56) Minimality Condition

A constituent can only enter into a syntactic relation with the minimal
(= closest) constituent of the relevant type above or below it

One implication of (56) is that the complementiser that/ø can only case-
mark the closest pro/nominal that it c‑commands in (55) – namely she.
Conversely, a second implication is that the pronoun him can only be
 1.3 Invisible structure and structural relations 31

case-marked by the closest case assigner that c‑commands him – namely

by the transitive (i.e. accusative case-assigning) verb help. Consequently,
the effect of the Minimality Condition (56) on case assignment in (55) is to
ensure that the complementiser that/ø assigns nominative case to she, and
that the transitive verb help assigns accusative case to him.
We can broaden our discussion of clause structure and case marking by
turning to consider the type of infinitive clause bold-printed below:

(57) He is arranging to see a cardiologist and for you to see an optician

The italicised infinitival clause for you to see an optician is a CP intro-

duced by the infinitival complementiser for. But given the assumption that
only likes can be coordinated, the bold-printed clause to see a cardiologist
must also be a CP. However, if this is so, why does it appear to contain no
complementiser, and no subject? The answer given in research which dates
back in spirit to work in the 1960s is that clauses like that bold-printed in
(57) contain a null infinitival complementiser, and a null pronominal DP
subject (denoted as PRO), so that the bold-printed clause has the structure
below (where I take the indefinite article a to be an ART/article constituent
which projects into an ARTP/article phrase):

(58) CP

C TP
ø
DP T′
PRO
T VP
to
V ARTP
see
ART NP
a cardiologist
In such structures, PRO is said to be assigned null case by the null infini-
tival complementiser which c‑commands it. Since case determines how a
pronoun is spelled out (e.g. a third person masculine singular pronoun is
spelled out as he if nominative, him if accusative, and his if genitive), it is
plausible to suppose that PRO (since it has a null spellout) carries null case.
To summarise: in Module 1.3, I have argued that syntactic structures
contain null constituents which have morphological, syntactic and seman-
tic properties of their own (so that, for example, the PRO subject in 58 will
carry null case, and be third person masculine singular by virtue of refer-
ring back to he). I also argued that the syntactic relation c‑command plays
a key role in regulating how syntactic operations and conditions apply.
You should now be able to tackle Exercise 1.3.
32 Background

1.4 Movement operations

In addition to the merge and adjunction operations discussed in Module 1.2,

the derivation/formation of syntactic structures can also involve movement
operations. In this section, I outline three different types of movement oper-
ation, termed A‑bar Movement, A‑Movement and Head Movement.

1.4.1 A‑bar Movement

One type of movement operation can be illustrated in terms of the exclama-
tive sentences below (sourced by me from live, unscripted radio and TV
broadcasts, and from the internet; capitals in 59b mark emphatic stress):

(59) a. What a career that she’s had! (Laura Robson, BBC Radio 5)
b. What a race that he ran as well! (Steve Cram, BBC2 TV)
c. What a mine of useless information that I am! (Terry Wogan, BBC
Radio 2)
d. How cool that she is! ([Link])
e. How often that I give it my all to finish a task or job!
([Link])

To understand what’s going on here, let’s take a closer look at (59a), sup-
posing that what is an exclamative quantifier and that what a career is a
QP/quantifier phrase (comprising Q‑what and the ARTP/article projection a
career). This QP functions as the (direct object) complement of the verb had,
and thus originates in the same position as the phrase such a good career
in ‘She’s had such a good career’. This being so, the derivation of (59a) will
proceed as follows.
The V/verb had merges with the QP what a career to form the VP had
what a career. The T/present tense auxiliary has first-merges with this VP
to form the T‑bar has had what a career, and second-merges with the pro-
nominal DP she to form the TP she has had what a career. The C/comple-
mentiser that then merges with this TP to form the C‑bar that she has had
what a career, with the structure shown below (simplified by not showing
the internal structure of QP):

(60) C′

C TP
that
DP T′
she
T VP
has
V QP
had what a career
 1.4 Movement operations 33

In an exclamative structure like (60), C attracts an exclamative wh‑XP to

move to the edge of the projection headed by C to become its specifier. If
we suppose that (in conformity with the Structure Dependence Principle
53 above) movement involves a head attracting some other constituent
which it c‑commands, it follows that C‑that can only attract an exclama-
tive wh‑XP that it c‑commands (thereby underlining that the relation
c‑command plays a central role in syntactic operations). Furthermore,
given the Minimality Condition (56), it follows that C‑that must attract the
closest exclamative wh‑XP that it c‑commands. The QP what a career is
the only exclamative wh‑XP c‑commanded by C‑that in (60), and conse-
quently C‑that attracts this QP to become its specifier, so resulting in the
movement arrowed below:

(61) CP

QP C′
what a
career C TP
that
DP T′
she
T VP
has
V QP
had what a
career

The arrowed movement of a wh‑XP (i.e. a maximal projection compris-

ing or containing a word beginning with wh‑) is traditionally termed
Wh‑Movement: in this case, the wh‑QP what a career moves from
comp‑VP (i.e. complement position within VP) into spec‑CP (i.e. specifier
position within CP).
Under the Copy Theory of Movement developed by Chomsky (1993),
movement is a composite operation which creates a copy of the constituent
to be moved, and then merges this newly created copy in a new position at
the top of the tree. On this view, Wh‑Movement in (60) involves creating a
copy of the QP what a career and second-merging it as the specifier of that.
This generates a syntactic structure like (61) above containing two copies of
the QP what a career: the original one in comp‑VP, and a second (italicised)
copy in spec‑CP. The syntactic structure in (61) is then handed over to the
PF/phonological component, and this contains spellout rules that specify
which copies of moved constituents are pronounced/overt, and which are
silent/null. In default of any other spellout rule applying (i.e. if no other
spellout rule applies), the following rule will apply:
34 Background

(62) Default Spellout Rule/DSR

For a constituent whose spellout is not determined in some other way,
the highest copy of the constituent is overtly pronounced, and any lower
copies are silent

This means that the higher copy of the QP what a career in (61) will be
spelled out overtly, but the lower copy will be silent, resulting in (63) below
(where strikethrough marks a silent copy):

(63) What a career that she has had what a career!

A point of detail to note, however, is that most speakers of standard

varieties of English don’t like the use of that in exclamatives, and prefer to
use a null complementiser ø instead – so that in place of (63) above they
use (64) below:

(64) What a career ø she has had what a career!

I made use of a (non-standard) structure containing the overt comple-

mentiser that in (61/63) above in order to underline the point that the QP
what a career moves to a position in which it becomes the specifier of a
complementiser (i.e. it moves to spec‑CP).
Wh‑Movement of a wh‑XP to spec‑CP is also found in other types of
clause, including interrogative clauses like that bracketed in (65a) below,
and relative clauses like that bracketed in (65b):

(65) a. I wonder [which film she was watching — ]

b. That’s the film [which she was watching — ]

On the assumption that Wh‑Movement involves movement of a wh‑XP

from a position below C to spec‑CP, the bracketed clauses in (65) above
will have the respective structures shown in (66a/66b) below (simplified
by not showing the internal structure of the DP which film, for space
reasons)

(66) CP

C′

TP

T′

VP

DP C DP T V DP
a. which film ø she was watching —
b. which ø she was watching —
 1.4 Movement operations 35

In both cases, the italicised wh‑DP which (film) originates in the gap
position ( — ) and from there moves to become the specifier of the null
complementiser ø which is the head C of CP. Under the Copy Theory of
Movement, the gap position in comp‑VP will contain a silent copy of the
moved wh‑DP which film/which.
Although our discussion thus far has been concerned with Wh‑
Movement, in reality Wh‑Movement is simply one instance of a wider class
of movement operations which move a maximal projection/phrase/XP to
spec‑CP. Other operations of this type include the Focusing operation by
which the italicised nominals move to spec‑CP (to become the specifier of a
null complementiser) in sentences such as those in (67b ) below in order to
mark them as focused (i.e. conveying new information), and the Topicali-
sation operation by which the italicised nominal moves to spec‑CP (again
becoming the specifier of a null complementiser) in a sentence such as the
second one in (68) below, in order to mark it as a topic conveying familiar/
old information;

(67) s p ea k e r a : What are the subjects you enjoy most and least?
s p ea k e r b : Syntax I enjoy most. Phonetics I enjoy least

(68) He had witnessed acts of horrific violence, mutilation and torture. Such
brutality, he prayed that he would never again encounter as long as he lived

A more generic term for operations which move a constituent into spec‑CP
is A‑bar Movement (so that Wh‑Movement, Focusing and Topicalisation
are specific instances of A‑bar Movement). It is important to note here
that the suffix -bar has a very different meaning from that found in cate-
gory labels like T‑bar (where it denotes an intermediate constituent larger
than T but smaller than TP). By contrast, in the term A‑bar Movement, the
-bar suffix marks negation, so that A‑bar Movement denotes ‘movement
to a non-A position’ (i.e. to a position which is not an A‑position). This of
course raises the question of what the difference is between an A‑position
and an A‑bar position. This can be illustrated in relation to a clause such as
that produced by speaker b in the dialogue below:

(69) s p ea k e r a : What don’t you understand?

s p ea k e r b : [CP Why [TP you are blaming yourself]]

An A‑position is one which can only be occupied by an argument (e.g. the

subject or complement of a verb) but not by an adjunct, which can control
agreement, and from which an anaphor (like yourself) can be bound; by
contrast, an A‑bar position is one which is not restricted to arguments
(so can contain either an adjunct or an argument), but cannot control
agreement, and cannot bind (i.e. serve as the antecedent of) an anaphor.
36 Background

Given these criteria, the spec‑TP position occupied by you in (69 b ) is an

A‑position because the pronoun you that it contains is an argument (of
the verb blaming), it controls the agreement properties of the auxiliary are,
and it serves as the antecedent of the anaphor yourself. By contrast, the
spec‑CP position occupied by why in (69 b ) is an A‑bar position because
why is an adjunct, does not control agreement with are, and cannot bind
an anaphor.

1.4.2 A‑Movement
Alongside A‑bar Movement, we find a second type of XP movement oper-
ation which is termed A‑Movement, since it involves movement into an
A‑position (like spec‑TP). In this connection, consider the position of the
QP/quantifier phrase several suspects in the three sentences below:

(70) a. They detained several suspects

b. There were detained several suspects
c. Several suspects were detained

(70a) is an active sentence, (70b) is an expletive passive sentence (i.e. one

with the expletive pronoun there as its subject), and (70c) is a non-expletive
passive sentence. In the active sentence (70a) and the expletive passive
(70b), the QP several suspects occupies comp‑VP (i.e. the complement posi-
tion within VP). But in the non-expletive passive (70c), the QP several
suspects occupies spec‑TP, in that it serves as the specifier of the past tense
auxiliary were. But how does the QP get from the comp‑VP position that
it occupies in (70a, 70b) to the spec‑TP position that it occupies in (70c)?
The answer given in work dating back in spirit to Chomsky (1955, 1957) is
that the QP several suspects originates in comp‑VP and from there moves
to spec‑TP via an operation traditionally referred to as Passivisation. Let’s
look more closely at passives like (70b, 70c).
The V/verb detained merges with the QP several suspects to form the VP
detained several suspects. The past tense T auxiliary were first-merges with
this VP to form the T‑bar were detained several suspects, which has the
structure below:

(71) T′

T VP
were
V QP
detained several suspects
T constituents carry a feature (traditionally called an EPP feature)
requiring them to second-merge with an appropriate nominal or pro-
nominal specifier/subject. One way of satisfying this EPP requirement is
 1.4 Movement operations 37

by merging the expletive pronominal DP there in spec‑TP. Merging the

resulting TP with a null declarative complementiser forms the expletive
passive structure below:

(72) CP

C TP
ø
DP T′
there
T VP
were
V QP
detained several suspects

The reason why the feature requiring T to have a specifier/subject is termed

an EPP feature is that in earlier work its function was seen as being to
satisfy an Extended Projection Principle (= EPP) requiring T to project not
only into T‑bar, but also to have an extended projection into a TP formed
by second-merging T with a nominal or pronominal subject.
However, an alternative way of satisfying the EPP feature on T is for T
to attract a nominal or pronominal DP/QP to become its subject. In con-
sequence of the Structure Dependence Principle (53) and the Minimality
Condition (56), T will attract the closest QP/DP that it c‑commands; and in
the case of the structure in (71) above, this is the QP several suspects. Ac-
cordingly, this QP moves from comp‑VP to spec‑TP, in the manner shown
by the arrow below:

(73) CP

C TP
ø
QP T′
several
suspects T VP
were
V QP
detained several
suspects

This movement is traditionally termed Passivisation, but (as we will see

in more detail in Chapter 2) it is a particular subtype of a more general
movement operation known as A‑Movement, because it involves move-
ment into an A‑position (spec‑TP). Under the Copy Theory of Movement,
Passivisation results in a structure containing two copies of the QP several
suspects (a lower one in comp‑VP, and a higher one in spec‑TP), but in
38 Background

accordance with the Default Spellout Rule (62), only the higher copy is
overtly spelled out in the phonology.

1.4.3 Head Movement

A more general conclusion to be drawn from our discussion so far in this
module is that there are two distinct types of XP movement operation
found in the grammars of natural languages: A‑bar Movement of an XP
from a position below C to spec‑CP; and A‑Movement of an XP from a
position below T to spec‑TP. However, it is not only XP constituents (i.e.
maximal projections/phrases) which can undergo movement, as the bold-
printed auxiliaries below illustrate:

(74) a. What colour have you chosen?

b. No leniency will the judge show

The italicised phrases what colour/no leniency are QPs headed by the
interrogative/negative quantifiers what/no. These QPs originate as the
complement of the verb chosen/show, as we see from the correspond-
ing in situ sentences (i.e. sentences in which no movement takes place)
below:

(75) a. You have chosen what colour?

b. The judge will show no leniency

In each of the two sentences in (75), the interrogative/negative QP origi-

nates in comp‑VP (as the complement of the verb chosen/show) and from
there moves to spec‑CP in sentences like (74). But in both (74a) and (74b),
movement of the italicised QP to spec‑CP triggers concomitant Inversion of
the bold-printed auxiliary have/will – in the sense that the auxiliary moves
from a position following the subject you/the judge in (75) to a position
preceding the subject in (74). But what kind of movement operation does
Auxiliary Inversion involve?
Since the bold-printed auxiliaries in (75) occupy the canonical
(post-subject) T position for finite auxiliaries, and since (via Inversion)
they end up in the same (pre-subject) C position in (74) as is occu-
pied by the complementiser that in (61) above, it is plausible to suppose
that Auxiliary Inversion involves T‑to‑C movement (i.e. movement of
an auxiliary from the T position following the subject to the C position
preceding the subject). On one classic implementation of this idea, the
derivation of (74a) proceeds as follows. The V chosen merges with the
QP what colour to form the VP chosen what colour. The (present tense) T
auxiliary have first-merges with this VP to form the T‑bar have chosen
what colour, and second-merges with the pronominal DP you to form the
 1.4 Movement operations 39

TP you have chosen what colour. The resulting TP is then merged with
a null complementiser (C‑ø) to form the C‑bar ø you have chosen what
colour. The null complementiser/C attracts the interrogative QP what
colour to become its specifier, and concomitantly attracts the T auxiliary
have to adjoin to the null C, so resulting in the two movement operations
arrowed below:

(76) CP

QP C′
what
colour C TP

T C DP T′
have ø you
T VP
have
V QP
chosen what
colour

On this view, movement of the interrogative QP what colour to spec‑CP

triggers concomitant movement of the auxiliary have from T to adjoin to
a null complementiser in C. In each case, only the highest/italicised copy
of the moved QP in spec‑CP (and the highest/bold-printed copy of the
auxiliary in C) is spelled out overtly in the phonology, in conformity with
the Default Spellout Rule (62). Note that, in conformity with the Structure
Dependence Principle (53i) and the Minimality Condition (56), C‑ø in (76)
attracts the closest auxiliary that it c‑commands (namely the T auxiliary
have) to adjoin to C‑ø.
The assumption that Auxiliary Inversion in sentences like (74) is trig-
gered when C has an interrogative or negative specifier raises the question
of what triggers Inversion of the bold-printed auxiliary in a yes–no ques-
tion like:

(77) Have you chosen the colour?

A traditional answer given in research dating back decades is that in yes–

no questions, C has a null yes–no question operator as its specifier, and it
is this operator which (by virtue of being interrogative) triggers Auxiliary
Inversion in main clause questions. It is called a yes–no question opera-
tor because it has the semantic function of operating on a proposition to
convert it into a yes–no question which questions the truth value of the
40 Background

proposition (e.g. 77 asks whether it is true that you have chosen the colour).
If we denote this abstract/silent yes–no question operator as OpYNQ (and
take it to be an ADVP like why/how/when/where), the yes–no question in
(77) will have the derivation below:

(78) CP

ADVP C′
OpYNQ
C TP
have+ø
DP T′
you
T VP
have
V DP
chosen the colour
On the assumption that a C with an interrogative or negative specifier
triggers Auxiliary Inversion, the T auxiliary have in (78) will undergo
Inversion and thereby adjoin to the null complementiser in C (as shown in
simplified form above), because the complementiser has an interrogative
operator as its specifier. Given the Default Spellout Rule (62), only the
higher copy of the auxiliary have (in C) will be overtly spelled out in the
phonology.
An interesting question arising from the yes–no question operator analy
sis is what function the operator serves. A plausible answer is that having
an interrogative operator on the edge of CP (i.e. inside CP but outside TP)
provides a way of signalling what type a given clause is (e.g. whether it is
interrogative, or exclamative, etc. in force). Reasoning along these lines,
let’s suppose that in the semantic component of the grammar, clauses are
typed (i.e. interpreted as being of a specific type) in accordance with the
Clause Typing Conditions in (79) below:

(79) Clause Typing Conditions

(i) A clause is interpreted as being of a given type (interrogative or
exclamative etc.) if it has a peripheral clause-typing specifier of the
relevant type
(ii) An indicative clause is interpreted as declarative in type by default if
it contains no peripheral clause-typing specifier

Note that a peripheral constituent is one positioned on the edge of the

clause periphery; and we can give a more concrete idea of what the clause
periphery comprises if we represent the structure in (78) in terms of the
highly simplified labelled brackets below:
 1.4 Movement operations 41

(80) [CP OpYNQ [C have+ø] [TP you [T have] chosen the colour]]

The periphery of a clause/CP comprises that part of the structure posi-

tioned above TP (i.e. the part boxed in 80): consequently, the bold-printed
constituents positioned on the edge of CP (i.e. inside CP but outside TP)
are in the clause periphery in (80). Given the clause-typing conditions in
(79), it follows from (79i) that the presence of the interrogative yes–no
question operator in spec‑CP in (78/80) will type the relevant clause as
having the force of a yes–no question. And in essentially the same way,
the presence of the exclamative QP what a career in spec‑CP in (61) above
will type (61) as an exclamative clause. Likewise, the presence of the
interrogative phrase which film in spec‑CP in (66a) will type the relevant
clause as a wh‑question. By contrast, the absence of any clause-typing
constituent in the clause periphery (= on the edge of CP) in structures
like (72) and (73) above will ensure that the relevant clauses are typed as
declarative in force by default (via 79ii), assuming that the null C marks
indicative mood.
Returning now to the main theme of this module: movement oper-
ations typically move a constituent (either a maximal projection XP
or a head X) to the edge of a functional projection. So, for example,
A‑bar Movement moves an XP from a position below C into spec‑CP;
A‑Movement moves an XP from a position below T into spec‑TP; and
Head Movement (in the case of Auxiliary Inversion) moves a head from
T into C. Still, if movement operations move an X or XP to the edge of
CP or TP, we’d expect to find a further type of Head Movement operation
by which a head moves from V to T. And indeed it has long been argued
that English does indeed have just such a V-to-T Movement operation
by which the verb b e (when used as a copula – i.e. a verb which links a
subject to a non-verbal predicate) can raise from the head V position of
VP into the head T position of TP. This can be illustrated by sentences
such as the following:

(81) a. He may not be at home

b. He is not at home

In (81a) the copular verb b e occupies the head V position of VP, and hence
follows not (which is taken in much recent work to be the specifier of a
NEGP/negative phrase constituent positioned between TP and VP). In (81b),
the verb be (in the guise of is) ends up in front of not, so can be taken to
have moved from V into T. Let’s take a closer look at what is involved in
this type of movement.
42 Background

Consider first (81a), which we can take to have the following structure:

(82) CP

C TP
ø
DP T′
he
T NEGP
may
ADVP NEG′
not
NEG VP
ø
V PP
be at home

In (81a/82) be occupies the head V position of VP, and hence follows the
NEGP containing the negative adverb not. But things are different in three
respects in (81b): firstly, there is no auxiliary like may in T; secondly, in
place of be we find the finite form is; and thirdly, is precedes not (whereas
in 81a/82, be follows not). How can we account for these differences?
A plausible answer is to suppose that b e has auxiliary-like properties
which enable it to raise from the V position which it occupies in (82) above
to adjoin to a present tense affix in T – in the manner shown by the arrow
in (83) below:

(83) CP

C TP
ø
DP T′
he
T NEGP
BE+Af
ADVP NEG′
not
NEG VP
ø
V PP
BE at home

Since the affix in T is present tense (and third person singular by agree-
ment with he), the resulting complex T head b e +Af will be spelled out in
the phonology as the irregular third person singular present tense form is
(and the overall sentence as 81b He is not at home). On the assumption that
only an affix that needs to attach to a verbal stem can attract the verb b e to
adjoin to it (and that T contains such an affix but NEG does not), movement
 1.4 Movement operations 43

from V to T will satisfy the Minimality Condition (56) (which requires a

constituent to move to the closest appropriate position above it), since the
verb be moves to adjoin to the closest head above it that contains the kind
of affix that attaches to a verb – namely T. It will also satisfy the Structure
Dependence Principle (53i), since (prior to movement) T‑Af c‑commands
the V‑be constituent that it attracts to adjoin to it.
Once be has raised to T, the b e +Af head thereby formed can undergo
Auxiliary Inversion in questions and thereby move to C, as in a yes–no
question such as Is he at home? with the (simplified) superficial structure
shown below:

(84) CP

ADVP C′
OpYNQ
C TP
BE+Af+ø
DP T′
he
T VP
BE+Af
V PP
BE at home
The verb be first moves from V to T (adjoining to a null present tense affix
in T), and then the resulting b e +Af structure (ultimately spelled out as is)
moves from T to C (adjoining to a null complementiser in C). The V and
T positions are empty after movement (and so end up as gaps), but under
the Copy Theory of Movement, these gaps represent null copies of moved
items.
In earlier varieties of English (e.g. in Shakespearean English), lexical/
main verbs could also raise from V to T, so resulting in auxiliariless finite
clauses like those in (85) below, where the (bold-printed) main verb is po-
sitioned in front of not: cf.

(85) a. I care not for her (Thurio, Two Gentlemen of Verona, [Link])
b. He heard not that (Julia, Two Gentlemen of Verona, [Link])
c. My master seeks not me (Speed, Two Gentlemen of Verona, I.i)
d. I know not where to hide my head (Trinculo, The Tempest, [Link])

In such sentences, the bold-printed verb raises from V (across the interven-
ing NEGP containing not) to adjoin to a present or past tense affix in T.
However, in present-day English, only a handful of verbs can now undergo
movement to T – namely items which have little (if any) lexical semantic
content and thus are auxiliary-like, such as b e and (in some varieties) some
uses of have .
44 Background

To summarise: I have argued in Module 1.4 that there are three main types of
syntactic movement operation: (i) A‑bar Movement of an XP from a position
below C into spec‑CP; (ii) A‑Movement of an XP from a position below T into
spec‑TP; and (iii) Head Movement from a lower head position to adjoin to a
higher one (e.g. from T to adjoin to null C, or from V to adjoin to an affix in T).
You should now be able to tackle Exercise 1.4.

1.5 Blocking mechanisms

Ever since the 1960s, linguists have argued that grammars need to contain
mechanisms which block illicit syntactic operations (termed constraints),
and also mechanisms which block illicit superficial syntactic structures
(termed filters). In this module, we’ll look at how these two different mech-
anisms can be used to block illicit operations and structures.

1.5.1 Constraints
Pioneering work by Chomsky (1964a) and Ross (1967) argued that principles
of UG/Universal Grammar impose constraints (i.e. restrictions) on the types
of syntactic operation which are permissible. Using a pictureseque metaphor,
Ross (1967) argued that certain types of syntactic structure are islands in
the sense that any constituent inside an island is marooned there and can’t
be got off the island by any movement operation. He proposed a family of
constraints on movement which can be characterised informally as follows:

(86) Island Constraints

No movement operation can extract any constituent out of an island
such as the following:
(i) a subject
(ii) an adjunct
(iii) a clause contained inside an NP (forming a complex NP)
(iv) a coordinate structure

These island constraints have the effect of blocking movement from the
gap to the italicised position in sentences such as the following:

(87) a. *The motorway, everyone knows that [the closure of — ] will cause
chaos
b. *The FBI, the CIA were working on the case [independently of — ]
c. *This course, there are [students [who hate — ]]
d. *The princess, the police are guarding [the prince and — ]

The reason is the following. In (87a), the italicised topic is extracted out
of a bracketed subject, in violation of the Subject Island Constraint (86i).
 1.5 Blocking mechanisms 45

In (87b), the italicised topic is extracted out of a bracketed ADVP which is

an adjunct to the VP working on the case, in violation of the Adjunct Island
Constraint (86ii). In (87c), the italicised topic is extracted out of a relative
clause who hate which is contained inside the NP students who hate, in
violation of the Complex NP Constraint (86iii). And in (87d), the italicised
topic is extracted out of a bracketed coordinate structure, in violation of the
Coordinate Structure Constraint (86iv).
In the decades since Ross’s original work, considerable research has been
undertaken into constraints. One strand of this research has sought to con-
flate several of Ross’s Constraints into a single, more general constraint. A
constraint along these lines (proposed by Cattell 1976, Cinque 1978, and
Huang 1982) can be characterised informally as follows:

(88) Constraint on Extraction Domains/CED

Only complements allow material to be extracted out of them, not
specifiers or adjuncts

Since subjects are specifiers, the ungrammaticality of (87a) can be attrib-

uted to a CED violation because the DP the motorway is extracted out of a
DP (the closure of the motorway) which is the specifier of the T auxiliary
will. And since (87b) involves extraction out of an adjunct, it too gives rise
to a CED violation. Similarly, CED is violated in (87c), because this course
is extracted out of a bracketed who-clause inside a bracketed NP containing
students, and in (87d), because the princess is extracted out of a bracketed
coordinate structure.
Much other work on constraints has focused on locality conditions which
impose restrictions on how far constituents can move in any single move-
ment operation. The Minimality Condition (outlined in 56 above) is one
of these, requiring a moved constituent to move to the closest appropriate
landing site above it. Another locality constraint (dating back to work by
Chomsky 1998) can be characterised informally as follows (where impene
trable means ‘inaccessible’, and – to be strictly accurate – below means
‘c‑commanded by’, and above means ‘c‑commanding’):

(89) Impenetrability Condition

Anything below the head C of a CP is impenetrable to anything above
the CP

The condition/constraint in (89) has the effect of barring syntactic opera-

tions from applying across the edge of a CP (i.e. across the head C of a CP
and any specifier that it has). The effect of the constraint can be illustrated
below:

(90) *They consider [CP that [TP each other should apologise]]
46 Background

In this sentence, each other is a reciprocal anaphor, and like all anaphors
it needs to be bound by (i.e. linked to) a local/nearby antecedent which
c‑commands it. However, while they c‑commands each other in (90), the
resulting sentence is ill-formed. Why should this be? A plausible answer
is that the Impenetrability Condition (89) prevents they from binding each
other across the edge of the intervening (bracketed) CP containing that.
And since there is no other possible binder/antecedent for each other, the
resulting sentence (90) is ungrammatical.
Locality constraints like the Minimality and Impenetrability Conditions
have important implications for how long-distance (i.e. cross-clausal)
movement works – as can be illustrated in relation to:

(91) What do you think [that he wants]?

Here, the interrogative wh‑QP what originates as the complement of the

verb wants, and subsequently (via Wh‑Movement/A‑bar Movement) moves
out of the bracketed wants-clause to the front of the think-clause. But how
does this long-distance movement (out of one clause into the other) take
place? At first sight, it might seem plausible to suppose that movement
takes place in a single step, in the manner shown in highly simplified form
below:

(92) [CP2 What [C2 do] you think [CP1 [C1 that] he wants what]]

That is, the wh‑QP what moves directly from being the complement of
the verb want to becoming the specifier of the inverted auxiliary do in the
main clause; and in accordance with the Default Spellout Rule (62), only
the higher copy of what (in spec‑CP2) is spelled out overtly.
However, the single-step movement arrowed in (92) violates locality
constraints on movement. For one thing, since Wh‑Movement involves
movement of a wh‑XP to become the specifier of a CP above it, the Mini-
mality Condition (56) requires a wh‑moved XP to become the specifier of
the closest CP above it – and in the case of (92), the closest CP above/to the
left of the position in which what originates (as the complement of want) is
CP1, and not CP2. Secondly, the single-step movement in (92) also violates
the Impenetrability Condition (89) – which bars movement across the edge
of an intervening CP – by moving what across the complementiser that
on the edge of CP1. In short, the single-step movement in (92) violates two
separate locality conditions (Minimality and Impenetrability), and hence
is untenable. And yet, the resulting sentence (What do you think that he
wants?) is fully grammatical. How come?
The answer is that although the single-step movement arrowed in (92)
is ruled out by locality constraints, there is an alternative way of deriving
 1.5 Blocking mechanisms 47

the sentence which satisfies locality requirements. This is to suppose that

the wh‑QP what moves in two successive steps, moving first to become the
specifier of the lower clause (CP1), and then to become the specifier of the
higher clause (CP2) – as shown by the two numbered arrows below:

(93) [CP2 What [C2 do] you think [CP1 what [C1 that] he wants what]]

(2) (1)

The wh‑QP what originates in comp‑VP (as the complement of the verb
want), and first of all moves into spec‑CP1: since CP1 is the closest CP
above the comp‑VP position in which what originates, this movement
does not violate the Minimality Condition (56). Nor does movement
(1) in (93) violate the Impenetrability Condition (89), since it involves
movement to a position inside CP1 (not to a position outside/above CP1).
Subsequent movement of what from spec‑CP1 to spec‑CP2 does not vio-
late the Minimality Condition either, since in moving to spec‑CP2 what
moves to the closest spec‑CP position above spec‑CP1. Nor does move-
ment (2) induce an impenetrability violation, since what moves from a
position above (not below) C1 to a position above CP1. In short, locality
(minimality and impenetrability) constraints permit long-distance move-
ment to take place only in a successive-cyclic fashion (i.e. in a succes-
sion of small steps, with the relevant constituent moving first to the
front of the lowest clause, then to the front of the next lowest clause,
and so on).
A different kind of constraint on movement is posited by Wexler & Culi-
cover (1980: 119), and can be outlined informally as follows:

(94) Freezing Principle

The constituents of a moved phrase are frozen internally within (and so
cannot be extracted out of) the moved phrase

How the principle works can be illustrated by the examples below:

(95) a. Which hotel do you think he is staying in?

b. In which hotel do you think he is staying?
c. *Which hotel do you think in he is staying?

In these sentences the wh‑phrase which hotel originates as the comple-

ment of the preposition in. In (95a), which hotel undergoes Wh‑Movement
on its own, stranding the preposition in at the end of the sentence. In
(95b), the preposition in is dragged (or, to use a technical term, pied-piped)
along with which hotel, so that the whole PP in which hotel undergoes
Wh‑Movement. Given the arguments presented in our discussion of (91–93)
above that long/cross-clausal Wh‑Movement is successive-cyclic, the PP in
48 Background

which hotel will move first to the edge of the CP in the embedded clause
before moving to the edge of the CP in the main clause in (95b). But this
assumption raises the question of what would prevent us from moving the
PP in which hotel to the front of the embedded clause CP1, and then moving
the wh‑phrase which hotel on its own to the front of the main clause CP2,
leaving the preposition in stranded in the bold-printed position below, at
the beginning of the embedded clause CP1:

(96)

[CP2 which hotel [C do] you think [CP1 in which hotel[C ø] he is staying in which hotel]]

Why does the derivation in (96) crash and lead to the ungrammatical out-
come in (95c)? The answer is that the movement shown by the upper arrow
violates the Freezing Principle (94), because which hotel has been extracted
out of a larger phrase (= the PP in which hotel) which has itself undergone
movement, in that the PP in which hotel undergoes the Wh‑Movement
operation indicated by the lower arrow.
Yet another kind of constraint on movement operations can be illustrated
by the following contrast:

(97) a. Nobody can have no secrets

b. *No secrets can nobody have

Let’s suppose that the expressions nobody and no secrets are both negative
QPs/quantifier phrases. Let’s also suppose that (97b) involves the negative
QP no secrets moving from its initial position as the complement of have to
the specifier position in the CP at the beginning of the sentence (and that
this triggers concomitant Auxiliary Inversion) – as shown by the arrow
below:

(98) *[CP No secrets [C can] nobody have —]

So why is the resulting sentence (97b) ungrammatical? The answer lies in a

constraint formulated by Abels (2012: 247) in the following terms:

(99) Intervention Condition

Likes cannot cross likes (likes = ‘constituents which are alike’)

The problem posed by the arrowed movement in (98) is that it involves one
negative QP (no secrets) crossing another negative QP (nobody), in violation
of the Intervention Condition (99). We will encounter a number of other
constraints on movement operations in later chapters of the book.
 1.5 Blocking mechanisms 49

1.5.2 Filters
In work dating back to the 1960s, it has been argued that (in addition to
imposing constraints on syntactic operations), natural language grammars
also impose constraints on the types of superficial syntactic structure that
they allow. These constraints on superficial structures have been termed
filters, because they have the effect of filtering out structures at the end of
a syntactic derivation that do not meet the relevant requirements.
By way of illustration, consider the clauses bracketed below (from sen-
tences I recorded from live, unscripted radio/TV broadcasts):

(100) a. It shows [what a fantastic squad that they have] (Gordon Strachan,
ITV)
b. You just wonder [how long that America can hang on] (Gary Lineker,
BBC1 TV)
c. I’m aware of the speed [with which that they work] (Tim Vickery,
BBC Radio 5)

In each case, the bracketed clause is a CP whose edge comprises a wh‑XP

(exclamative in the case of what a fantastic squad in 100a, interrogative
in the case of how long in 100b, and relative in the case of with which in
100c) immediately followed by the complementiser that – as shown in the
structures below (simplified by not showing the internal structure of TP):

(101) CP

C′

wh-XP C TP
a. It just shows what a fantastic squad that they have
b. You just wonder how long that America can hang on
c. I’m aware of the speed with which that they work
However, while I have recorded hundreds of similar wh+that structures
from radio, TV and the internet, it is nonetheless true that speakers of
standard varieties of English typically don’t like such structures, and regard
them as ungrammatical. Why?
A traditional answer (dating back to work by Chomsky & Lasnik 1977)
is that standard varieties of English impose a requirement on superficial
syntactic structures that can be outlined informally as follows:

(102) Doubly Filled COMP Filter/DFCF

At the end of a syntactic derivation, any superficial structure in which
the edge of a projection headed by an overt complementiser is doubly
filled (i.e. contains some other overt constituent) is filtered out as
ungrammatical
50 Background

(Note that the edge of a projection XP comprises its head X and any other
constituent/s of XP above/c‑commanding X; most commonly, the edge of XP
will comprise the head X and any specifier that it has.) Speakers with DFCF in
their grammar find wh+that structures like (100/101) ungrammatical, because
the edge of the relevant CP is doubly filled (by virtue of comprising an overt
complementiser that with an overt wh‑XP as its specifier). Such speakers
avoid violating DFCF by using a structure in which the relevant CP contains
a null C‑ø in place of C‑that, resulting in the structures bracketed below:

(103) a. It shows [CP what a fantastic squad [C ø] they have]

b. You just wonder [CP how long [C ø] America can hang on]
c. I’m aware of the speed [CP with which [C ø] they work]

Since the clauses bracketed in (103) do not contain an overt complement-

iser, they do not give rise to any DFCF violation. By contrast, speakers (like
me) of varieties of English which do not have DFCF in their grammar accept
both wh+that CPs like those in (100/101), and wh+ø CPs like those in (103).
Thus, varieties of English can differ with respect to whether specific filters
(like DFCF) operate in their grammar.
A second type of filter which operates in standard varieties of English
(again proposed by Chomsky & Lasnik 1977: 451) can be characterised in-
formally as follows:

(104) COMP‑Trace Filter/CTF

Any structure at the end of a syntactic derivation in which an overt
COMP/complementiser is immediately adjacent to and c‑commands a
trace (i.e. a gap/null copy left behind by a moved constituent) is filtered
out as ill-formed

Types of structure which CTF allows/disallows can be illustrated below:

(105) a. Who do you think that he should help — ?

b. *Who do you think that — should help him?

In (105a), the fronted wh‑QP who originates as the complement of the verb
help, and from there moves first to spec‑CP in its own clause, and then to
spec‑CP in the main clause – as shown in simplified schematic form below:

(106) Who do you think CP

QP C′
who
C TP
that
DP T′
he should help who
 1.5 Blocking mechanisms 51

The two-step Wh‑Movement operation arrowed above is forced by the

Minimality Condition (56) and the Impenetrability Condition (89). It results
in a superficial syntactic structure (106) which does not violate the COMP-
Trace Filter (104), because the italicised constituent immediately following
(and c‑commanded by) the bold-printed complementiser that is the overt
pronoun he, not a trace/null copy of a moved constituent. Since the result-
ing structure (106) does not violate CTF, the corresponding sentence (105a)
is grammatical.
But now compare what happens in (105b). Here, Wh‑Movement results
in the pronoun who undergoing the (two-step) Wh‑Movement operation
arrowed below:

(107) Who do you think CP

QP C′
who
C TP
that
QP T′
who should help him

More specifically, who first moves from spec‑TP to spec‑CP in the comple-
ment clause (i.e. the should-clause), and then moves to spec‑CP in the main
clause. But this results in a structure (107) in which the overt complement-
iser that immediately precedes and c‑commands a trace/null copy of the
moved pronoun who. This results in a violation of the COMP‑Trace Filter
(104) which causes the resulting structure to be filtered out as ungram-
matical. The corresponding sentence (105b) is therefore ungrammatical, by
virtue of violating CTF.
While CTF violations lead to ungrammaticality in standard varieties of
English, it should be noted that there are non-standard varieties which
do not have CTF in their grammar, and thus allow structures such as
those in (108) below, where an italicised constituent has been extract-
ed out of a position (marked by the gap) and leaves behind a trace/null
copy which immediately follows and is c‑commanded by a bold-printed
complementiser:

(108) a. Michael Owen, I’m sure that — will be touted round many clubs (Stan
Collymore, talkSPORT Radio)
b. Who did you feel that — ran the dressing room? (Michael Vaughan,
BBC Radio 5)
c. What are we hoping [that — can come from this]? (Jonny Gould,
talkSPORT Radio)
52 Background

d. I probably played under 25 managers, which I’m not sure whether —

is a good thing (Danny Mills, BBC Radio 5)
e. Chamak it looks as if — will be joining West Ham (Jason Cundy,
talkSPORT Radio)

This suggests that the restrictions on the class of admissible syntactic struc-
tures imposed by filters may be variety-specific. It is also interesting to
note that both of the filters discussed here (the Doubly Filled COMP Filter
in 102, and the COMP‑Trace Filter in 104) relate to properties of overt
complementisers.
To summarise: Module 1.5 has shown how constraints limit the ways
in which syntactic operations can apply, and how filters limit the kinds of
superficial syntactic structure which they can give rise to.
You should now be able to tackle Exercise 1.5.

1.6 Summary

This chapter has set out to provide an outline of key concepts and con-
structs underlying the classic CP+TP+VP model of clause structure devel-
oped in the 1980s.
In Module 1.1, I compared descriptive and prescriptive approaches to syn-
tax (§1.1.1), noting that contemporary work in Linguistics takes a purely
descriptive stance. I went on (§1.1.2) to discuss the relative merits of three
different types of data used in syntactic analysis: corpus data, introspective
grammaticality judgements, and experimental data. I then turned (§1.1.3) to
outline the key role played by categories and features in syntactic analysis.
In Module 1.2, we looked at the nature of syntactic structures, and the
operations which generate/build them. We began by looking at the opera-
tion merge (§1.2.1), by which a head combines with a maximal projection
to form a larger constituent. We saw that in more complex structures, a
head X first-merges with a following complement to form an intermediate
X-bar projection, and then second-merges with a specifier to form a maxi-
mal projection XP. So, for example, in a sentence like He is helping her, the
verb helping merges with the pronoun/pronominal DP her to form the VP
helping her. The (present tense) T auxiliary is first-merges with this VP to
form the T‑bar is helping her, and then T‑is second-merges with the pro-
noun/pronominal DP he to form the TP He is helping her. As this example
illustrates, syntactic structures are formed in a bottom-up fashion, in the
sense that lower parts of the structure are formed before higher parts. This
particular model of syntax is known as X-bar syntax (§1.2.2). In addition
to merge, syntactic structures can also be generated by a separate type of
 1.6 Summary 53

operation termed adjunction (§1.2.3). There are two types of adjunction.

One type (called head adjunction) involves adjoining one head to anoth-
er to form a complex head: for example, by adjoining the prepositional
particle up to the verb pick we form the complex verb pick up. A second
type of adjunction (termed phrasal adjunction) involves adjoining a phrase
(more precisely, a maximal projection/XP) to a constituent larger than a
head (i.e. to an intermediate projection/X-bar or to a maximal projection/
XP) to form an even larger constituent of the same type. For example, if we
take the verb phrase/VP examine it and adjoin the prepositional phrase/PP
under a microscope to the right of this VP, we form the larger VP italicised
in ‘He will examine it under a microscope’. We briefly looked at a number
of differences between adjunction and merge.
In Module 1.3, we saw that parts of the structure of a sentence can be
invisible, because syntactic structures can contain null constituents. One
way this can come about is through ellipsis (a process whereby a constitu-
ent present in the syntax is silent/unpronounced in the phonology): for
example, in order to avoid unnecessary repetition, we can omit the <brack-
eted> occurrence of the noun phrase cans of beer in a sentence like ‘I had
four cans of beer, and you had three <cans of beer>’ (§1.3.1). Some types
of head can be inherently null: for example, in a sentence like ‘People don’t
trust Boris’, people can be argued to be a QP/quantifier phrase headed by
an inherently null quantifier Q‑ø with a meaning paraphrasable as ‘an un-
specified number of’; likewise, Boris can be argued to be a DP headed by
an inherently null definite determiner (so that Boris has a meaning roughly
paraphrasable as ‘the specific individual Boris’). And the overall sentence
can be argued to be a CP headed by an inherently null complementiser
which serves the function of marking the clause as indicative in mood and
declarative in force (i.e. as having the force of a statement).
I went on to argue that hierarchical structure (and the relation c‑command
in particular) plays a central role in syntactic operations and conditions
(§1.3.2), and we saw that this can be captured in terms of the following
principle:

(53) Structure Dependence Principle

Syntactic operations and conditions are structure dependent in the
sense that they are sensitive to hierarchical structure, and canonically
involve a c‑command relation between a given constituent and another
constituent that it c‑commands. For example:
(i) movement involves a head attracting (i.e. triggering movement of)
another constituent that it c‑commands
(ii) case-marking involves a head assigning case to a constituent that it
c‑commands
54 Background

(iii) anaphora involves an anaphor (like himself) being bound by (i.e.

taking its reference from) a constituent that c‑commands it
(iv) polarity items are subject to a condition that they be c‑commanded
by an appropriate licenser (e.g. any is licensed by being
c‑commanded by not in ‘I did not ask for any help’)

The relation c‑command was defined as follows:

(54) C‑command
One constituent X c‑commands another Y if the two are independent
(i.e. neither contains the other), and if the mother of X contains Y

I also argued that other principles of Universal Grammar determine how

syntactic operations apply, and that these include the following condition
(where above means ‘c‑commanding’, and below means ‘c‑commanded
by’):

(56) Minimality Condition

A constituent can only enter into a syntactic relation with the minimal
(= closest) constituent of the relevant type above or below it

We can illustrate the role played by (53, 54, 56) by looking at how
case-marking applies in relation to a sentence such as I am keen for them
to help him. Each of the pronouns will be assigned case by the clos-
est case-assigning head which c‑commands it. Consequently, him will be
assigned accusative case by the transitive verb help; them will be assigned
accusative case by the transitive complementiser for; and I will be assigned
nominative case by the invisible null complementiser above it (given the
assumption made here that finite clauses are CPs headed by an overt or
null complementiser).
In Module 1.4, we examined three different types of movement oper-
ation. We began by looking at A‑bar Movement (§1.4.1) and saw that this
involves C attracting an XP below it (e.g. one that is wh‑marked, or topic-
alised or focused) to become its specifier (and thereby move to spec‑CP). In
consequence of the Structure Dependence Principle (53) and the Minimality
Condition (56), C will attract the closest XP of the relevant type to become
its specifier. I noted that under the Copy Theory of Movement developed by
Chomsky (1993), movement is a composite operation which creates a copy
of the constituent to be moved, and then merges this newly created copy
in a new position at the top of the tree. Spellout rules in the phonology
determine which copies of a moved constituent are spelled out overtly and
which are given a silent/null spellout (and so are unpronounced). In default
of any other spellout rule applying (i.e. if no other spellout rule applies), the
following spellout rule will apply:
 1.6 Summary 55

(62) Default Spellout Rule/DSR

For a constituent whose spellout is not determined in some other way,
the highest copy of the constituent is overtly pronounced, and any lower
copies are silent

We went on to look at a second type of XP‑movement operation known

as A‑Movement, whereby (in conformity with the Structure Dependence
Principle and the Minimality Condition) T attracts the closest DP/QP that
it c‑commands to become its specifier (§1.4.2). Thus, in a passive sentence
such as Several demonstrators were arrested, the QP several demonstra-
tors originates in comp‑VP as the complement of the verb arrested, and is
attracted by the T auxiliary were to move to spec‑TP, so deriving Several
demonstrators were arrested — (where the position marked by the gap con-
tains a null copy of the passivised QP several demonstrators).
We also looked at a third type of movement operation termed Head Move-
ment (§1.4.3), whereby a head attracts another head that it c‑commands
to adjoin to it. One particular instance of Head Movement is Auxiliary
Inversion – an operation by which a null complementiser (C‑ø) that has an
interrogative or negative XP as its specifier attracts a T auxiliary to adjoin
to C‑ø. Thus, in a sentence such as What decision have you come to? the
QP what decision originates as the complement of the preposition to, and
the auxiliary have originates in the head T position of TP (as in the in situ
structure You have come to what decision?). A null complementiser above
TP attracts (a copy of) the QP what decision to move to spec‑CP by A‑bar
Movement, and concomitantly attracts (a copy of) the auxiliary have to
undergo Inversion and thereby adjoin to C‑ø via Head Movement, as shown
in simplified form below (where the upper/dotted arrow marks Head Move-
ment, and the lower/dashed arrow marks A‑bar Movement):

(109) [CP what decision [C have+ø] [TP you [T have] come to what decision]]

Lower copies of the two moved constituents receive a null spellout in the
phonology (marked by strikethrough in 109) via the Default Spellout Rule (62).
An interesting question posed by the claim that a null C triggers Inver-
sion (i.e. attracts a T constituent below C to adjoin to C) when C has an
interrogative or negative specifier is how to deal with yes–no questions like
Have you come to a decision? An answer dating back in spirit to work in the
1960s is that in such cases, C has a null yes–no question operator (OpYNQ)
as its specifier, and that (by virtue of having an interrogative specifier), C
triggers Auxiliary Inversion (i.e. attracts a present/past tense auxiliary in
T to adjoin to it) in essentially the same way as in a wh‑question. On this
56 Background

view, the periphery of a yes–no question like Have you come to a decision?
will be a CP with the structure shown in highly simplified form below:

(110) [CP OpYNQ [C have+ø] you come to a decision]

However, a question raised by this analysis is what semantic function is

served by the yes–no question operator on the edge of CP. The answer
given here is that the yes–no question operator serves to type the relevant
clause as being a yes–no question, in accordance with the Clause Typing
Conditions below:

(79) Clause Typing Conditions

(i) A clause is interpreted as being of a given type (interrogative or
exclamative etc.) if it has a peripheral clause-typing specifier of the
relevant type
(ii) An indicative clause is interpreted as declarative in type by default if
it contains no peripheral clause-typing specifier

It would then follow that the presence of the interrogative yes–no ques-
tion operator in spec‑CP in (110) will type the relevant clause as having
the force of a yes–no question. Likewise, the presence of an interrogative
wh‑XP in spec‑CP in a structure like (109) will type the relevant clause as
a wh‑question; and the presence of an exclamative wh‑XP in spec‑CP in a
sentence like What a great time we had! will type the sentence as exclama-
tive in force.
In Module 1.5, I noted that work dating back to the 1960s has posited
that grammars contain mechanisms which block illicit syntactic operations
and structures. One such mechanism are constraints which govern how
syntactic operations can (and can’t) apply. An early constraint of this kind
(deriving from work by Cattell 1976, Cinque 1978, and Huang 1982) is the
following:

(88) Constraint on Extraction Domains/CED

Only complements allow material to be extracted out of them, not
specifiers or adjuncts

Another (dating back to work by Wexler & Culicover 1980: 119) is the
following:

(94) Freezing Principle

The constituents of a moved phrase are frozen internally within (and so
cannot be extracted out of) the moved phrase

In addition, numerous locality constraints have been proposed which

limit how far a constituent can move in any single movement operation
 1.6 Summary 57

(§1.5.1). These include the Minimality Condition in (56) above, and the
Impenetrability Condition in (89) below:

(89) Impenetrability Condition

Anything below the head C of a CP is impenetrable to anything above
the CP

The Minimality Condition requires a moved constituent to move to the

closest position of the appropriate kind above it, while the Impenetrability
Condition bars movement across the edge of an intervening CP (i.e. across a
complementiser and any specifier that it has). In consequence of such con-
straints, Wh‑Movement in a sentence like Where do you think that he has
gone? moves where from its initial position as the complement of gone in
two successive steps: where moves first to spec‑CP in the that-clause, and
then from there into spec‑CP in the do-clause.
A different kind of constraint on movement operations is the following:

(99) Intervention Condition

Likes cannot cross likes (likes = ‘constituents which are alike’):

In a sentence like Nobody would do nothing if they saw someone drowning,

this prevents the negative QP nothing from being moved across the nega-
tive QP nobody to form *Nothing would nobody do if they saw someone
drowning.
In addition to imposing restrictions on syntactic operations, grammars
also impose restrictions on the class of syntactic structures they produce:
such restrictions are termed filters (since they filter out illicit structures as
ungrammatical). Two such filters (discussed in §1.5.2) are the following:

(102) Doubly Filled COMP Filter/DFCF

At the end of a syntactic derivation, any superficial structure in which
the edge of a projection headed by an overt complementiser is doubly
filled (i.e. contains some other overt constituent) is filtered out as
ungrammatical

(104) COMP‑Trace Filter/CTF

Any structure at the end of a syntactic derivation in which an overt
COMP/complementiser is immediately adjacent to and c‑commands a
trace (i.e. a gap/null copy left behind by a moved constituent) is filtered
out as ill-formed

DFCF filters out for example, ‘I’m not sure [CP how amenable that [TP he
will be]]’, because the edge of the bracketed CP is doubly filled, in that
it contains an overt complementiser (that) with an overt specifier (how
amenable). CTF filters out for example, ‘What do you think [CP that [TP — has
58 Background

made him change his mind]]?’ because the overt complementiser that
immediately precedes and c‑commands a gap which represents a trace/
null copy of the moved wh‑QP what. As noted in the text discussion, both
DFCF and CTF concern properties of complementisers, and both seem to be
variety-specific (in the sense that they hold in standard English, but not in
some other varieties).

1.7 Bibliographical notes

For examples of the prescriptive approach to grammar discussed in Module

1.1, see Heffer (2011), and Gwynne (2013). On the who/whom distinction,
see Sobin (1997) and Lasnik & Sobin (2000). For discussion of the rel-
ative merits and reliability of different methods of collecting linguistic
data, see Schütze (1996), Cowart (1997), Hoffmann (2011: ch.2), Weskott &
Fanselow (2011), and Schütze & Sprouse (2014). On the drawbacks of col-
lecting linguistic data from a corpus (or from the web), see Schütze (2009).
For a defence of the use of introspective judgement data rather than other
sources of data, see Newmeyer (2003, 2005, 2006a, 2006b) on usage-based
data, and Sprouse (2011), Sprouse, Schütze & Almeida (2013), and Sprouse
& Almeida (2011a, 2011b, 2012, 2013) on experimental data. For evidence
that the grammaticality judgements of linguists about sentences may dif-
fer from those of non-linguists, see Dąbrowska (2010). On grammatical
categories in English see Quirk et al. (1985), Biber et al. (1999), Huddleston
& Pullum (2002), and Radford (2016: 60–110). On determiners, see Giusti
(1997), Adger (2003), Spinillo (2004), and Isac (2006). On pronouns see
Cardinaletti & Starke (1999), Wiltschko (1998, 2001), Freidin & Vergnaud
(2001), and Déchaine & Wiltschko (2002). On treating personal pronouns
as DPs, see Postal (1966), Abney (1987), Longobardi (1994), and Lyons
(1999). On grammatical features and their relation to grammatical catego-
ries, see Chomsky (1965, 1970), Radford (1997a, 1997b, 2004a, 2016) and
Ramat (2014); on gender features in English, see Namai (2000).
The claim made in Module 1.2 that phrases are formed by a merge oper-
ation which combines heads with complements and specifiers dates back to
the model of X-bar Syntax developed in Jackendoff (1974, 1977a, 1977b):
for a textbook introduction, see Radford (1981, 1988). The idea that syntac-
tic structures are headed dates back in spirit to Bloomfield (1935), though it
should be noted that Bloomfield assumed that some structures are headless.
The idea that all syntactic structure is binary-branching was put forward
and defended in Kayne (1981, 1984): for a review of evidence for and
against binary-branching structures, see Stradmann (2013). The idea that
clauses contain a TP projection headed by a T constituent containing an
 1.7 Bibliographical notes 59

auxiliary or affix in finite clauses and the particle to in infinitive clauses

dates back in spirit to Chomsky (1981, 1986b); the idea that (non-defec-
tive) clauses also contain a CP projection is due to Fassi-Fehri (1980),
Chomsky (1981, 1986b) and Stowell (1981). In earlier work (e.g. Chomsky
1957, 1965) nominal expressions like the nose or several people were taken
to have the status of NP/Noun Phrase; however, in much work since the
mid-1960s they have been analysed as having the status of DP/determiner
phrase or QP/quantifier phrase constituents (see e.g. Fukui 1986, Fukui
& Speas 1986, Hellan 1986, Abney 1987, Löbel 1989, and Olsen 1989 –
though it should be noted that the DP analysis is not without problems: see
for example Bruening 2009, and Bruening et al. 2015). An alternative to
the bottom-up model of syntax developed by Chomsky and outlined in the
main text is the top-down model presented in Phillips (1996, 2003), Rich-
ards (1999), Shan & Barker (2006), and Bruening (2014): but see Chomsky
(2007) for arguments that the bottom-up/top-down dichotomy may be a
false one.
For a textbook account of (head and phrasal) adjunction operations,
see Radford (2020: 108–151). The type of adjunction operation dealt with
here was termed ‘Chomsky-adjunction’ by Ross (1967: 259, fn. 12) on the
grounds that it was ‘first noted by Chomsky’, and this term was widely
used in earlier work. The treatment of verb+particle structures as complex
verbs involving a particle adjoined to a verb dates back to Chomsky (1955)
and Fraser (1965), and was widely adopted in much subsequent work (e.g.
Radford 1988, Johnson 1991, Neeleman 1994, Haegeman & Guéron 1999,
Dehé 2002, etc.): for a critical review of a range of analyses of verb+parti-
cle structures, see Capelle (2005), Elenbaas (2009), and Larsen (2014). The
analysis of TP adjuncts presented here dates back in spirit to earlier work
by Ross (1967: 289), Baltin (1978: 156, and 1982: 16–22), Lasnik & Saito
(1992: 81) and Haegeman & Guéron (1999: 102). The analysis of PPs and
ADVPs which modify VPs as VP adjuncts dates back to Emonds (1970:
187), who argues that the italicised PP in ‘John told us the game was can-
celled at the airport’ is an adjunct to the VP told us the game was cancelled.
For a more detailed textbook account of various types of null constit-
uent discussed in Module 1.3, see Radford (2016: 172–244), or Radford
(2020: 152–208). The idea that certain types of null constituent can arise
via ellipsis dates back to Hankamer (1971), Hankamer & Sag (1976), Sag
(1980), Kuno (1981), Pesetsky (1982a), Hardt (1993), McCawley (1993),
Lobeck (1995), Schwarz (1999, 2000), Merchant (1999, 2001, 2002, 2003,
2004a, 2008a, 2008b, 2013a, 2013b), Johnson (2001) etc. The idea that
bare nominals (i.e. noun expressions not modified by an overt determiner
or quantifier) can contain a null determiner/null quantifier has a long
history, dating back to a suggestion made by Chomsky (1965: 108) which
60 Background

was taken up and extended in later work by Abney (1987), Bernstein (1993,
2001), and Longobardi (1994, 1996, 2001). On the nature of quantified ex-
pressions see Löbel (1989), Giusti (1991a), and Shlonsky (1991). On deter-
miners and determiner phrases, see Abney (1987), Bernstein (1993), Giusti
(1997), Alexiadou & Wilder (1998), Lyons (1999), Zamparelli (2000), Gro-
hmann & Haegeman (2002), and Ticio (2003, 2005). The idea that apparent-
ly complementiserless clauses contain a null complementiser dates back in
spirit several decades (see e.g. Stockwell et al. 1973: 599): for discussion of
factors governing the use of null or overt complementisers in finite claus-
es, see Hawkins (2001), and Nomura (2006). The idea that certain infini-
tive clauses are introduced by null complementiser dates back to Bresnan
(1970); on treating infinitive complements of verbs like want as CPs headed
by a null counterpart of for, see Sawada (1995) and Nomura (2006). The
Affix Hopping account of verb morphology dates back in spirit to Chom-
sky (1955, 1957), and is revised in Lasnik (1981). The c‑command relation
dates back to Klima (1964): a defence of its primitive nature can be found
in Frank & Vijay-Shanker (2001). However, it should be noted that Bruen-
ing (2014) argues for alternative accounts of many phenomena traditional-
ly handled in terms of c‑command. On the idea that finite complementisers
assign nominative case, see Chomsky (2001). The idea that PRO subjects in
English carry null case derives from Chomsky & Lasnik (1993), Chomsky
(1995), and Martin (1996, 2001). On the claim that null case is assigned
to a PRO subject by C, see Rizzi (1997: 304) and Collins (2005: 104). On
the (Relativised) Minimality Condition, see Rizzi (1990, 2003), Chomsky
(1995), Friedmann et al. (2009), and Takano (2013).
For a more detailed textbook account of the A‑bar Movement operation
discussed in Module 1.4, see Radford (2016: 317–477), or Radford (2020:
265–331). The idea that A‑bar Movement involves C attracting a (specific
kind of) XP to second-merge with C and thereby move to spec‑CP derives
from Chomsky (1981, 1986b, 1998, 2008). The Copy Theory of Movement
derives from Chomsky (1995); on why (generally) only the highest copy of a
moved constituent is overtly spelled out, see Nunes (1995, 1999, 2001, 2004).
On the A‑Movement operation by which T attracts a DP/QP to move from a
position below T to spec‑TP (e.g. in passives), see the detailed discussion in
Chapter 2 of this book. On the nature of the EPP feature of T which requires
T to have a specifier/subject, see Chomsky (1982, 1995, 1998), Rothstein
(1983), Alexiadou & Anagnostopoulou (1998), Déprez (2000), Grohmann et
al. (2000), Holmberg (2000), Kiss (2001), Bošković (2002), Roberts & Roussou
(2002), Rosengren (2002), Haeberli (2003), Miyagawa (2005, 2006a, 2010),
van Craenenbroeck & den Dikken (2006), Landau (2007), and Lin (2011).
For a textbook account of Head Movement, see Radford (2016: 245–316),
or Radford (2020: 209–64). The idea that Auxiliary Inversion involves a
 1.7 Bibliographical notes 61

null C attracting a finite auxiliary to adjoin to it dates back to Chom-

sky (1981); on earlier analyses, see Radford (2018: 11–17). The view that
yes–no questions contain a null yes–no question operator is put forward
in Katz & Postal (1964), Bresnan (1970), Larson (1985), Grimshaw (1993),
Roberts (1993a), den Dikken (2006), and Haegeman (2012). On the notion of
clause typing, see Cheng (1997); the idea that indicative clauses are inter-
preted as declarative by default is suggested by Roberts & Roussou (2002).
The analysis of Be ‑Raising whereby finite forms of the copula b e involve b e
raising from V to adjoin to a tense affix in T dates back in spirit to Klima
(1964); see Nomura (2006: 303–14) for discussion of av e / e ‑Raising in
subjunctive clauses.
The idea outlined in Module 1.5 that syntactic operations are regulated
by putatively universal principles/conditions/constraints dates back to pi-
oneering work by Chomsky (1964a) and Ross (1967). The Impenetrability
Condition (preventing operations applying across the edge of a CP) is a sim-
plified version of the Phase Impenetrability Condition of Chomsky (1998);
its historical roots lie in the Subjacency Condition of Chomsky (1973)
(amended by Rizzi 1982), the Barrierhood Condition of Chomsky (1986b),
and the Relativised Minimality Condition of Rizzi (1990). On the Inter-
vention Condition (barring likes from crossing likes), see Starke (2001),
Rizzi (2004a), Endo (2007a), Friedmann et al. (2009), Haegeman (2012), and
Abels (2012).
The idea that grammars contain filters which rule out illicit superficial
structures at the end of a syntactic derivation dates back to pioneering
work by Perlmutter (1968, 1971). On the Doubly Filled COMP Filter, see
Chomsky & Lasnik (1977), Seppänen & Trotta (2000), Koopman (2000),
Koopman & Szabolsci (2000), Zwicky (2002), Collins (2007), Baltin (2010),
Collins & Radford (2015), and Radford (2018). On the COMP‑Trace Filter,
see Perlmutter (1968, 1971), Chomsky & Lasnik (1977), Pesetsky (1982b),
Sobin (1987, 2002, 2009), Browning (1996), Szczegielniak (1999), Rous-
sou (2000), Pesetsky & Torrego (2001, 2004), Ackema & Neeleman (2003),
Kandybowicz (2006), Lohndal (2009), Ackema (2010), Ha (2010), Llinàs-
Grau & Fernández Sánchez (2013), Sato & Dobashi (2013), Erlewine (2014,
2017a), Bošković (2016), Pesetsky (2016), Sobin (2016), and Douglas (2017).
The approach adopted in this book can be described as representational
(in that more emphasis is placed on getting the precise details of syntactic
structures right), rather than algorithmic (in the sense that less attention
is devoted to the mechanical details and formal properties of operations
like Merge, Agree, and Adjoin that apply in the course of derivations): see
Müller (2015) on this distinction. For an alternative (more formal and algo-
rithmic) approach, see especially Collins & Stabler (2016), and additionally
Chomsky (1995, 2013, 2014, 2015), Citko (2005), Di Sciullo & Isac (2008),
62 Background

Hornstein & Nunes (2008), Hornstein (2009), Fukui (2011), Larson (2014),
Merchant (2014, 2019), Epstein et al. (2015), Collins (2017), Müller (2017),
and Zyman (2022).

1.8 Workbook

In the exercises below, an [M] after an example sentence indicates that a

model answer is provided in the free-to-download Students’ Answerbook
(and in the Teachers’ Answerbook as well). You will find it helpful to look at
this model answer before attempting to tackle other examples in the exer-
cise. Some other exercise examples have an [S] after them, indicating that
they are for self-study: this means that after reading the relevant module
(and those preceding it), and looking at the helpful hints and model answer
associated with the exercise, students should be able to analyse these self-
study examples on their own, and then compare their answers with those
provided in the Students’ Answerbook. The remaining exercise examples
(i.e. those not marked with [M] or [S] after them) are intended for teachers
to use as the basis for hands-on problem-solving work in seminars, classes,
or assignments: answers to all of these non-self-study examples are pro-
vided in the Teachers’ Answerbook. Note that Exercise 1.1 tests you on
Module 1.1, Exercise 1.2 on Module 1.2, and so on; and that each exercise
in a given chapter presupposes familiarity with material covered in earlier
chapters and in earlier modules of the same chapter.

EXERCISE 1.1

This exercise is designed to be attempted after reading Module 1.1.

Comment on the potential problems posed by relying on usage data to

determine grammaticality in relation to the following example sentences
(recorded by me from live, unscripted radio/TV broadcasts). Say what (if
anything) is odd about them, and what potential problems they highlight
with usage data:

1 We look at Arsenal and we see what the squad they have (Steve Claridge, BBC
Radio 5) [M]
2 They have not withstood up to the pressure (Graeme Souness, Sky Sports TV)
3 The speed that which they go at these days leaves no margin for error (Darren
Gough, talkSPORT Radio) [S]
4 As Gary Neville was saying at the start of our coverage, that it has usually been
easy for Manchester United at White Hart Lane (Martin Tyler, Sky Sports TV)
 1.8 Workbook 63

5 One of the first things you’re taught as a batsman is to don’t leave your
ground while the ball is live (Dominic Cork, talkSPORT Radio) [S]
6 First of all, let’s me bring you an update on this dreadful fire we heard about
(Rachel Burden, BBC Radio 5) [S]
7 What would’ve England asked from you? (Dominic Cork, talkSPORT Radio) [S]
8 You do feel sorry for him as the rain run downs the back of his neck (Adrian
Chiles, ITV)
9 We entirely agree that there needs to be changes (Angela Knight, BBC Radio 5)
10 This is someone who now finds themself in line for considerable
compensation (Reporter, BBC Radio 5)

In addition, comment on the issue arising from using experimental meth-

ods to determine grammaticality in relation to the experiment outlined
below. As noted in Module 1.1, in everyday English we find structures
like Who by? where the preposition by and its object who are ‘inverted’.
These are said to be instances of Swiping (an acronym for ‘Sluicing With
Inverted Prepositions In North Germanic’, where Sluicing is a type of ellip-
sis, involving deleting redundant material like that marked by strikethrough
in a sentence such as He is coming back, but I’m not sure when he is coming
back). Numerous published studies have claimed that Swiping is only gram-
matical in wh‑questions (like the when-question in the preceding sentence),
not in other types of clause. In order to verify whether this is indeed the
case, I asked Philip Hofmeister (and he kindly agreed) to include Swiping
sentences like those in 11 below in an online experiment he ran to test the
acceptability of a wide range of different types of sentence structure. Each
of the sentences in 11 below contains a bracketed elliptical clause in which
the italicised object of a bold-printed preposition is positioned in front of
the preposition:

11 a. I wonder where she bought that awful tie, and [who for] [M]
b. Whenever we argue and [whatever about], we always make up
afterwards
c. I’m amazed at how much he bought on eBay and [how little for]!
d. I’m going away, but [not long for]
e. The fewer presents we send and [the fewer people to], the happier
Scrooge will be
f. So hard has he trained and [so long for] that he is sure to win the race

The italicised object of the preposition is interrogative in 11a, unconditional

(in the terminology of Rawlins 2008) in 11b, exclamative in 11c, nega-
tive in 11d, comparative in 11e, and consecutive in 11f. Sixty-four native
speakers of American English were asked to rate the acceptability of a set
of sentences including those in 11 using a seven-point numerical scale on
64 Background

which 7 denotes ‘extremely natural’ and 1 denotes ‘extremely unnatural’.

The mean scores for each sentence were as follows: 5.13 for the interroga-
tive 11a; 4.95 for the unconditional 11b; 4.50 for the exclamative 11c; 4.27
for the negative 11d; 3.92 for the comparative 11e; and 2.69 for the con-
secutive 11f. (The above results are reported in Radford & Iwasaki 2015;
note that this is an abridged version of an exercise in Radford 2020.)

Helpful hints
In relation to the sentences in 1–10, bear in mind the possibility that
some of them may involve accidental production errors, perhaps resulting
from a memory lapse of some kind (e.g. forgetting how you started your
sentence). This sometimes results in the formation of what are termed
‘blends’, that is, structures in which parts of two different (but related)
structures are blended/combined together. Bear in mind, too, that some
structures which strike you as odd may be used in other varieties of
English, even if not in your own. In relation to the acceptability scores
reported for each of the sentences in 11, discuss the issue they pose for the
assumption in much work in traditional syntax that sentences are either
grammatical or ungrammatical (in that a sentence marked by a prefixed
star/asterisk is ungrammatical, and one not so marked is grammatical).

EXERCISE 1.2

This exercise is designed to be attempted after reading Module 1.2 and the
module preceding it.

Discuss the derivation of the sentences produced by speaker b below:

1 s p eaker a : What did he say about your offer?

s p eaker b : That he will carefully consider it [M]
2 s p eaker a : What did he ask you?
s p eaker b : If she really did enjoy the meal
3 s p eaker a : What do the police say?
s p eaker b : That in all probability someone is lying [S]
4 s p eaker a : What did John say?
s p eaker b : That he may reluctantly give up the job soon [S]
5 s p eaker a : What did you ask the engineers?
s p eaker b : Whether they had previously run into this problem
6 s p eaker a : What was the advice you gave to Boris?
s p eaker b : For him immediately to apologise unreservedly
 1.8 Workbook 65

Helpful hints
Discuss the set of merge and adjunction operations used to form the
relevant sentences, and say how merge and adjunction differ from each
other. Comment on any additional points of interest arising in relation
to your analysis – for example, why in 1 the personal pronouns he/it
are DPs, and why carefully is an ADVP. In 3, treat in all probability as
a PP, but don’t concern yourself with its internal structure. In 4 and 5,
be mindful of the distinction between particle verbs (i.e. verbs like pick
up in which a prepositional particle is an adjunct to the verb pick), and
prepositional verbs (i.e. verbs like run which have a prepositional phrase
as their complement in a sentence like ‘He ran up the hill’): note that
prepositional verbs allow the P to be fronted along with its complement
(as in ‘Up which hill did he run?’), and also allow straight to intervene
between V and P (as in ‘He will run straight up the hill’), but particle
verbs don’t exhibit either of these properties.

EXERCISE 1.3

This exercise is designed to be attempted after reading Module 1.3 and the
modules preceding it.

Discuss the derivation of the sentences below, paying particular attention

to the syntax of the null constituents in 1–3, and to the case-marking of
(overt or null) personal pronouns in 4–6:

1 John will now undergo tests [M]

2 Students are going home this weekend [S]
3 Boris thinks politicians should show leadership
4 She intended to examine them thoroughly [S]
5 He must be keen for us to vote for him
6 I would like you to be nice to them

Helpful hints
Bear in mind the following assumptions made in the book. (Non-
defective) clauses are CPs headed by an overt or null C/complementiser;
in relation to 6, note that like in this kind of use is traditionally classed as
a for‑deletion verb (i.e. a verb which selects a CP complement headed by
a null counterpart of the complementiser for). Finite TPs are headed either
by a finite auxiliary in T, or by an affix in T which lowers onto the closest
verb below it in the phonology by Affix Hopping; infinitival TPs are
headed by the infinitive particle to. Definite nominal arguments are DPs
66 Background

headed by an overt or null determiner, and indefinite nominal arguments

can be QPs headed by a null quantifier (with a meaning paraphrasable
as ‘an unspecified quantity of’). Adverbial (pro)nominals are PPs headed
by a null preposition; since prepositional phrases can often contain the
modifiers straight/right (as in ‘straight to the point’ or ‘right on time’) one
test of PP status is to see whether a given constituent can be modified
by straight or right. (Pro)nominal constituents are assigned case by the
closest case-assigning head above/c‑commanding them, with transitive
heads assigning accusative case, finite complementisers assigning
nominative case, and a null infinitival complementiser assigning null case
(to a PRO subject).

EXERCISE 1.4

This exercise is designed to be attempted after reading Module 1.4 and the
modules preceding it.

Discuss the merge, movement and adjunction operations involved in the

derivation of the following sentences:

1 What chaos there was! [M]

2 No concessions will they make
3 Where can he now go? [S]
4 How was the parcel wrapped?
5 Were there any complaints? [S]
6 Could anyone imagine what a struggle it was?

Helpful hints
Bear in mind the following assumptions made in the book. Expletive there
sentences involve a pronominal DP there in spec‑TP, with a QP associate
that originates in comp‑VP (as the complement of an appropriate verb).
A‑bar Movement (e.g. Wh‑Movement) involves movement of an XP
from a position below C into spec‑CP; A‑Movement (e.g. Passivisation)
involves movement of an XP from a position below T into spec‑TP; Head
Movement involves movement of a lower head to adjoin to a higher one
(e.g. an auxiliary moving from T to C, or verb like b e from V to T). A
null C with an interrogative or negative specifier triggers concomitant
Auxiliary Inversion/T‑to‑C movement); yes–no questions contain a null
interrogative operator in spec‑CP. Movement involves copying, and (in
consequence of the Default Spellout Condition), only the highest copies of
moved constituents are (generally) overtly spelled out in the phonology.
 1.8 Workbook 67

In consequence of Clause Typing Conditions, clauses are interpreted as

being interrogative/exclamative etc. in force if they contain a clause-
typing specifier of the relevant kind in their periphery (i.e. in spec‑CP);
indicative clauses with no clause-typing specifier in their periphery are
interpreted as declarative in type by default.
In 6, take the phrase what a struggle to be an exclamative QP, headed
by the quantifier what. Its internal structure need not concern you: but
one possibility is that the phrase what a struggle is derived by merging
the ART/article a with the NP struggle to form the ARTP/article phrase
a struggle, and then merging the exclamative quantifier what with this
ARTP to form the QP what a struggle.

EXERCISE 1.5

This exercise is designed to be attempted after reading Module 1.5 and the
modules preceding it.

Discuss the derivation of the sentences below, and the role played by
constraints and/or filters in accounting for their ungrammaticality:

1 *Who did he ask if was fired? [M]

2 *What document were which parts of redacted? [S]
3 *Few things would I say never should you try [S]
4 *Who did you wonder where had gone to?
5 *The weekend, she asked who he was with at
6 *The accident, they know which car that the driver of caused

Helpful hints
Consider the role played by constraints in blocking illicit syntactic
operations, especially the following (N.B. in 8 and 9, below = c‑commanded
by; above = c‑commanding):

7 Constraint on Extraction Domains/CED

Only complements allow material to be extracted out of them, not specifiers
or adjuncts
8 Minimality Condition
A constituent can only enter into a syntactic relation with the minimal
(= closest) constituent of the relevant type above or below it
9 Impenetrability Condition
Anything below the head C of a CP is impenetrable to anything above the CP
10 Intervention Condition
Likes cannot cross likes (likes = ‘constituents which are alike’)
68 Background

Consider also the possible role of filters in blocking illicit superficial

syntactic structures, especially the following:

11 Doubly Filled COMP Filter/DFCF

At the end of a syntactic derivation, any superficial structure in which the
edge of a projection headed by an overt complementiser is doubly filled (i.e.
contains some other overt constituent) is filtered out as ungrammatical
12 COMP‑Trace Filter/CTF
Any structure at the end of a syntactic derivation in which an overt COMP/
complementiser is immediately adjacent to and c‑commands a trace (i.e. a
gap/null copy left behind by a moved constituent) is filtered out as ill-formed

Assume that all yes–no question clauses (including those introduced by

the complementiser if/whether) have a null yes–no question operator
(OpYNQ) in spec‑CP. Take the word of to be a K/case particle which merges
with a nominal or pronominal complement to form a KP/case phrase
(rather than treating of as a preposition – though this assumption does
not affect the analysis in any material way). For the purposes of this
exercise, treat which? as an interrogative Q/quantifier (like what?) –
though a case could be made for treating which? as an interrogative D/
determiner because it is definite/specific in sense, in that for example,
which book means ‘which book out of a specific set’. In relation to 5,
consider the possibility that there is an Anti-Iteration Filter which bars
iteration of likes: that is, it bars superficial structures containing likes
(e.g. words of the same form or type) which are immediately adjacent to
each other.
2 A‑Movement

2.0 Overview

In this chapter, we take a closer look at the syntax of subjects. Subjects

generally occupy the specifier position within TP and remain there
(unless the subject undergoes A‑bar Movement and so moves to spec‑CP).
However, here I will argue that subjects originate internally within the
verb phrase as arguments of verbs, and are subsequently raised into the
specifier position within TP, with the relevant movement operation being
triggered by an EPP feature carried by T. Since spec‑TP is an A‑position
(i.e. an agreement-related position occupied by an argument which can
bind an anaphor), the operation by which subjects move into spec‑TP is
known as A‑Movement.

2.1 VP-Internal Subject Hypothesis

This module outlines evidence for positing that subjects originate intern-
ally within VPs (a claim known as the VP-Internal Subject Hypothesis), and
from there subsequently raise to their superficial position in spec‑TP

2.1.1 Subjects in Belfast and Standard English

Let’s begin our discussion of the syntax of subjects by looking at some
interesting data from Belfast English. Alongside Standard English struc-
tures like (1a, 1b), Belfast English also has structures like (2a, 2b):

(1) a. Some students should get distinctions

b. Lots of students have missed the classes

(2) a. There should some students get distinctions

b. There have lots of students missed the classes

Sentences like (2a, 2b) are called expletive structures because they contain
the expletive pronominal DP there. (The fact that there is not a locative DP
meaning ‘that place’ in this kind of use is shown by the impossibility of
replacing it by locative here or questioning it by the interrogative locative
70 A‑Movement

where? or focusing it by assigning it contrastive stress.) For the time being,

let’s concentrate on the derivation of Belfast English sentences like (2a, 2b)
before turning to consider the derivation of Standard English sentences like
(1a, 1b).
One question to ask about the sentences in (2a, 2b) is where the expletive
pronoun there is positioned. Since there immediately precedes the tensed
auxiliary should/have, a reasonable conjecture is that there is the subject/
specifier of should/have and hence occupies spec‑TP. If so, we’d expect
to find that the auxiliary can move in front of the expletive subject (via
Auxiliary Inversion, i.e. T‑to‑C Movement) in questions – and this is indeed
the case, as the Belfast English sentences in (3) below illustrate:

(3) a. Should there some students get distinctions?

b. Have there lots of students missed the classes?

But where are the underlined quantified expressions some students/lots

of students positioned in (3)? Since they precede the verbs get/missed
and since subjects precede verbs, it seems reasonable to conclude that
the expressions some students/lots of students function as the subjects of
the verbs get/missed and (since subjects are typically specifiers) occupy
spec‑VP (i.e. specifier position within VP). If these assumptions are cor-
rect, (2a) will have the structure (4) below (simplified by not showing
the internal structure of the expressions some students/distinctions: we
can take both of these to be QP/Quantifier Phrase constituents, headed
by the overt quantifier some in one case and by a null quantifier [Q ø] in
the other):

(4) CP

C TP
ø
DP T′
there
T VP
should
QP V′
some students
V QP
get ø distinctions

The analysis in (4) claims that the sentence contains two subjects/specifiers:
the pronominal DP there is the specifier (and subject) of should, and the QP
some students is the specifier (and subject) of get. The two subjects serve
two different functions: the expression some students has the semantic
function of denoting the recipients of the distinctions, and so could be said
to be the semantic subject of the sentence; by contrast, the pronoun there
 2.1 VP Internal Subject Hypothesis 71

satisfies the syntactic requirement of the EPP feature on T (requiring T to

have a specifier which is a noun or pronoun expression), and so could be
said to be the syntactic subject of the sentence.
Given the assumptions in (4), sentence (2a) will be derived as follows.
The verb get first-merges with the QP ø distinctions to form the V‑bar
(intermediate verb expression) get ø distinctions, and get then second-
merges with the QP some students (itself formed by merging the quantifier
some with the NP students), so deriving the VP some students get ø distinc-
tions. The past-tense auxiliary T‑should first-merges with this VP forming
the T‑bar should some students get ø distinctions. Let’s suppose that every
T constituent has an EPP feature requiring it to second-merge with a noun
or pronoun expression that serves as its specifier (and subject). In sentences
like (2a, 2b) in Belfast English, the EPP requirement for T to have a specifier
can be satisfied by second-merging T‑should with expletive there to form
the TP There should some students get ø distinctions. The resulting TP is
then merged with a null indicative complementiser, forming the CP in (4)
above. In the semantic component, this CP is interpreted as declarative in
type by default, in accordance with the Clause Typing Condition (5ii) below
(introduced in §1.4.3):

(5) Clause Typing Conditions

(i) A clause is interpreted as being of a given type (interrogative or
exclamative etc.) if it contains a peripheral clause-typing specifier of
the relevant type
(ii) An indicative clause is interpreted as declarative in type by default if
it contains no peripheral clause-typing specifier

Our discussion of subjects in Belfast English here invites questions

about the derivation of the corresponding Standard English sentence (1a)
Some students should get distinctions. Let’s suppose that the derivation
of (1a) starts out in the same way as that of (2a). So, once again, the verb
get first-merges with the QP ø distinctions to form the V‑bar get ø distinc-
tions, and then get second-merges with the QP some students to derive
the VP some students get ø distinctions. This VP is in turn first-merged
with the past-tense auxiliary T‑should, forming the T‑bar should some
students get ø distinctions. As before, let’s assume that T‑should has an
EPP feature requiring it to project a structural subject/specifier. But let’s
also suppose that the requirement for T‑should to have a specifier of its
own cannot be satisfied by merging expletive there in spec‑TP because in
standard varieties of English there can generally only occur in structures
containing an intransitive existential verb like be, become, exist, occur,
arise, remain etc. Instead, the EPP requirement for T to have a subject is
satisfied by moving the subject some students from its original position
72 A‑Movement

in spec‑VP, and merging it in a new position in spec‑TP – as shown by

the arrow in (6) below:

(6) CP

C TP
ø
QP T′
some students
T VP
should
QP V′
some students
V QP
get ø distinctions

Since spec‑TP is an A‑position, the kind of movement operation illustrated

by the arrow in (6) is termed A‑Movement. This is in contrast to the kind
of operation which moves a focused, topicalised or wh‑marked constituent
to spec‑CP, which (since spec‑CP is an A‑bar position) is known generically
as A‑bar Movement.
Given that other movement operations (e.g. Head Movement and A‑bar
Movement) are composite operations involving copying, we would ex-
pect the same to be true of A‑Movement. This would mean that when
the QP some students moves to its italicised position in spec‑TP in (6)
above, it leaves behind a copy of itself in spec‑VP. Only the higher copy
of the QP (= the one in spec‑TP) is spelled out overtly, in consequence
of the Default Spellout Rule, which was given the following formulation
in §1.4.1:

(7) Default Spellout Rule/DSR

For a constituent whose spellout is not determined in some other way,
the highest copy of the constituent is overtly pronounced, and any lower
copies are silent

One piece of evidence in support of a copying analysis of A‑Movement

comes from scope properties in relation to sentences such as (8a) below,
which will have the syntactic structure shown in simplified form in (8b) if
everyone originates as the specifier of the verb finished and is then raised
up (by A‑Movement) to become the specifier of the present tense auxiliary
has (as shown by the arrow in 8b):

(8) a. Everyone hasn’t finished it yet

b. [CP [C ø] [TP Everyone [T has] [NEGP [NEG n’t] [VP everyone [V finished] it yet]]]]
 2.1 VP Internal Subject Hypothesis 73

For some speakers, sentences like (8a) are ambiguous between (i) a reading
on which the quantifier expression everyone has scope over the negative
particle n’t so that the sentence means much the same as ‘Everyone is in
the position of not having finished it yet’, and (ii) another reading on which
everyone falls within the scope of the negative particle (so that the sentence
means much the same as ‘Not everyone is yet in the position of having fin-
ished it’). We can account for this scope ambiguity in a principled fashion
if we suppose that A‑Movement involves copying, that scope is defined in
terms of c‑command (so that a scope-bearing constituent has scope over a
scope-sensitive constituent which it c‑commands), and that the scope of a
universally quantified expression like everyone in negative structures like
(8b) can be determined either in relation to the initial position of everyone
or in relation to its final position. In (8b) everyone is initially merged in a
position (marked by strikethrough) in which it is c‑commanded by (and so
falls within the scope of) n’t; but via A‑Movement it ends up in an (ital-
icised) position in which it c‑commands (and so has scope over) n’t. The
scope ambiguity in (8a) therefore reflects the different positions occupied
by the initial and final copies of everyone in structure in (8b).

2.1.2 Further evidence for VP‑internal subjects

The claim made in §2.1.1 that (non-expletive) subjects like some students/
lots of students in sentences like (1) originate internally within the VP con-
taining the relevant verb (and from there move into spec‑TP in sentences
like 1 above) is known in the relevant literature as the VP-Internal Subject
Hypothesis (= VPISH), and this has been almost universally adopted in
research since the mid-1980s. Let’s take a look at some of the evidence in
support of VPISH.
One piece of evidence comes from the syntax of quantifiers in sentences
such as following:

(9) a. All the students have done the assignment

b. The students have all done the assignment

In both sentences, the universal quantifier all is construed as modifying the

DP the students. So how does all come to end up in two different positions
in (9a) and (9b)? A plausible answer is that the QP all the students origin-
ates in spec‑VP as the subject of done the assignment, and either the whole
QP can undergo A‑Movement to spec‑TP as shown in simplified form in
(10a) below, or alternatively the DP the students can undergo A‑Movement
on its own, leaving the quantifier all behind in the spec‑VP position that it
occupies in (10b):
74 A‑Movement

(10)a. CP

C TP
ø
QP T′
all the
students T VP
have
QP V′
all the
students V DP
done the assignment

b. CP

C TP
ø
DP T′
the stu-
dents T VP
have
QP V′

Q DP V DP
all the stu- done the assignment
dents

In a structure like (10b), all is said to be stranded, in the sense that it is

detached from (and occupies a lower position than) the higher italicised
DP the students that it modifies. The fact that the quantifier is stranded in
spec‑VP in (10b) provides us with evidence that the QP all the students ori-
ginates in spec‑VP.
An additional piece of evidence that subjects originate internally within
VP comes from the syntax of idioms. We can define idioms as expressions
(like those italicised below) which have an idiosyncratic meaning that is not
a purely compositional function of the meaning of their individual parts:

(11) a. Let’s have a couple of drinks to break the ice

b. Be careful not to upset the applecart
c. The president must bite the bullet

There seems to be a constraint that only a string of words which forms a

unitary constituent (i.e. a non-discontinuous constituent) can be an idiom.
So, while we find idioms like those in (11) which are of the form verb+
complement (but where the subject isn’t part of the idiom), we don’t find
 2.1 VP Internal Subject Hypothesis 75

idioms of the form subject+verb where the verb has a complement which isn’t
part of the idiom: this is because in subject+verb+complement structures, the
verb and its complement form a unitary constituent (a V‑bar), whereas the
subject and the verb do not – and only unitary constituents can be idioms.
In the light of the constraint that an idiom is a unitary constituent with
an idiosyncratic interpretation, consider idioms such as the following:

(12) a. All hell broke loose (= ‘There was pandemonium’)

b. The shit hit the fan (‘The situation became chaotic’)
c. The cat got his tongue (= ‘He was speechless’)

In (12), not only is the choice of verb and complement fixed, but so too is
the choice of subject. In such idioms, we can’t replace the subject, verb or
complement by near synonyms – as we see from the fact that sentences like
(13) below are ungrammatical (on the intended idiomatic interpretation):

(13) a. *The whole inferno escaped

b. *Camel dung was sucked into the air conditioning
c. *A furry feline bit his lingual articulator

However, what is puzzling about idioms like (12) is that one or more (itali-
cised) auxiliaries can freely be positioned between the subject and verb, as
we see from the examples below:

(14) a. All hell had broken loose

b. All hell could break loose
c. All hell would have broken loose
d. All hell might have been breaking loose

How can we reconcile our earlier claim that only a string of words which
form a unitary constituent can constitute an idiom with the fact that all
hell … break loose is a discontinuous string in (14), since the subject all hell
and the expression break loose are separated by the intervening auxiliaries
had/could/would have/might have been? To phrase the question another
way: how can we account for the fact that although the choice of subject,
verb and complement is fixed, the choice of auxiliary is not?
The VP-Internal Subject Hypothesis provides us with a principled answer,
if we suppose that subjects originate internally within VP, and that clausal
idioms like those in (12) are VP idioms which require a fixed choice of head,
complement and specifier in the VP containing them. For instance, in the
case of (12a), the relevant VP idiom requires the specific word break as its
head verb, the specific adjective loose as its complement, and the specific
quantifier phrase all hell as its subject/specifier. We can then account for
the fact that all hell surfaces in front of the auxiliary had in (14a) by posit-
ing that the QP all hell originates in spec‑VP as the subject of break loose,
76 A‑Movement

and is then raised (via A‑Movement) into spec‑TP to become the subject
of had broken loose. Given these assumptions, sentence (14a) All hell had
broken loose will be derived as follows.
The (perfect participle) verb broken merges with the AP/adjectival pro-
jection loose to form the idiomatic V‑bar broken loose. This is then merged
with its QP subject all hell to form the idiomatic VP all hell broken loose.
The resulting VP is merged with the (past tense) auxiliary T‑had to form the
T‑bar had all hell broken loose. Since finite auxiliaries carry an EPP feature
requiring them to have a nominal constituent like QP or DP as their subject/
specifier, the QP all hell moves from being the specifier of break to becom-
ing the specifier of had – as shown by the arrow below:

(15) TP

QP T′
all hell
T VP
had
QP V′
all hell
V AP
broken loose

We can then say that (in the relevant idiom) all hell must be the sister of
broken loose, and that this condition will be met only if all hell originates
in spec‑VP as the specifier (and sister) of the V‑bar broken loose. We can
account for how the subject all hell comes to be separated from its predicate
broken loose by positing that subjects originate internally within VP and
from there raise to spec‑TP (via A‑Movement) across an intervening T con-
stituent like had, so that the subject and predicate thereby come to be sep-
arated from each other – movement of the subject to spec‑TP being driven
by an EPP feature carried by the T auxiliary had requiring had to have a
nominal or pronominal subject. Subsequently, the TP in (15) is merged with
a null indicative complementiser, so deriving the structure associated with
(14a) All hell had broken loose: the resulting CP is interpreted as declarative
by default in accordance with the Clause Typing Condition (5ii).
To summarise: in Module 2.1 we have seen that data from expletive
structures in Belfast English, quantifier scope, floating quantifiers and
idioms provide evidence in support of the VP-Internal Subject Hypothesis/
VPISH, whereby subjects originate internally within the verb phrase/VP,
and from there move into spec‑TP via an operation known as A‑Movement.
In the next module, we will see that considerations relating to argument
structure provide further support for VPISH.
You should now be able to tackle Exercise 2.1.
 2.2 Argument structure and theta roles 77

2.2 Argument structure and theta roles

The assumption that subjects originate internally within VP ties up in

interesting ways with a traditional idea from predicate logic maintaining
that propositions (which can be thought of as representing the substantive
semantic content of clauses) comprise a predicate and a set of arguments.
In this module, we’ll explore key ideas embodied in this work.

2.2.1 Predicates and arguments

Simplifying somewhat, we can say that a predicate is an expression denot-
ing an activity or event, and an argument is an expression denoting a par-
ticipant in the relevant activity or event. For example, in sentences such
as those in (16) below, the italicised verbs are predicates and the bracketed
expressions represent their arguments:

(16) a. [The guests] have arrived

b. [The police] have arrested [the suspect]

In other words, the arguments of a verb are typically its subject and com-
plement. A verb like arrive when used with a single [bracketed] argument
in a structure like (16a) is said to be a one-place predicate; and a verb
like arrest when used with two [bracketed] arguments in a structure like
(16b) is said to be a two-place predicate. It has been widely assumed in
work spanning more than half a century that complements of verbs are
contained within a projection of the verb – for example, the suspect in
(16b) is the direct object complement of arrested and is contained within
the verb phrase headed by the verb arrested (serving as the complement
of the verb). Under the VP-Internal Subject Hypothesis, we can go fur-
ther than this and hypothesise that all the arguments of a predicate (not
just its complement, but also its subject) originate within a projection of
the predicate. Such an assumption allows us to maintain that there is a
uniform mapping/relationship between syntactic structure and semantic
argument structure – more specifically, between the position in which
arguments are initially merged in a syntactic structure and their semantic
function.
To see what this means in practice, consider the derivation of (16b)
The police have arrested the suspect. The verb arrested first-merges with
its direct object complement the suspect (a DP formed by merging the
determiner the with the NP suspect) to form the V‑bar arrested the sus-
pect; the verb arrested then second-merges with the subject DP the police
(formed by merging the determiner the with the NP police) to form the VP
shown in (17) below (simplified by not showing the internal structure of
the two DPs):
78 A‑Movement

(17) VP

DP V′
the police
V DP
arrested the suspect
In a structure such as (17), the complement the suspect is said to be the
internal argument of the verb arrested (in the sense that it is the argument
contained within the immediate V‑bar projection of the verb, and hence is
a sister of the verb), whereas the subject the police is the external argument
of the verb arrested (in that it occupies a position external to the V‑bar con-
stituent which is the immediate projection of the verb arrested, and is an
aunt of the verb). The VP in (17) is then merged with the present tense aux-
iliary T‑have, forming the T‑bar have the police arrested the suspect. T‑have
has an EPP feature requiring it to second-merge with a subject of its own,
and this EPP feature allows have to attract the DP the police to move from
being the subject of arrested to becoming the subject of T‑have, so forming
The police have the police arrested the suspect. Merging the resulting TP
with a null complementiser (marking the sentence as indicative in mood) in
turn derives the structure shown in simplified form in (18) below, with the
arrow showing the A‑Movement operation involved:

(18) CP

C TP
ø
DP T′
the police
T VP
have
DP V′
the police
V DP
arrested the suspect

Under the analysis in (18), the argument structure of the verb arrest is dir-
ectly reflected in the internal structure of the VP which it heads, since the
suspect is the internal (direct object) argument of arrested and the police
was initially merged as its external (subject) argument – and indeed a silent
copy of the police is left behind in spec‑VP, marking the spec‑VP position
as associated with the police. The overall CP is interpreted as declarative in
force in accordance with (5ii).
However, there is an important sense in which it is not enough simply
to say that in a sentence such as (16b) The police have arrested the sus-
pect the verb arrest is a predicate which has two arguments – the internal
 2.2 Argument structure and theta roles 79

argument the suspect and the external argument the police. After all, such a
description fails to account for these two arguments playing very different
semantic roles in relation to the act of arrest – that is, it fails to account
for the police being the individuals who perform the act (and hence get to
handcuff suspects and bundle them into the back of a police vehicle), and
the suspect being the person who suffers the consequences of the act (e.g.
being handcuffed and bundled into the back of a police vehicle). Hence,
any adequate account of argument structure should provide a description
of the semantic role which each argument plays. In the next section, we’ll
look at how this can be done.

2.2.2 Theta roles

In research spanning more than half a century, linguists have attempted to
devise a universal typology of the semantic roles played by arguments in
relation to their predicates. In (19) below are listed a number of terms used
to describe some of these roles (the convention being that terms denot-
ing semantic roles are c a p ita i s ed ), and for each role an informal gloss is
given, together with an illustrative example sentence (in which the itali-
cised expression has the semantic role specified):

(19) List of roles played by arguments with respect to their predicates

Role Gloss Example
the e Entity undergoing the effect of Mary fell over
some action
ag e nt Entity instigating some action Debbie killed Harry
e x pe r i e nc e r Entity experiencing some I like syntax
psychological state
o c at i ve Location in which something is He hid it there
situated or takes place
g oa Entity representing the John went home
destination of some other entity
s o u rc e Entity from which something He returned from
moves Paris
i n st r u ent Means by which something He hit it with a
comes about hammer

We can illustrate how the terminology in (19) can be used to describe the
semantic roles played by arguments in terms of the following examples:

(20) a. [The FBI] arrested [Larry Luckless]

[ag e nt ] [the e ]
b. [The suspect] received [a caution]
[g oa ] [the e ]
80 A‑Movement

c. [The audience] enjoyed [the play]

[experiencer ] [t h e e ]
d. [The president] went [to Boston]
[the e ] [g oa ]
e. [They] stayed [in a hotel]
[the e ] [ o c at i ve ]
f. [The noise] came [from the car]
[the e ] [s o u rc e ]

Given that – as we see from these examples – the the e role is a central one,
it has become customary to refer to the relevant semantic roles as thematic
roles; and since the Greek letter θ (= theta) corresponds to th in English and
the word thematic begins with th, it has also become standard practice to
use the synonymous expression theta role or θ‑role (with theta pronounced
[θi:tǝ] by some and [θei:tǝ] by others). Using this terminology, we can say
(for example) that in (20a) the FBI is the agent argument of the predicate
arrested, and that Larry Luckless is the the e argument of arrested.
Thematic relations (like ag e nt and t h e e ) have been argued to play a
central role in the description of a range of linguistic phenomena. For
example, it has been claimed that the distribution of certain types of adverb
is thematically determined. For instance, Gruber (1976) maintains that
adverbs like deliberately can only be associated with ag e nt arguments:

(21) a. John (= agent ) deliberately rolled the ball down the hill
b. *The ball (= the e ) deliberately rolled down the hill

Likewise, Fillmore (1970) argues that the adverb personally (in the use illus-
trated below) can only be associated with an (italicised) experiencer argument:

(22) a. Personally, I (= experience r ) don’t like roses

b. Personally, your proposal doesn’t interest me (= e x pe r i e nc e r )
c. *Personally, I (= agent ) hit you
d. *Personally, you hit me (= t h e e )

And in much the same vein, Anderson (1977: 267–71) argues that only
verbs with ag e nt subjects have by‑phrase nominal counterparts:

(23) a. The mayor (= agent ) protested

b. the protest by the mayor

(24) a. The mayor (= the e ) died

b. *the death by the mayor

Many other aspects of sentences have also been argued to be subject to the-
matic constraints. For example, selection restrictions (i.e. restrictions which
hold between a predicate and its choice of arguments) are a function of
 2.2 Argument structure and theta roles 81

the semantic properties of the predicate on the one hand, and the thematic
role of its arguments on the other. This can be illustrated by comparing the
following sentences:

(25) a. John rolled the ball down the hill

b. The ball rolled down the hill

Although the ball has a different constituent structure status in these two
sentences (= complement of the verb rolled in 25a, but the subject of rolled
in 25b), it has the same thematic function of t h e e (i.e. entity which under-
goes the rolling motion) in both cases. Accordingly, the ball is subject to
the same selection restrictions in the two uses – as illustrated below (where
! marks semantic/pragmatic anomaly):

(26) a. John rolled the ball/the rock/!the theory/!sincerity down the hill
b. The ball/the rock/!the theory/!sincerity rolled down the hill

Such data suggest that selection restrictions correlate with thematic struc-
ture rather than syntactic structure.
In a similar vein, Fillmore (1970) argues that constraints on coordinate
structures like (27c) below suggest that only constituents with the same
thematic function can readily be coordinated:

(27) a. John (= agent ) broke the window

b. A hammer (= instru ent ) broke the window
c. !John (= agent ) and a hammer (= instr u e n t ) broke the window

And Jackendoff (1972) argues that a number of constraints on passive

structures can be accounted for in thematic terms. For example, he argues
(1972: 44) that the ill-formedness of passive sentences like:

(28) a. *Five dollars are cost by this book

b. *Two hundred pounds are weighed by Bill

is attributable to violation of the following constraint (formulated in the-

matic terms):

(29) Passive Thematic Hierarchy Condition

The passive by‑phrase must be higher on the Thematic Hierarchy than
the superficial subject

The hierarchy referred to in (29) is the following:

(30) Thematic Hierarchy

ag e nt > ocative / source / goa > the e

Jackendoff maintains that the by‑phrase in both examples in (28) is a t h e e

argument of the relevant verb, whereas the superficial subject is a o c at i ve
82 A‑Movement

argument. Since t h e e is lower on the hierarchy (30) than o c at i ve , sen-

tences like (28) violate the condition (29) and so are ungrammatical.

2.2.3 Theta-marking
Having seen that theta roles play a key role in accounting for a variety
of aspects of the use of sentences, let’s reflect more generally on how
theta-marking (i.e. assignment of theta roles) works, and what constraints
it is subject to. In order to determine this, let’s take another look at the
sentences in (20) above, repeated as (31) below:

(31) a. [The FBI] arrested [Larry Luckless]

[agent ] [t h e e ]
b. [The suspect] received [a caution]
[goa ] [t h e e ]
c. [The audience] enjoyed [the play]
[experiencer ] [t h e e ]
d. [The president] went [to Boston]
[the e ] [g oa ]
e. [They] stayed [in a hotel]
[the e ] [ o c at i ve ]
f. [The noise] came [from the car]
[the e] [s o u rc e ]

If we look closely at the examples in (31), we see a fairly obvious pattern

emerging. Each of the bracketed argument expressions in (31) carries one
and only one θ‑role, and no two arguments of any predicate carry the same
θ‑role: for example, in (31a), the FBI is assigned the ag e nt role, and Larry
Luckless the t h e e role. Chomsky (1981) suggested that these thematic
properties of arguments are the consequence of a principle of Universal
Grammar outlined below:

(32) Theta Criterion/θ‑criterion

Each argument bears one and only one θ‑role, and each θ‑role is
assigned to one and only one argument (Chomsky 1981: 36)

A principle along the lines of (32) has been assumed (in some form or other)
in much subsequent work.
However, a question which arises from (32) is how θ‑roles are assigned
to arguments. In V‑bar constituents of the form verb+complement, the the-
matic role of the complement is determined by the semantic properties of
the verb. As examples like (31a–31c) above illustrate, the θ-role associated
with complements is often that of t h e e (though this is not always the
case – e.g. the complement me of the verb bother in Personally, it doesn’t
bother me has the thematic role of e x pe r i e nc er ).
 2.2 Argument structure and theta roles 83

But how are subjects assigned theta roles? Marantz (1984: 23ff.) and
Chomsky (1986a: 59–60) argue that although verbs directly assign theta
roles to their internal arguments (i.e. complements), it is not the verb
but rather the whole verb+complement (i.e. V‑bar) constituent which
determines the theta role assigned to its external argument. The evi-
dence they adduce in support of this conclusion comes from sentences
such as:

(33) a. John threw a ball

b. John threw a fit

(34) a. John broke the window

b. John broke his arm

Although the subject of the verb threw in both (33a) and (33b), John plays
a different thematic role in the two sentences – that of ag e nt in the case
of threw a ball, but that of e x pe r i e nc e r in threw a fit. Likewise, although
the subject of the verb broke in both (34a) and (34b), John plays the role
of agent in (34a) but that of e x pe r i e nc e r on the most natural (acciden-
tal arm-breaking) interpretation of (34b). From examples such as these,
Marantz and Chomsky conclude that the thematic role of the subject is not
determined by the verb alone, but rather is compositionally determined
by the whole verb+complement structure – that is, by V‑bar. On this view,
a verb assigns a θ-role directly to its internal argument, but only indir-
ectly (as a compositional function of the semantic properties of the overall
V‑bar) to its external argument. To use the relevant technical terminology,
we can say that predicates directly θ-mark their complements, but indir-
ectly θ-mark their subjects/specifiers.
A related observation is that auxiliaries seem to play no part in deter-
mining the assignment of theta roles to subjects, as can be illustrated in
relation to:

(35) a. John will throw a ball/a fit

b. John was throwing the ball/a fit
c. John had been throwing the ball/a fit
d. John might have been throwing the ball/a fit

Here, the thematic role of the subject John is determined purely by the
choice of V‑bar constituent (i.e. whether it is throw the ball or throw a fit),
and is not affected in any way by the presence of the italicised auxiliaries.
Clearly, any theory of θ-marking should offer us a principled answer to
questions such as the following: How are θ-roles assigned? Why do some
constituents (e.g. verbs) play a key role in θ-marking, while others (e.g.
auxiliaries) do not?
84 A‑Movement

We can provide a principled answer to these questions if we suppose that

theta roles are assigned to arguments in accordance with the hypothesis
below:

(36) Predicate-Internal Theta-Marking Hypothesis

An argument is theta-marked internally within a projection of a
predicate, either directly by being merged with the predicate, or indirectly
by being merged with an intermediate projection of the predicate

In the light of (36), consider our earlier sentence (16b) The police have
arrested the suspect. Here, the DP complement the suspect will be assigned
the θ-role of the e argument of arrested when the verb arrested merges
with this DP to form the V‑bar arrested the suspect. In a similar fashion, the
subject DP the police will be assigned the θ-role of agent when the V‑bar
arrested the suspect is merged with the subject DP to form the VP the police
arrested the suspect. The resulting VP is then merged with the auxiliary have
to form the T‑bar have the police arrested the suspect. Because T has an EPP
feature, T‑have attracts the subject DP the police to raise to spec‑TP, deriv-
ing The police have the police arrested the suspect. However, the subject the
police does not receive a theta role from the auxiliary have, since auxiliaries
are not predicates (unlike main verbs) and hence do not theta-mark their
subjects. The resulting TP is ultimately merged with a null complementiser
to form the CP shown in (18) above, which is the structure associated with
(16b) The police have arrested the suspect. Since both arguments of arrested
(the complement DP the suspect and the subject DP the police) are assigned
a single theta role, and each has a different theta role (the police being an
agent , and the suspect a the e ), the Theta Criterion (32) is satisfied.
To summarise: our discussion in Module 2.2 suggests that the semantic
role played by arguments with respect to their associated predicates can be
described in terms of a set of theta roles, and that thematic considerations
lend further support to the VP-Internal Subject Hypothesis. By positing
that subjects originate internally within VP, we can arrive at a unitary and
principled account of θ-marking in terms of the operation merge, in that
an argument is θ-marked when merged with a predicate or with an inter-
mediate projection of a predicate.
You should now be able to tackle Exercise 2.2.

2.3 Unaccusative structures

The overall conclusion to be drawn from our discussion hitherto is that

subjects originate internally within VP, as theta-marked arguments of the
verb. In all the structures we have looked at so far, the verb phrase has
 2.3 Unaccusative structures 85

contained both a complement and a specifier (the specifier being the sub-
ject of the verb). However, in this module we look at a special class of VPs
which contain a verb and a complement but no specifier, and where it is the
complement of the verb which subsequently moves to spec‑TP.

2.3.1 Unaccusative predicates

One class of verb phrases which have a complement but no specifier are those
headed by a special subclass of intransitive verbs which have the property
that they can be used both in expletive structures such as (37a, 38a, 39a)
below, and in non-expletive structures like those in (37b, 38b, 39b):

(37) a. There have arisen several complications

b. Several complications have arisen

(38) a. There could have occurred a diplomatic incident

b. A diplomatic incident could have occurred

(39) a. There does still remain some hope of finding survivors

b. Some hope of finding survivors does still remain

I shall argue here that the italicised arguments originate as the comple-
ments of the bold-printed verbs in both types of structure: they remain in
situ as the complement of the verb in the expletive (a) structures, but raise
to become the subject of the underlined T auxiliary have/could/does in the
non-expletive (b) structures. However, the bold-printed verbs in (37–39)
differ from transitive verbs taking a nominal expression as their comple-
ment in that they don’t assign accusative case to their complement, but
rather nominative case. Although the nominative-accusative case contrast
is invisible on noun expressions in English, it is visible in languages like
Icelandic with a richer case system. And in Icelandic, the (italicised) com-
plement of an unaccusative verb receives nominative (= n o ) case even in
expletive structures where it remains in situ and follows the verb – as the
following example illustrates:

(40) Þad hafa komið nokkrirNOM gestirNOM

There have come some guests

And it may be that the nominative case borne by unaccusative arguments

can be illustrated from English by use of I in archaic structures such as:

(41) There but for the grace of God go I

Because they don’t assign accusative case to their complements, verbs like
those bold-printed in (37–39) are known as unaccusative predicates.
Given that the complements of unaccusative predicates are not assigned
accusative case, it would be only natural to question whether they actually
86 A‑Movement

are complements (rather than e.g. subjects which end up positioned after
the unaccusative verb rather than in front of it). In fact, there is strong
syntactic evidence in support of analysing them as complements. Part of
the evidence comes from their behaviour in relation to a constraint on
movement operations (dating back to work by Cattell 1976, Cinque 1978,
and Huang 1982) which was given the following characterisation in §1.5.1:

(42) Constraint on Extraction Domains/CED

Only complements allow material to be extracted out of them, not
specifiers or adjuncts

We can illustrate how CED works in terms of the following contrasts:

(43) a. He was taking [pictures of who]?

b. Who was he taking [pictures of who]?

(44) a. [Part of what] has broken?

b. *What has [part of what] broken?

(45) a. He was angry [because she did what]?

b. *What was he angry [because she did what]?

(43a, 44a, 45a) are questions in which the wh‑pronoun who/what remains
in situ, while (43b, 44b, 45b) are their Wh‑Movement counterparts. In (43b),
who is extracted out of a bracketed nominal expression which is the com-
plement of the verb taking, and yields a grammatical outcome because
there is no violation of CED (extraction out of complements being per-
mitted by CED). By contrast, in (44b) what is extracted out of a bracketed
expression which is the subject (and hence specifier) of the auxiliary has,
and since CED blocks extraction out of specifiers, the resulting sentence
is ungrammatical. Likewise, in (45b) what is extracted out of a bracketed
adjunct clause, and since CED blocks extraction out of adjuncts, the sen-
tence is ungrammatical.
In the light of CED, consider a sentence such as:

(46) How many survivors does there remain [some hope of finding — ]?

Here, the wh‑phrase how many survivors has been extracted (via
Wh‑Movement) out of the bracketed expression some hope of finding how
many survivors, moving from the gap position marked by — to the italicised
position. Given that the Condition on Extraction Domains tells us that only
complements allow material to be extracted out of them, it follows that
the bracketed expression in (46) must be the complement of the unaccusa-
tive verb remain. By extension, we can assume that the italicised expres-
sions in (37–39) likewise originate as the complements of the bold-printed
unaccusative verbs.
 2.3 Unaccusative structures 87

A further argument supporting the claim that unaccusative subjects

are initially merged as complements comes from observations about
quantifier stranding in the West Ulster variety of English. McCloskey
(2000) notes that West Ulster English allows wh‑questions such as those
below which have the interpretation ‘What are all the things that you
got for Christmas?’:

(47) a. What all did you get for Christmas?

b. What did you get all for Christmas?

He argues that when the universal quantifier all is used to modify a wh‑word
like what, Wh‑Movement can either move the whole expression what all to
the front of the sentence (as in 47a) above, or can move the word what on
its own, thereby stranding the quantifier in situ (as in 47b). In the light of
his observation, consider the following sentence:

(48) What happened all at the party last night?

The fact that the quantifier all is stranded in a position following the
unaccusative verb happened suggests that the wh‑expression what all
originates in postverbal position as the complement of the verb hap-
pened. More generally, sentences like (48) provide empirical evidence
in support of positing that unaccusative subjects are initially merged as
complements.

2.3.2 Unaccusatives compared with other predicates

Although unaccusative verbs allow their arguments to be positioned after
them in expletive structures like (37a, 38a, 39a) above, there are other types
of predicate which do not – as we see from the ungrammaticality of sen-
tences such as (49) below:

(49) a. *When the Snail Rail train arrived five hours late, there complained
many passengers
b. *In the dentist’s surgery, there groaned a toothless patient
c. *Every time General Wynott Nukem goes past, there salutes a guard
at the gate

Intransitive verbs like complain/groan/salute are known as unergative

verbs: they differ from unaccusative verbs in that the subject of an
unergative verb has the thematic role of an ag e n t argument, whereas
the subject of an unaccusative verb has the thematic role of a t h e e
argument.
In addition to those already noted, there are a number of other import-
ant syntactic differences between unaccusative verbs and other types of
verb (e.g. unergative verbs or transitive verbs). For example, Henry (1995)
88 A‑Movement

observes that in one dialect of Belfast English (which she calls dialect A)
unaccusative verbs like those bold-printed below can have (italicised) post-
verbal subjects in imperative structures like:

(50) a. Leave you now!

b. Arrive you before 6 o’clock!
c. Be going you out of the door when he arrives!

By contrast, other (e.g. unergative or transitive) verbs don’t allow post-

verbal imperative subjects, so that imperatives such as those below are
ungrammatical in the relevant dialect:

(51) a. *Read you that book!

b. *Eat you up!
c. *Always laugh you at his jokes!

Additional evidence for positing that unaccusative verbs are syntactically

distinct from other verbs comes from auxiliary selection facts in relation
to earlier stages of English when there were two perfect aspect auxiliaries
(have and b e ), each taking a complement headed by a specific kind of verb.
Unaccusative verbs differed from transitive or unergative verbs in being
used with the perfect auxiliary b e , as the sentences below (taken from vari-
ous plays by Shakespeare) illustrate:

(52) a. Mistress Page is come with me (Mrs Ford, Merry Wives of Windsor,
V.v)
b. How chance thou art returned so soon? (Antipholus, Comedy of
Errors, [Link])
c. I am arriv’d for fruitful Lombardy (Lucentio, Taming of the Shrew,
I.i)
d. You shall hear I am run away (Countess, All’s Well That Ends Well,
[Link])
e. She is fallen into a pit of ink (Leonato, Much Ado About Nothing,
IV.i)
f. Did he not say my brother was fled? (Don Pedro, Much Ado About
Nothing, V.i)
g. And now is he become a man (Margaret, Much Ado About Nothing,
[Link])
h. Why are you grown so rude? (Hermia, Midsummer Night’s Dream,
[Link])

We find a similar contrast with the counterparts of perfect h av e / b e in a

number of other languages – for example, Italian and French (Burzio 1986),
Sardinian (Jones 1994), German and Dutch (Haegeman 1994), and Danish
(Spencer 1991).
 2.3 Unaccusative structures 89

A further difference between unaccusative predicates and others relates

to the adjectival use of their perfect participle forms. As the examples below
indicate, (italicised) perfect participle forms of unaccusative verbs can be
used adjectivally (to modify a noun), for example, in sentences such as:
(53) a. The train arrived at platform 4 is the delayed 8.28 for London Euston
b. The vice squad arrested a businessman recently returned from Thailand
c. Several facts recently come to light point to his guilt
d. Brigadier Bungle is something of a fallen hero

By contrast, perfect participle forms of (active) transitive verbs (like those

in 54a, 54b below) or unergative verbs (like those in 54c, 54d) cannot be
used in the same way, as we see from the ungrammaticality of examples
like the following:

(54) a. *The man committed suicide was a neighbour of mine

b. *The thief stolen the jewels was never captured
c. *The man overdosed was Joe Doe
d. *The yawned student eventually fell asleep in class

In this respect, unaccusative verbs resemble passive participles, which can

also be used adjectivally (cf. a changed man, a battered wife, a woman
arrested for shoplifting, etc.).
Further support for the subjects of unaccusative predicates originating as
complements comes from thematic considerations. In this respect, consider
alternations such as the following:

(55) a. A torpedo sank the ship

b. The ship sank

In (55a) the subject a torpedo is an i n st r u e n t argument of sank (rep-

resenting the means used to destroy the ship), and the complement the
ship is a the e argument (representing the entity undergoing the effects of
sinking). In (55b), the ship is again a t h e e argument, but this time it is the
subject of sank, rather than its complement. Baker (1988) argued at length
that there is a uniform mapping between thematic and syntactic structure,
and he captured this in terms of a principle of Universal Grammar/UG
which can be outlined informally as follows:

(56) Uniform Theta Assignment Hypothesis/UTAH

Constituents which fulfil the same thematic role with respect to a given
predicate occupy the same initial position in the syntax.

So, since the ship is a t h e e argument of sank in both sentences in (55),

and since it is the complement of sank in (55a), it is plausible to suppose
(on the basis of UTAH) that it originates as the complement of sank in (55b)
90 A‑Movement

also, and from there moves into spec‑TP via A‑Movement, in the manner
shown by the arrow below:

(57) CP

C TP
ø
DP T′
the ship
T VP
Af
V DP
sink the ship

In the PF component, the higher (italicised) copy of the DP the ship will receive
an overt spellout and the lower copy a silent spellout (marked by strikethrough),
in accordance with the following spellout rule (repeated from 7 above):

(58) Default Spellout Rule/DSR

For a constituent whose spellout is not determined in some other way,
the highest copy of the constituent is pronounced at PF, and any lower
copies are silent

In addition, the past tense affix in T will lower onto the verb sink in (57) via
Affix Hopping, so that this is spelled out at PF as the past tense form sank.
Verbs which (when used intransitively) are unaccusative predicates in-
clude motion predicates such as arrive/come/fall/go/leave/return/rise etc. and
existential predicates (i.e. verbs indicating the existence or coming into be-
ing of a state of affairs, or a change of state, or the cessation of a state) like
appear/arise/be/become/begin/change/die/exist/happen/occur/remain/start/
stay/stop/turn. An unaccusative verb like fall typically allows inanimate as
well as animate subjects (cf. My wife fell off a stool/The temperature fell
sharply), whereas an unergative predicate like complain typically allows only
an animate subject (cf. My wife complained/!The temperature complained).

2.3.3 The syntax of unaccusative subjects

As we have seen in §2.3.2, there is empirical evidence that unaccusative
subjects behave differently from subjects of other (e.g. unergative or tran-
sitive) verbs. Why should this be? The answer given by the VP Internal
Subject Hypothesis is that the subjects of unaccusative verbs do not orig-
inate as the subjects of their associated verbs at all, but rather as their
complements, and that unaccusative structures with postverbal arguments
involve leaving the relevant argument in situ in VP-complement position –
for example, in unaccusative expletive structures such as (37–39) above,
 2.3 Unaccusative structures 91

and in Belfast English unaccusative imperatives such as (50). This being so,
a sentence such as (37a) There have arisen several complications will have
the following derivation (slightly simplified). The QP several complications
is merged as the complement of the unaccusative verb arisen, forming
the VP arisen several complications. This VP is merged with the auxiliary
T‑have, thereby forming the T‑bar below:

(59) T′

T VP
have
V QP
arisen several complications

The finite T auxiliary have carries an EPP feature (not shown in 59) which
requires it to have a pro/nominal constituent as its specifier. One way of
satisfying this requirement is by second-merging T‑have with the expletive
pronominal DP there in spec‑TP, and the resulting TP there have arisen
several complications is then merged with a null complementiser (marking
indicative mood) to form the CP shown in (60):

(60) CP

C TP
ø
DP T′
there
T VP
have
V QP
arisen several complications
The CP in (60) is spelled out at PF as (37a) There have arisen several com-
plications, and is interpreted as declarative in force in conformity with the
Default Clause Typing Condition (5ii).
However, an alternative way for the T constituent in (59) to satisfy the
EPP requirement for it to have a pro/nominal specifier is not to use an
expletive, but instead allow T to attract a pro/nominal to move to spec‑TP.
Let’s suppose that (in such a case) T attracts the closest pro/nominal which
it c‑commands, in consequence of a principle (termed the Relativised
Minimality Condition in Rizzi 1990, 2003) which was given the following
characterisation in §1.3.2:

(61) Minimality Condition/MC

A constituent can only enter into a syntactic relation with the minimal
(= closest) constituent of the relevant type above/below it
92 A‑Movement

(where above/below mean ‘c‑commanding/c‑commanded by’). Since the

closest nominal below T in (59) is the QP several complications (not the NP
complications), T attracts this QP to move to spec‑TP in the manner shown
in (62):

(62) CP

C TP
ø
QP T′
several complications
T VP
have
V QP
arisen several complications

The type of movement involved is the familiar A‑Movement operation

which moves an argument from an A‑position lower down in a sentence
to an A‑position higher up in the sentence: in the case of (62), the arrowed
A‑Movement operation moves the QP several complications from comp‑VP
(i.e. complement position within VP) into spec‑TP (i.e. specifier position
within TP). In the PF component, the higher (italicised) copy of the moved
QP receives an overt spellout, and the lower copy (marked by strikethrough)
a null spellout, in accordance with the Default Spellout Rule (58) above.
Consequently, the structure in (62) is spelled out at PF as (37b) Several com-
plications have arisen.
The derivation in (62) illustrates a characteristic property of A‑Movement.
Since comp‑VP is a thematic A‑position/θ‑position (i.e. an A‑position
which is assigned a theta role by the relevant predicate/verb) but spec‑TP is
a θ‑bar position (i.e. an A‑position to which no theta role is assigned), we
can see that an A‑Movement structure like (61) involves movement from a
thematic A‑position within VP into a non-thematic A‑position in spec‑TP.
This results in the formation of an A‑chain (i.e. a movement chain created
by A‑Movement) which satisfies the Theta Criterion (32) by virtue of the
relevant A‑chain being assigned only one theta role: more specifically, the
null copy of the QP several complications at the foot of the A‑chain in (62)
is assigned the θ‑role of t h e e complement of arisen, but the overt copy
of the same QP at the head of the movement chain is in a θ‑bar position
in spec‑TP where it receives no additional θ‑role (but instead inherits the
θ‑role of its lower copy). This suggests that the Theta Criterion (32) can be
taken to be a condition on A‑chains, requiring each A‑chain to carry one
and only one theta role.
 2.3 Unaccusative structures 93

The A‑Movement analysis of unaccusative subjects in (62) above allows

us to provide an interesting account of sentence pairs like that below:

(63) a. All hope of finding survivors has gone

b. All hope has gone of finding survivors

Since go is an unaccusative verb, the QP all hope of finding survivors will

originate as the complement of gone. Merging gone with this QP will derive
the VP gone all hope of finding survivors. The resulting VP is merged with
the T auxiliary has to form the T‑bar has gone all hope of finding survivors.
The QP all hope of finding survivors is attracted by the EPP feature on T‑has
to raise to the bold-printed spec‑TP position, leaving a copy behind in the
italicised position in which it originated – as shown in (64):

(64) CP

C TP
ø
QP2 T′

Q NP T VP
all has
N KP V QP1
hope of finding gone
survivors Q NP
all
N KP
hope of finding
survivors

In the case of (63a), the higher copy of the moved QP all hope of finding
survivors is spelled out in the bold-printed spec‑TP position which it moves
to, and the lower copy (QP1) in the italicised VP-complement position is
given a null spellout (in accordance with the Default Spellout Rule 58) – as
shown in simplified form in (65):

(65) CP

TP

T′

VP

C QP2 T V QP2
ø all hope has gone all hope
of finding survivors of finding survivors
94 A‑Movement

By contrast, in the case of (63b), the quantifier all and the noun hope are
spelled out in the bold-printed position that they move to in (64), but the
KP of finding survivors is spelled out in the VP-complement position in
which it originates – as shown in (66):

(66) CP

TP

T′

VP

C QP2 T V QP2
ø all hope has gone all hope
of finding survivors of finding survivors
The exceptional spellout pattern in (66) is the result of the application of a
special Low Spellout Rule which is given the following informal character-
isation in Radford (2020: 277):

(67) Low Spellout Rule/LSR

A KP/PP/CP which is the lowest/rightmost constituent of a larger phrase
that undergoes movement can be spelled out on a lower copy of the
moved phrase

(67) allows the KP/case phrase of finding survivors to be spelled out on the lower
(italicised) copy of the moved QP, while the remainder of the QP is spelled out
on the higher (bold-printed) copy, giving rise to discontinuous/split spellout.
This split spellout pattern provides evidence in support of taking A‑Movement
(like other movement operations) to be a composite operation involving copy-
ing and deletion. An interesting observation about sentences like (63) is that
they allow the archaic use of be as a perfect auxiliary (as in All hope is now
gone of finding survivors), so underlining their unaccusative status.
To summarise: in Module 2.3 we have looked at the syntax of unaccusa-
tive predicates like arise/remain/occur etc. We saw that the thematic argu-
ment of an unaccusative verb originates as its complement but differs from
the complement of a transitive verb in that (in a finite clause) it receives
nominative rather than accusative case. We examined a number of further
differences between unaccusative predicates and other types of predicate
(e.g. in relation to the position of subjects in Belfast English imperatives,
and auxiliary selection in earlier varieties of English). We also looked at
how split spellout patterns can arise in unaccusative clauses like All hope
has now gone of finding survivors.
You should now be able to tackle Exercise 2.3.
 2.4 Passive structures 95

2.4 Passive structures

In this module, we turn to examine the syntax of passive structures. We will

see that passives are similar in certain respects to unaccusative structures.
We begin by looking at simple passive structures (i.e. those which involve
Passivisation applying internally within a single clause).

2.4.1 Simple passives

Traditional grammarians maintain that the bold-printed verbs in sentences
such as (68a, 69a, 70a) below are in the active voice, whereas the italicised
verbs in the corresponding sentences in (68b, 69b, 70b) are in the passive
voice (so that the italicised verbs have the status of passive participles):

(68) a. Hundreds of passers-by saw the attack

b. The attack was seen by hundreds of passers-by

(69) a. Lex Luthor stole the kryptonite

b. The kryptonite was stolen by Lex Luthor

(70) a. They took everything

b. Everything was taken

There are four main properties which differentiate passive sentences

from their active counterparts. One is that passive (though not active)
sentences generally require the auxiliary b e (as in 70b ‘Everything was
taken’). Another is that the main verb in passive sentences is in the pas-
sive participle form (cf. seen/stolen/taken), which is often the same as the
perfect participle form (cf. He has stolen it). A third is that passive sen-
tences may contain a by‑phrase in which the complement of by plays the
same thematic role as the subject in the corresponding active sentence: for
example, hundreds of passers-by in the active structure (68a) serves as the
subject of saw the attack, whereas in the passive structure (68b) it serves
as the complement of the preposition by (though in both cases it has the
thematic role of e x p e r i e n c e r argument of see). The fourth difference is
that the expression which serves as the complement of an active verb sur-
faces as the subject in the corresponding passive structure: for example,
the attack is the complement of saw in the active structure (68a), but is
the subject of was in the passive structure (68b). Since this chapter is con-
cerned with A‑Movement (and hence the syntax of subjects), we focus on
the syntax of the superficial subjects of passive sentences, setting aside the
derivation of by‑phrases.
Passive predicates resemble unaccusatives in that alongside structures
like those in (71a–73a) below containing (bold-printed) preverbal subjects,
they also allow expletive structures like (71b–73b) in which the relevant
96 A‑Movement

argument can occupy the italicised postverbal position (providing it is an

indefinite expression):
(71) a. No evidence of any corruption was found
b. There was found no evidence of any corruption

(72) a. Several cases of syntactophobia have been reported

b. There have been reported several cases of syntactophobia

(73) a. A significant change of policy has been announced

b. There has been announced a significant change of policy

How can we account for the dual position of the highlighted constituents
in such structures?
The answer given within the framework utilised here is that a passive
subject is initially merged as the thematic complement of the main verb (i.e.
it originates as the complement of the main verb as in 71b, 72b, 73b) and
so receives the θ‑role which the relevant verb assigns to its complement.
It then moves from comp‑VP (i.e. complement position within VP) into
spec‑TP in passive sentences such as (71a, 72a, 73a), via a specific type of
A‑Movement operation traditionally termed Passivisation.
On this view, the derivation of sentences like (71) will proceed as fol-
lows. The quantifier any merges with the NP corruption to form the QP any
corruption. The resulting QP is then merged with the case particle/K of to
form the KP/case phrase of any corruption. This KP in turn is merged with
the noun evidence to form the NP evidence of any corruption. The resulting
NP is merged with the negative quantifier no to form the QP no evidence of
any corruption. This QP is merged as the complement of the passive verb
found (and thereby assigned the thematic role of the e argument of found)
to form the VP found no evidence of any corruption. The VP thereby formed
is merged with the T auxiliary was forming the T‑bar was found no evidence
of any corruption. The auxiliary T‑was carries an EPP feature requiring it to
have a specifier. This requirement can be satisfied by merging the expletive
pronominal DP there in spec‑TP, deriving the TP There was found no evidence
of any corruption. Merging this TP with a null complementiser marking indi-
cative mood will derive the structure shown in simplified form in (74) below:

(74) CP

C TP
ø
DP T′
there
T VP
was
V QP
found no evidence of any corruption
 2.4 Passive structures 97

However, an alternative way of satisfying the EPP feature of T‑was is not

to merge there in spec‑TP, but rather to passivise the QP no evidence of any
corruption – that is, move it from being the thematic complement of found to
becoming the specifier/subject of T‑was. Merging the resulting TP with a null
complementiser derives the CP shown in simplified form in (75) below (with the
dotted arrow showing the A‑Movement operation involved in Passivisation):

(75) CP

C TP
ø
QP T′
no evidence
of any corruption T VP
was
V QP
found no evidence
of any corruption

The arrowed Passivisation operation by which QP moves from being the

complement of the verb found to becoming the specifier of T‑was is a par-
ticular instance of the more general A‑Movement operation which serves
to create structural subjects (i.e. to move some constituent from a position
below T into spec‑TP in order to satisfy the EPP feature of T).
Note that an assumption implicit in the analyses in (74) and (75) is that
verb phrases headed by intransitive passive participles remain subjectless
throughout the derivation, because the T auxiliary was is the head which
requires a structural subject by virtue of its EPP feature, not the verb found
(suggesting that it is functional heads like T and C which trigger movement,
not lexical heads like V). Note also that the A‑Movement operation in (75)
satisfies the Theta Criterion (32) because the QP no evidence of any corrup-
tion is assigned a θ‑role (as the e argument of the verb found) in its original
position in comp‑VP, but is not assigned a second theta role when it moves
to spec‑TP, because spec‑TP is a θ‑bar position. The resulting A‑chain thus
carries one and only one θ‑role, and hence satisfies the Theta Criterion.
In the case of (71a) No evidence of any corruption was found, the whole of
the QP no evidence of any corruption is spelled out in the italicised spec‑TP
position in (75) at the head (i.e. on the highest link) of the movement chain,
and all the material in the comp‑VP position at the foot (i.e. on the lowest
link) of the movement chain is silent, in accordance with the Default Spell-
out Rule/DSR (58) above. However, an alternative possibility is for the Low
Spellout Rule/LSR (67) to apply. LSR would allow the KP of any corruption
to be spelled out on the lower link of the chain (as shown in italics in 76),
98 A‑Movement

and the remainder of the moved QP would then be spelled out on the higher
link of the chain by DSR (as shown in bold print), so resulting in the fol-
lowing split spellout pattern:

(76) CP

TP

T′

VP

C QP T V QP
ø no evidence was found no evidence
of any corruption of any corruption

And (76) is the structure associated with the sentence No evidence was
found of any corruption. Note that the postverbal position of the KP of any
corruption in (76) provides us with evidence that the subject of a passive
predicate originates in comp‑VP.
Split spellout passive structures like (76) provide evidence that passive
subjects originate as complements. Further evidence comes from the distri-
bution of idiomatic nominals like those italicised in (77)–(79):

(77) a. They paid little heed to what he said

b. Little heed was paid to what he said

(78) a. They paid due homage to General Ghouly

b. Due homage was paid to General Ghouly

(79) a. The FBI kept close tabs on the CIA

b. Close tabs were kept on the CIA by the FBI

In expressions such as pay heed/homage to and keep tabs on, the verb
pay/keep and the noun expression containing heed/homage/tabs together
form an idiom. Given the arguments in §2.1.2 that idioms are unitary con-
stituents, it is apparent that the bold-printed verb and the italicised noun
expression must form a unitary constituent when they are first introduced
into the derivation. This will clearly be the case if we suppose that the noun
expression originates as the complement of the associated verb (as in 77a,
78a, 79a), and becomes the subject of the passive auxiliary was/were in
(77b, 78b, 79b) via Passivisation/A‑Movement.
Additional evidence that passive subjects are initially merged as com-
plements comes from quantifier stranding in West Ulster English structures
such as the following (from McCloskey 2000: 72):

(80) What was said all at the meeting?

 2.4 Passive structures 99

Recall from our earlier discussion of sentences like (47, 48) that McCloskey
argues that stranded quantifiers modifying wh‑expressions are left behind
via movement of the wh‑expression without the quantifier. This being so,
sentences such as (80) provide evidence that what all originates as the com-
plement of the passive participle said (with what subsequently being pas-
sivised on its own, stranding all) – and more generally, that passive subjects
are initially merged as thematic complements.
A claim which is implicit in the hypothesis that passive subjects originate
as thematic complements is that the subjects of passive verbs and the com-
plements of active verbs have the same thematic function. Evidence that
this is indeed the case comes from the observation that the two are subject
to the same selection restrictions on the choice of expression which can oc-
cupy the relevant position, as we see from sentences such as the following
(where ?/?!/! mark increasing degrees of anomaly):

(81) a. The students/?the camels/?!The flowers/!The ideas were arrested

b. They arrested the students/?the camels/?!the flowers/!the ideas

We can account for this if we suppose that selection restrictions on the

choice of admissible arguments for a given predicate depend jointly on the
semantic properties of the predicate and the thematic role of the argument:
it will then follow that two expressions which fulfil the same thematic role
in respect of a given predicate will be subject to the same restrictions on
argument choice. Since passive subjects like those italicised in (81a) origin-
ate as complements, they will have the same θ‑role (and so be subject to
the same pragmatic restrictions on argument choice) as active complements
like those italicised in (81b).
We can arrive at the same conclusion (that passive subjects originate as
thematic complements) on theoretical grounds. Earlier, we hypothesised
that thematic structure is mapped into syntactic structure in a uniform
fashion, in consequence of the Uniform Theta Assignment Hypothesis/
UTAH, outlined in (56) above. Consequently, it follows from UTAH that if
passive subjects have the same theta role as active complements, they must
originate in the same comp‑VP position as active complements.

2.4.2 Cross-clausal passives

Thus far, the instances of Passivisation which we have looked at have
been clause-internal in the sense that they have involved movement from
comp‑VP to spec‑TP within the same clause. However, Passivisation is also
able to apply across certain types of clause boundary – for example, across
the infinitive clause boundary bracketed in (82):

(82) a. There are alleged [to have been stolen several portraits of the queen]
b. Several portraits of the queen are alleged [to have been stolen]
100 A‑Movement

It seems clear that the italicised QP several portraits of the queen is the
thematic complement of the bold-printed verb stolen in the bracketed
infinitive clause. In (82a) this QP remains in situ as the complement of the
bold-printed verb; but in (82b) it moves to become the structural subject of
the auxiliary are. Let’s look rather more closely at the derivation of these
two sentences.
(82a) has the following derivation (slightly simplified). The noun por-
traits merges with the KP/case phrase of the queen to form the NP portraits
of the queen. The quantifier several merges with this NP to form the QP
several portraits of the queen. The verb stolen merges with (and assigns the
thematic role of t h e e argument to) its QP complement several portraits
of the queen, thereby forming the VP stolen several portraits of the queen.
The resulting VP is then merged with the passive VOICE auxiliary been to
form the VOICEP been stolen several portraits of the queen. This VOICEP
is in turn merged with the perfect aspect auxiliary/PERF have to form the
PERFP have been stolen several portraits of the queen. The PERFP thereby
formed is merged with the infinitival tense particle T‑to, forming the T‑bar
to have been stolen several portraits of the queen. Infinitival to (like all T
constituents) has an EPP feature which requires it to have a noun or pro-
noun expression as its subject/specifier. One way of satisfying this require-
ment is for expletive there to be merged in spec‑TP, forming the TP shown
in (83) below:

(83) TP

DP T′
there
T PERFP
to
PERF VOICEP
have
VOICE VP
been
V QP
stolen
Q NP
several
N KP
portraits of the queen
The resulting TP is merged as the complement of the (passive participle)
verb alleged to form the VP alleged there to have been stolen several por-
traits of the queen. This VP is then merged with the present-tense auxil-
iary T‑are to form the T‑bar are alleged there to have been stolen several
 2.4 Passive structures 101

portraits of the queen. T‑are has an EPP feature which can attract expletive
there to become its subject (via Passivisation/A‑Movement), so forming the
TP there are alleged there to have been stolen several portraits of the queen.
This TP is then merged with a null C (marking indicative mood), so deriving
the following structure (simplified by not showing the internal structure of
VOICEP, which is as shown in 83 above):

(84) CP
C TP
ø
DP T′
there
T VP
are
V TP
alleged
DP T′
there
T PERFP
to
PERF VOICEP
have been stolen several
portraits of the queen

The overall CP is interpreted as declarative in type by the Default Clause

Typing Condition (5ii), and is spelled out at PF as (82a) There are alleged to
have been stolen several portraits of the queen.
Now consider the derivation of (82b) Several portraits of the queen are
alleged to have been stolen. Assume that the derivation proceeds as for
(82a) until we reach the stage where we have formed the T‑bar to have
been stolen several portraits of the queen. But this time, suppose that the
requirements of the EPP feature on T‑to are satisfied by moving the QP
several portraits of the queen to spec‑TP to become the subject of T‑to, so
forming the TP several portraits of the queen to have been stolen. The re-
sulting TP is merged with the passive verb alleged to form the VP alleged
several portraits of the queen to have been stolen. This VP is then merged
with the present-tense T auxiliary are to form the T‑bar are alleged several
portraits of the queen to have been stolen. Since the T auxiliary are has an
EPP feature, it can attract the closest noun or pronoun expression (= the
QP headed by several) to become its subject/specifier, so forming the TP
several portraits of the queen are alleged to have been stolen. The resulting
TP is then merged with a null C (marking indicative mood), forming the CP
below (simplified by abbreviating lower copies of the QP several portraits
of the queen to spotq, in order to save space):
102 A‑Movement

(85) CP
C TP2
ø
QP T′
several
portraits T VP
of the are
queen V TP1
alleged
QP T′
(ii) spotq
T PERFP
to
PERF VOICEP
have
V VP
been
V QP
stolen spotq
(i)

Our discussion illustrates that (in consequence of the EPP feature on each T),
cross-clausal Passivisation takes place in a successive-cyclic fashion (i.e. in a
succession of short steps), moving the passivised constituent one TP at a time.
More specifically, cross-clausal Passivisation involves multiple applications
of the familiar A‑Movement operation: consequently, each of the two num-
bered movements arrowed in (85) is a particular instance of A‑Movement.
In the previous section (§2.4.1), Passivisation was treated as an operation
by which the complement of a VP is raised to become the subject/speci-
fier of a TP above it: this holds true of the lower Passivisation operation
numbered (i) in (85) above, where the passivised QP several portraits of the
queen moves from being the complement of the verb stolen to becoming
the specifier of TP1. However, it is not true of the higher Passivisation oper-
ation numbered (ii), since this involves the QP several portraits of the queen
moving from spec‑TP1 to spec‑TP2. This shows that Passivisation is an in-
stance of a more general A‑Movement operation which moves a constitu-
ent from a lower (complement or specifier) A‑position into a higher one.
And indeed independent evidence that subjects/specifiers can be passiv-
ised comes from sentences such as:

(86) a. All hell was said to have broken loose

b. The shit is expected to hit the fan

In (86a, 86b) the italicised passivised nominal is an idiomatic expression which

originates as the subject of the bold-printed expression – demonstrating that
Passivisation can target subjects as well as complements. The derivation of
(86b) proceeds as follows. The verb hit merges with the DP the fan to form
 2.4 Passive structures 103

the V‑bar hit the fan. This V‑bar in turn merges with the DP the shit to form
the (idiomatic) VP the shit hit the fan. The resulting VP is merged with the
infinitival T constituent to, forming the T‑bar to the shit hit the fan. In con-
formity with the Minimality Condition (61), the EPP feature on T‑to attracts
the closest (pro)nominal constituent (= DP the shit) to become its subject, so
forming the TP the shit to the shit hit the fan. The resulting TP is merged
as the complement of the passive verb expected, forming the VP expected
the shit to the shit hit the fan. This VP is then merged as the complement
of the present-tense T auxiliary is, forming the T‑bar is expected the shit to
the shit hit the fan. In conformity with the Minimality Condition (61), the
EPP feature on the T auxiliary is attracts the closest nominal (= the shit) to
become the subject of T‑is, so forming the TP The shit is expected the shit to
the shit hit the fan. Merging this TP with a null indicative complementiser
derives (87) below, with the structure being interpreted as declarative in
force via the Default Clause Typing Condition (5ii) above:

(87) CP

C TP
ø
DP T′
the shit
T VP
is
V TP
expected
DP T′
the shit
T VP
to
DP V′
the shit
V DP
hit the fan

Although a structure like (87) – and likewise the structure in (85) above –
is traditionally said to involve long(-distance) Passivisation (because it
moves a constituent out of a lower clause into a higher one), in actual fact
it involves two separate local A‑Movement operations by which the idi-
omatic nominal the shit moves first to become the specifier of the lower
TP (thereby becoming the subject of infinitival to) and then to become the
specifier of the higher TP (thereby becoming the subject of is).

2.4.3 Constraints on Passivisation

The analyses of Passivisation in (84, 85, 87) satisfy a number of famil-
iar constraints/conditions. One is the Minimality Condition in (61) above,
repeated as (88) below (where above/below mean ‘c‑commanding/c‑
commanded by’):
104 A‑Movement

(88) Minimality Condition

A constituent can only enter into a syntactic relation with the minimal
(= closest) constituent of the relevant type above/below it

It follows from this condition that, just as an XP undergoing A‑bar

Movement is attracted by the closest C head above it, so too an XP under-
going A‑Movement is attracted by the closest T head above it. So, for
example, A‑Movement in (87) above conforms to the Minimality Condition,
in that the passivised DP the shit is first attracted to become the specifier
of the closest T constituent c‑commanding it (= the embedded clause T
constituent containing infinitival to), and then attracted to become the spe-
cifier of the next closest T constituent c‑commanding it (= the main clause
T constituent containing is).
A key assumption made in (84–87) is that the to‑infinitive complement
of the verb alleged/expected is a TP and not a CP. This is in line with the
traditional assumption that allege/expect are ECM/Exceptional Case Mark-
ing verbs when used with an infinitival complement, and that the infini-
tive complement clause is a defective clause (lacking the CP layer found in
complete clauses) and hence a TP (for reasons outlined in Radford 2020:
192–94). We have independent evidence in support of this claim from con-
trasts such as the following (where — denotes the gap left behind by Pas-
sivisation of you):

(89) a. They didn’t intend [TP you to get hurt]

b. You weren’t intended [TP — to get hurt]

(90) a. They didn’t intend [CP for you to get hurt]

b. *You weren’t intended [CP for — to get hurt]

These examples show that an italicised constituent contained within a TP

complement like that bracketed in (89) can passivise, but not one contained
within a CP complement like that bracketed in (90). Passivisation of a con-
stituent inside a CP is prevented by a constraint (deriving from Chomsky
1998) which was given the following informal characterisation in §1.5.1
(where below can be characterised more precisely as ‘c‑commanded by’ and
above as ‘c‑commanding):

(91) Impenetrability Condition

Anything below the head C of a CP is impenetrable to anything above
the CP

(In simpler terms, this condition prevents syntactic operations from apply-
ing across the edge of an intervening CP.) Since you in (90b) occupies the
gap position below the complementiser for prior to moving to become the
subject of the main clause, it is impenetrable to the T auxiliary weren’t
 2.4 Passive structures 105

(and so cannot be attracted by T‑weren’t to become its subject) because

T‑weren’t c‑commands the complementiser for; or, in simpler terms, the
intervening complementiser for forms an impenetrability barrier to the
relevant movement. Conversely, however, the fact that the shit can passiv-
ise in (87) suggests that the to‑infinitive complement of expected must be
a TP, and not a CP.
The assumption that the A‑Movement operation involved in Passivisa-
tion obeys the Impenetrability Condition offers potential insight into why
Passivisation is possible out of a nonfinite clause like that bracketed in
(92a) below, but not out of a finite clause like that bracketed in (92b):

(92) a. The prisoners were said [ — to have been tortured — ]

b. *The prisoners were said [that/ø — were tortured — ]

In both cases, the DP the prisoners originates as the complement of tor-

tured, then (via one application of A‑Movement) becomes the subject of the
bracketed complement clause, and subsequently (via another application of
A‑Movement) moves into the italicised position as the subject of the main
clause auxiliary were. If we suppose that – as in (84, 85, 87) above – the
infinitive complement in (92a) is a TP, nothing prevents movement out of
the bracketed TP. But now suppose that ‘all finite clauses are CPs’ (Radford
2020: 180). This means that the bracketed complement clause in (92b) will
be a CP (irrespective of whether introduced by an overt or null comple-
mentiser), and that Passivisation in (92b) will involve the two A‑Movement
operations arrowed (i) and (ii) below:

(93) CP2
C TP2
ø
DP T′
the prisoners
T VP
were
V CP1
said
C TP1
that/ø
DP T′
(ii) the prisoners
T VP
were
V DP
tortured the prisoners

(i)

However, there are problems which arise with the derivation in (93).
106 A‑Movement

One is that the Impenetrability Condition (91) will block movement (ii)
from spec‑TP1 to spec‑TP2, since this involves illicit movement across the
edge of a CP (i.e. across the overt/null complementiser which is the head C
of CP1). Moreover, if that is used, (93) will also violate a filter (proposed by
Chomsky & Lasnik 1977: 451) which was characterised as follows in §1.5.2:

(94) COMP‑Trace Filter/CTF

Any structure at the end of a syntactic derivation in which an overt
COMP/complementiser is immediately adjacent to and c‑commands a
trace (i.e. a gap/null copy left behind by a moved constituent) is filtered
out as ill-formed

The violation of CTF arises in (93) if that is used because that c‑com-
mands and immediately precedes a null copy/trace of the moved subject
the prisoners.
To summarise: in Module 2.4, we began by looking at the structure of
simple passive clauses, and I argued that a passive subject originates as
the thematic complement of a subjectless passive participle, and is raised
into spec‑TP (via A‑Movement) in order to satisfy the EPP feature of T. We
also saw that Passivisation can be a cross-clausal/long-distance operation
involving movement of an argument contained within an infinitival TP
which is the complement of a passive participle. I observed that (in conse-
quence of every T carrying an EPP feature requiring it to have a subject),
the passivised nominal moves in a successive-cyclic fashion, first into the
closest spec‑TP position above it, then into the next closest spec‑TP pos-
ition above it (and so on). I went on to argue that Passivisation obeys trad-
itional constraints on syntactic operations (including the Minimality and
Impenetrability Conditions), and the COMP‑Trace Filter.
You should now be able to tackle Exercise 2.4.

2.5 Raising and Control structures

In this module, we turn to examine the syntax of seemingly subjectless

infinitive clauses. We will see that these are of two types. On the one hand,
in a sentence like Jim does seem to support them, the superficial subject Jim
originates as the specifier/thematic subject of the VP headed by the verb
support, and from there moves to become the specifier of the TP headed by
infinitival to; subsequently, it raises to become the specifier of the higher
TP headed by the auxiliary does, leaving behind a trace/null copy of the
raised subject Jim in the infinitival clause (and for this reason infinitives
of this kind are called Raising structures). By contrast, in a sentence such
as ‘Jim does want to support them’ the infinitive complement has a null
 2.5 Raising and Control structures 107

PRO subject which originates as the specifier/thematic subject of support

them and from there moves to become the specifier of infinitival to; this
PRO subject is controlled by (i.e. refers back to) the main clause subject Jim,
and for this reason infinitives like this are referred to as Control structures.
Thus, the two types of infinitive differ in respect of whether the null subject
of the infinitive clause is a trace (i.e. a copy of a raised subject), or PRO. In
this module, we look at differences between Raising and Control Structures,
starting with a look at the former.

2.5.1 Raising structures

The Raising operation by which an argument expression is moved out of
one clause to become the subject of another clause can be illustrated by the
(b) examples in (95–98) below:

(95) a. There does seem [to remain some hope of peace]

b. Some hope of peace does seem [to remain]

(96) a. There does appear [to have been made remarkably little progress]
b. Remarkably little progress does appear [to have been made]

(97) a. It would seem [that Senator Slyme has been lying to Congress]
b. Senator Slyme would seem [to have been lying to Congress]

(98) a. It would appear [that they have underestimated her]

b. They would appear [to have underestimated her]

In (95), the italicised expression some hope of peace is the thematic comple-
ment of the unaccusative predicate remain; it remains in situ in the exple-
tive structure (95a), but raises to become the subject of the seem-clause in
(95b). In (96), the italicised expression remarkably little progress is the the-
matic complement of the passive verb made; it remains in situ in the exple-
tive structure (96a) but raises to become the subject of the appear-clause in
(96b). In (97), the italicised expression Senator Slyme is the thematic sub-
ject of the verb lying: if the complement clause is a finite clause as in (97a),
it surfaces as the subject of the complement clause; but if the complement
clause is infinitival as in (97b), it surfaces as the subject of the seem clause.
Likewise, in (98) the italicised pronoun they is the thematic subject of the
verb underestimate: if the complement clause is finite as in (98a), it surfaces
as the subject of the complement clause; if the complement clause is infini-
tival as in (98b), it surfaces as the subject of the appear clause.
Examples like (95–98) suggest that verbs like seem and appear resemble
passive verbs in that they allow an expression which is a theta-marked ar-
gument of a predicate in a lower clause to raise to become the subject of the
(head T of TP in the) seem-/appear-clause. Given this assumption, sentence
108 A‑Movement

(95b) will have the following simplified derivation. The unaccusative verb
remain merges with the QP some hope of peace to form the VP remain
some hope of ø peace. This VP is then merged with the infinitival tense
particle to, forming the T‑bar to remain some hope of peace. Infinitival to
(like all T constituents) has an EPP feature requiring it to have a noun or
pronoun expression as its subject. One way of satisfying this requirement
is by merging the expletive pronoun there in spec‑TP, so deriving the TP
there to remain some hope of peace. The resulting infinitival TP is then
merged with the verb seem to form the VP seem there to remain some hope
of peace. This VP in turn is merged with the present-tense auxiliary does to
form the T‑bar does seem there to remain some hope of ø peace. Like all T
constituents, T‑does has an EPP feature which enables it to attract the clos-
est noun or pronoun expression (= the expletive pronoun there) to become
its subject, so forming the TP There does seem there to remain some hope
of peace. This TP is then merged with a null indicative complementiser, so
deriving the structure shown in simplified form below:

(99) CP

C TP2
ø
DP T′
there
T VP
does
V TP1
seem
DP T′
there
T VP
to
V QP
remain some hope of peace
In accordance with the Default Spellout Rule (58), only the higher copy of
there is spelled out overtly, and the overall structure in (99) is spelled out as
(95a) There does seem to remain some hope of peace. The sentence is inter-
preted as declarative in force via the default Clause Typing Condition (5ii).
Now consider the derivation of (95b) Some hope of peace does seem to re-
main. Assume that this proceeds as for (95a) until we have formed the T‑bar
to remain some hope of peace. Since infinitival to (like all T constituents
in English) has an EPP feature, it can (in conformity with the Minimality
Condition 61) attract the closest nominal (= some hope of peace) to become
its subject, so forming the TP some hope of peace to remain some hope of
peace. The resulting TP is then merged with the verb seem to form the VP
seem some hope of peace to remain some hope of ø peace. This VP is in turn
 2.5 Raising and Control structures 109

merged with the present-tense T auxiliary does to form the T‑bar does seem
some hope of peace to remain some hope of peace. Like all T‑constituents,
the T auxiliary does has an EPP feature which allows it to attract the closest
nominal (= some hope of peace) to become its subject, so forming the TP
some hope of peace does seem some hope of peace to remain some hope
of peace. Merging this TP with a null indicative complementiser forms the
following structure (with arrows indicating A‑Movement operations which
take place in the course of the derivation):

(100) CP

C TP2
ø
QP T′
some hope
of peace T VP
does
V TP1
seem
QP T′
some hope
of peace T VP
to
V QP
remain some hope
of peace

The movement operation denoted by the upper arrow by which the QP some
hope of peace moves from being the subject of a lower TP to becoming the
subject of a higher TP in (100) is traditionally known as (Subject) Raising,
because it raises a noun or pronoun expression from being the subject of a
lower clause to becoming the subject of a higher clause; but in reality it is
yet another instance of the more general A‑Movement operation by which
T attracts the closest nominal which it c‑commands to become its specifier
(and thereby move to spec‑TP). Words like seem/appear (when used with
an infinitival complement) have the property that the subject of the seem/
appear-clause is created by being raised out of a complement clause, and
so (for this reason) they are known as Raising predicates. The parallels
between Raising in structures like (100) and long-distance Passivisation in
structures like (87) are self-evident.
Although all constituents of the QP some hope of peace are spelled out
on the highest link in the movement chain in (100) above (in accordance
with the Default Spellout Rule in 58), an alternative possibility allowed
for by the Low Spellout Rule (67) is that the KP/case phrase of peace is
spelled out on the lowest link in the movement chain, and the remaining
110 A‑Movement

constituents of the QP (bold-printed below) on the highest link, so giving

rise to the split spellout pattern below:

(101) Some hope does seem to remain of peace

The fact that the KP of peace is positioned after the verb remain in (101)
provides additional evidence that the QP some hope of peace originates as
the complement of remain.
Note in addition that the assumption that the nominal some hope of
peace moves from its underlying position as the complement of the verb
remain into its superficial position as the subject of does in two successive
steps (moving one TP at a time) means that A‑Movement in (100) obeys
the Minimality Condition (61), which requires a constituent undergoing
A‑Movement to become the specifier of the closest T constituent above it
each time it moves. The derivation in (100) also obeys the Theta Criterion
(32), since the QP some hope of peace originates in a θ‑position in comp‑VP
(as the the e argument of the verb remain) and from there moves to a θ‑bar
position as specifier of T‑to, before going on to move to a further θ‑bar
position as specifier of T‑does. Since the resulting A‑chain is assigned only
one theta role, it obeys the Theta Criterion.

2.5.2 Control structures

It might at first sight seem tempting to conclude from our discussion of
cross-clausal Passivisation structures like (87) and Raising structures like
(100) that all clauses containing a structure of the form verb+to+infinitive
have a derivation in which some expression is raised out of the infinitive
complement to become the subject of the main clause. However, any such
conclusion would be undermined by the traditional claim that some verbs
which take to+infinitive complements are Control predicates whose com-
plement contains a PRO subject which is controlled by (i.e. takes its refer-
ence from) an antecedent in the matrix/higher clause. In this connection,
consider the difference between the two types of infinitive structure below:

(102) a. He does seem [to scare them]

b. He does want [to scare them]

As they are used in (102), the verb seem is a Raising predicate, but the verb
want is a Control predicate. We will see that this reflects the fact that the
verbs seem and want differ in respect of their argument structure. We can
illustrate this by sketching the derivation of the two sentences.
In the Raising structure (102a), the verb scare merges with (and assigns
the ex pe r i e nc e r θ‑role to) its internal argument/thematic complement
them. The resulting V‑bar scare them then merges with (and assigns the
agent θ‑role to) its external argument/thematic subject he. The resulting
 2.5 Raising and Control structures 111

VP he scare them is subsequently merged with the infinitival tense particle

to, so forming the T‑bar to he scare them. Like all T constituents, T‑to has
an EPP feature which enables it to attract the closest noun or pronoun ex-
pression to become its subject, so forming the TP he to he scare them. This
TP in turn merges with the Raising verb seem to form the VP seem he to he
scare them. The resulting VP is then merged with the T auxiliary does. The
EPP feature carried by T‑does enables it to attract he to become its subject,
so forming the TP he does seem he to he scare them. This TP is merged with
a null indicative C, so deriving the structure shown in simplified form in
(103) below (where the personal pronouns he/them are analysed as pro-
nominal DPs):

(103) CP

C TP
ø
DP T′
he
T VP
does
V TP
seem
DP T′
he
T VP
to
DP V′
he
V DP
scare them
The successive-cyclic (one-TP-at-a-time) application of A‑Movement in
(103) satisfies the Minimality Condition (61), which requires an A‑moved
constituent to become the specifier of the closest T above it each time it
moves: accordingly, he first becomes the specifier of the embedded clause T
constituent to before moving on to become the specifier of the main clause
T constituent does.
A key assumption made in the Raising analysis in (103) is that the verb
seem (as used there) is a one-place predicate whose only argument is its
infinitival TP complement, to which it assigns an appropriate θ‑role – that
of the e argument of seem. This means that the VP headed by seem has no
thematic subject: note, in particular, that the verb seem does not θ‑mark
the pronoun he, since he is θ‑marked by scare, and the Theta Criterion (32)
rules out the possibility of any argument being θ‑marked by more than one
predicate. Nor does the VP headed by seem have a structural subject at any
stage of derivation, since he raises to become first the subject of the lower
112 A‑Movement

TP headed by infinitival to, and then to become the subject of the higher TP
headed by the present-tense T auxiliary does (with the movement in both
cases being driven by an EPP feature on T). This underlines the point that
syntactic movement operations typically move a constituent to the edge
(i.e. head or specifier position) of a functional projection like TP or CP.
The property of having a t h e e complement but no thematic subject
means that Raising predicates like seem have essentially the same themat-
ic properties as unaccusative verbs like come discussed in Module 2.3. Not
surprisingly, therefore, many intransitive verbs like those below allow a
dual use as (i) an unaccusative verb with an italicised nominal complement
as in each (a) example in (104–107) below, and (ii) a Raising verb with an
underlined infinitival TP complement as in the corresponding (b) example:

(104) a. There appeared a grotesque face at the window

b. His attitude appears to have changed

(105) a. There has never before happened anything quite so momentous

b. It happened to be raining at the time

(106) a. There remain doubts about his competence

b. That remains to be seen

(107) a. There began a long period of negotiation

b. The situation began to get worse

In such sentences, a Raising verb is a one-place predicate with a TP as its

complement, whereas an unaccusative verb is a one-place predicate with a
noun or pronoun expression as its complement.
Having looked at the Raising infinitive structure (102a), let’s now turn
to consider the derivation of the Control infinitive structure (102b) He does
want to scare them. As before, the verb scare merges with (and assigns the
experie nc e r θ‑role to) its internal argument/thematic complement them.
The resulting V‑bar scare them then merges with (and assigns the ag e n t
θ‑role to) its external argument. Given the traditional assumption that
Control infinitives have a particular kind of null pronominal subject known
as ‘big PRO’ (because PRO is written in upper case letters), the thematic
subject of scare them will be PRO, and this will be merged in spec‑VP (in
accordance with the VP-Internal Subject Hypothesis), and thereby be as-
signed the θ‑role of ag e nt argument of scare. The resulting VP PRO scare
them then merges with the infinitival tense particle T‑to, forming the T‑bar
to PRO scare them. The EPP feature of infinitival to enables it to attract
PRO to become its subject, so forming the TP PRO to PRO scare them. Giv-
en the assumption that Control infinitives are CPs (for reasons set out in
Radford 2020: 207–10), this TP will in turn merge with a null infinitival
 2.5 Raising and Control structures 113

complementiser to form the CP ø PRO to PRO scare them. The CP thereby

formed serves as the internal argument (and thematic complement) of the
verb want, so is merged with want and thereby assigned the θ‑role of t h e e
argument of want. The resulting V‑bar want ø PRO to PRO scare them then
merges with its external argument (and thematic subject) he, assigning he
the thematic role of e x pe r i e nc e r argument of want. The resulting VP he
want ø PRO to PRO scare them is then merged with the T auxiliary does,
forming the T‑bar does he want ø PRO to PRO scare them. The EPP feature
carried by T‑does enables it to attract the closest noun or pronoun expres-
sion which it c‑commands (= he) to become its subject, so forming the TP
he does he want ø PRO to PRO scare them. Merging the resulting TP with a
null indicative complementiser to form a CP which is interpreted as declar-
ative in force by default forms the structure shown in simplified form in
(108) below (with arrows marking A‑Movement operations):

(108) CP2

C2 TP
ø
DP T′
he
T VP
does
DP V′
he
V CP 1
want
C1 TP
ø
DP T′
PRO
T VP
to
DP V′
PRO
V DP
scare them

The resulting structure satisfies the Theta Criterion (32) (which requires each
argument to be assigned a single θ‑role, and each θ‑role to be assigned to
a single argument), in that he is the e x pe r i enc er argument of want, the
bracketed CP is the t h e e complement of want, PRO is the ag e n t argument
of scare, and them the e x pe r i e nc er argument of scare. The main clause
subject he is assigned nominative case by the finite complementiser in C2,
and the complement clause subject PRO is assigned null case by the null
infinitival complementiser in C1.
114 A‑Movement

The analysis of Control predicates presented here assumes that the

PRO subject of a Control infinitive like that bracketed in (102b) He does
want to scare them is initially merged in spec‑VP. The requirement for
PRO to be generated in spec‑VP follows from the VP-Internal Subject
Hypothesis and the Predicate-Internal Theta-Marking Hypothesis which
posit that arguments are generated and theta-marked internally to a
projection of their predicate, so that PRO (by virtue of being the ag e n t
argument of scare) is generated as the specifier of the VP headed by
scare. However, given that all T constituents in English have an EPP
feature, the EPP feature on infinitival to will attract PRO to move out
of the VP in which it originates into the specifier position of the TP
headed by infinitival to. Since movement of an empty category like PRO
out of the VP into spec‑TP is by nature invisible, an obvious question
which arises is whether there is any empirical evidence that PRO moves
to spec‑TP rather than remaining in situ within VP. The answer, as we
shall see, is ‘Yes’.
One piece of evidence suggesting that PRO does indeed move to spec‑TP
in Control infinitives comes from the syntax of constituents like those itali-
cised in (109) below which have the property that they are construed as
modifying a bold-printed antecedent which is not immediately adjacent to
them:

(109) a. They were both priding themselves on their achievements

b. I don’t myself think that Hubert Hedge-Hogge should manage the
hedgefund
c. He was personally held responsible

In (109a), both is a floating quantifier (and each/all can be used in a simi-

lar fashion); in (109b), myself is a floating emphatic reflexive; and in
(109c), personally is an argument-oriented adverb (construed as modify-
ing an argument, in this case he). In each sentence in (109), the italicised
expression is construed as modifying the bold-printed subject of the clause.
Contrasts such as those in (110) and (111) suggest that a floating modifier
must be c‑commanded by its bold-printed antecedent:

(110) a. Two republican senators were themselves thought to have been

implicated
b. *There were themselves thought to have been implicated two
republican senators

(111) a. Two republican senators are both thought to have been implicated
b. *There are both thought to have been implicated two republican
senators
 2.5 Raising and Control structures 115

In the light of the requirement for a floating modifier to be c‑commanded

by its antecedent, consider the syntax of the bracketed clauses in the fol-
lowing sentences:

(112) a. [To both be betrayed by their friends] would be disastrous for Romeo
and Juliet
b. [To themselves be indicted] would be unfair on the company directors
c. It was upsetting [to personally have been accused of corruption]

In each of these examples, the bracketed clause is a Control clause con-

taining a PRO argument. In each case, PRO is the thematic complement
of a passive participle (viz. betrayed/indicted/accused). Hence, if Control to
had no EPP feature and PRO remained in situ in comp‑VP, the TP in the
bracketed infinitive complement in (112b) would have the skeletal struc-
ture (113a) below (if we take be to be a passive VOICE auxiliary heading a
VOICEP projection); but if (as claimed here) Control to has an EPP feature,
this will trigger movement of PRO to become the structural subject of to,
resulting in the structure (113b):

(113) a. [CP [C ø] [TP [T to] [VOICEP themselves [VOICE be] [VP [V indicted] PRO]]]]
b. [CP [C ø] [TP PRO [T to] [VOICEP themselves [VOICE be] [VP [V indicted] PRO]]]]

Given the requirement for a floating emphatic reflexive to be c‑commanded

by its antecedent, and given that PRO is the intended antecedent of them-
selves in (113), it is clear that (113a) cannot be the right structure, since PRO
does not c‑command themselves in (113a). By contrast, movement of PRO
to spec‑TP in (113b) means that PRO will indeed c‑command themselves,
so correctly predicting that (112b) is grammatical. Thus, sentences such
as (112) provide empirical evidence that PRO moves to spec‑TP in Control
clauses, and thereby becomes the subject of infinitival to. This in turn is
consistent with the more general claim that all T‑constituents in English
have an EPP feature requiring them to have a subject/specifier.

2.5.3 Differences between Raising and Control

An important syntactic difference between Raising and Control structures
concerns the status of the infinitival complement of the two types of verb.
In this connection, compare the derivation of the Raising structure in
(103) above with that of the Control structure in (108). Whereas the infin-
itive complement of the Raising verb seem in (103) is a TP, by contrast the
infinitive complement of the Control verb want in (108) is a CP. Theoretical
considerations lend support to this conclusion. Since the subject of a
Raising verb is raised out of its infinitive complement, the infinitive com-
plement of a Raising verb cannot be a CP, because the Impenetrability
116 A‑Movement

Condition (91) prevents A‑Movement from applying across a CP bound-

ary. Conversely, if (as claimed in §2.3.2) PRO is assigned null case by a
c‑commanding null complementiser, it follows that the infinitive com-
plement of the Control verb want in (108) must be a CP headed by a null
complementiser.
Having compared the syntax of Control predicates with that of Rais-
ing predicates, let’s look briefly at the question of how we can determine
whether a given predicate which selects an infinitival to complement is a
Control predicate or a Raising predicate. There are a number of differences
between them which are a direct reflection of the different thematic prop-
erties of these two types of predicate. For example, Raising predicates like
seem can have expletive it/there subjects, whereas Control predicates like
want cannot: cf.

(114) a. It seems/*wants to be assumed that he lied to Congress

b. There seem/*want to remain several unanswered questions

(The expletive nature of it in 114a is shown by the fact that it cannot be

substituted by a referential pronoun like this/that, or questioned by what?
Likewise, the expletive nature of there in 114b is shown by the fact that it
cannot be substituted by a referential locative pronoun like here, or ques-
tioned by where?) This is because Control predicates like want are two-place
predicates which project a thematic subject (an e x pe r i e nc e r in the case of
want, so that the subject of want must be an expression denoting a sentient
entity capable of experiencing desires), and non-referential expressions like
expletive it/there clearly do not denote sentient entities and so cannot serve
as the thematic subject of a verb like want. By contrast, Raising predicates
like seem do not theta-mark their subject, and hence impose no restrictions
on the choice of structural subject in their clause, so allowing a (non-the-
matic) expletive subject.
Similarly, Raising predicates like seem (but not Control predicates like
want) allow idiomatic subjects such as those italicised below:

(115) a. All hell seems/*wants to break loose whenever they meet

b. The fur seems/*wants to fly whenever they meet
c. The cat seems/*wants to get his tongue whenever they meet

The ungrammaticality of sentences like *All hell wants to break loose can
be attributed to the fact that want is a Control predicate, and hence (in
order to derive such a structure) it would be necessary to assume that all
hell originates as the subject of want, and that break loose has a separate
PRO subject of its own: but this would violate the requirement that (in its
idiomatic use) all hell can only occur as the subject of break loose, and con-
versely break loose (in its idiomatic use) only allows all hell as its subject.
 2.5 Raising and Control structures 117

By contrast, All hell seems to break loose is grammatical because seem is a

Raising predicate, and so all hell will originate as the subject of break loose
and then be raised up to become the subject of (the present tense affix in T
in) the seem clause.
A further property which differentiates the two types of predicate is that
Raising predicates like seem allow long Passivisation with preservation of
synonymy, so that (116a) below is synonymous with (116b):

(116) a. John seems to have helped Mary

b. =Mary seems to have been helped by John

By contrast, Control predicates like want do not preserve synonymy in long

passives, as we see from the fact that (117a) below is not synonymous with
(117b):

(117) a. John wants to help Mary

b. ≠Mary wants to be helped by John

Moreover, there are selection restrictions on the choice of subject which

Control predicates like want allow (in that the subject generally has to be
an animate being, not an inanimate entity) – as we see from (118) below
(where ! marks anomaly):

(118) My cat/!My gesture wants to be appreciated

By contrast, Raising predicates like seem freely allow animate or inanimate

subjects:

(119) My cat/My gesture seems to have been appreciated

The different properties of the two types of predicate stem from the fact that
Control predicates like want θ‑mark their subjects, whereas Raising predi-
cates like seem do not: so, since want selects an experiencer subject as its
external argument (and a prototypical experiencer is an animate being),
want allows an animate subject like my cat, but not an inanimate sub-
ject like my gesture. By contrast, since Raising predicates like seem do not
θ‑mark their subjects, they impose no restrictions on their choice of subject.
An important point to note is that although our discussion of Raising
and Control predicates has revolved around verbs, a parallel distinction is
found in adjectives. For example, in sentences such as (120), the adjective
likely is a Raising predicate and keen a Control predicate:

(120) a. John is likely to win the race

b. John is keen to win the race

We can see this from the fact that likely allows expletive and idiomatic
subjects, but keen does not: cf.
118 A‑Movement

(121) a. There is likely/*keen to be a strike

b. All hell is likely/*keen to break loose

Moreover, Raising structures like (120a) typically allow an impersonal finite

clause paraphrase in which the likely-clause has an (italicised) non-the-
matic expletive subject:

(122) a. It is likely that John will win the race

b. There is every likelihood that John will win the race

By contrast, Control structures like (120b) typically allow a personal finite

clause paraphrase in which the main and complement clauses have (coref-
erential) personal/thematic subjects (like those italicised below):

(123) John is keen that he should win the race

The fact that adjectives can have the same range of complements as verbs
is one reason why I have talked about different types of predicate (drawing
a distinction between Raising and Control predicates) rather than different
types of verb.
A final point to note is the following. Although a distinction has been
drawn here between Raising and Control predicates, some predicates have
a range of different uses, and are able to function as Raising predicates in
one use, but as Control predicates in another. This can be illustrated by the
sentences below:

(124) a. John promised to do his duty

b. There is promising to be a spectacular end to the season

The verb promise in (124a) shows the hallmarks of a Control predicate, as

we see below:

(125) a. *It promised that John would do his duty

b. John promised that he would do his duty
c. John solemnly promised to do his duty
d. What John promised was to do his duty

Thus, promise requires an animate subject like John in (124a) and does not
allow an expletive subject like it in (125a). Moreover, (124a) has a personal
(John … he) paraphrase in (125b), its subject can be modified by a the-
matic adverb like solemnly (i.e. one which can modify a thematic subject) in
(125c), and its complement can appear in focus position in a pseudo-cleft
like (125d).
By contrast, promise in (124b) behaves like a Raising predicate in al-
lowing an expletive pronoun like there as its subject in (126a) below, in
not allowing the use of a thematic adverb like solemnly in (126b), and
 2.6 Mixed structures 119

in not allowing its complement to be focused in a pseudo-cleft sentence

like (126c):

(126) a. *There is promising that there will be a spectacular end to the season
b. *There is solemnly promising to be a spectacular end to the season
c. *What there is promising is to be a spectacular end to the season

So, while there are clear differences between Raising and Control predicates
(as we have seen), we need to bear in mind that some predicates exhibit
dual behaviour (functioning as Raising predicates in one type of use, and
as Control predicates in another).
You should now be able to tackle Exercise 2.5.

2.6 Mixed structures

Modules 2.3–2.5 in this chapter have provided an introduction to four differ-

ent types of structure: unaccusative, passive, raising and control structures.
Although our discussion of these structures above dealt with each separately, it
is important to note that any given sentence can involve two or more of them,
thereby giving rise to mixed structures. In this module, I’ll briefly illustrate this
by looking at just two (of the many) different types of mixed structure.

2.6.1 A mixed Passive, Control and Raising sentence

An example which illustrates how different types of structure can be com-
bined within a single sentence is the following:

(127) He seems to be keen to be promoted

As we saw in §2.5.3, the verb seem behaves like a Raising predicate in not
theta-marking its subject, and hence allows an unrestricted choice of sub-
jects (including an expletive subject in 128a below and an idiomatic subject
in 128b), and also allows cross-clausal Passivisation in (128c):

(128) a. There seems to be no easy answer

b. All hell seems to have broken loose
c. The plan seems [to have been changed — ]

By contrast, the adjective keen functions as a Control predicate, in that it assigns

the theta role of experiencer to its subject, and hence keen does not allow
non-thematic (e.g. expletive or idiomatic) subjects as in (129a, 129b) below,
nor does it allow long/cross-clausal Passivisation in sentences like (129c):

(129) a. *There is keen to be peace

b. *All hell is keen to break loose
c. *The plan is keen [to be changed — ]
120 A‑Movement

As a Control predicate, keen selects/takes a CP complement with a PRO sub-

ject; but because to be promoted is a passive clause, PRO originates as the
complement of the passive participle promoted, and from there moves to
become the specifier/subject of infinitival to in to be promoted; the control-
ler of PRO (i.e. the constituent which PRO refers back to) is the pronoun he.
Evidence that keen takes a CP complement comes from the observation
that when it has a to‑infinitive complement like that italicised below, its
infinitive complement can be coordinated with a (bold-printed) infinitival
CP headed by the infinitival complementiser for:

(130) He is keen to go to Rome, and for me to go with him

Moreover, to be promoted in our example sentence (127) is paraphrasable as

a CP headed by C‑for, or as a CP headed by C‑that – as below:

(131) a. He seems to be keen [CP [C for] himself to be promoted]

b. He seems to be keen [CP [C that] he should be promoted]

Given these (and other) assumptions, our example sentence (127) will be
derived as follows.
The PRO subject of the infinitive to be promoted will originate as the
the e complement of the passive participle promoted and be attracted by
the EPP feature on the infinitive particle to in T to move to spec‑TP where
it is assigned null case by the null infinitival complementiser heading CP,
thereby deriving the structure in (132) below (where the arrow marks the
Passivisation operation which PRO undergoes, and the position out of
which PRO moves is shown as a gap):

(132) CP

C TP
ø
DP T′
PRO
T VOICEP
to
VOICE VP
be
V DP
promoted

The resulting CP is then merged as the complement of the A/adjective keen,

forming the AP keen ø PRO to be promoted —. The verb be is first-merged
with this AP, and second-merged with the pronoun/pronominal DP he,
forming the VP he be keen ø PRO to be promoted — (where he has the
 2.6 Mixed structures 121

thematic role of the e x pe r i e nc e r argument of be keen). The resulting VP is

merged with the infinitive particle T‑to, and the EPP feature on T‑to attracts
he to move (via the A‑Movement operation arrowed below) to become its
specifier, so deriving the following structure (simplified by not showing the
internal structure of CP, which is given in 132 above):

(133) TP

DP T′
he
T VP
to
DP V′
he
V AP
be
A CP
keen to be promoted

The resulting TP is then merged with the verb seem, forming a VP which
can be represented in simplified form as seem he to be keen to be promoted.
This VP is then merged with a T head containing a present tense affix
(T‑Af), and this affix carries an EPP feature which attracts the pronoun he
to undergo the Raising operation arrowed in (134) below, and thereby move
to spec‑TP. Merging the TP thereby formed with a null complementiser
forms the CP below (simplified by not showing the internal structure of the
VP be keen to be promoted):

(134) CP

C TP
ø
DP T′
he
T VP
Af
V TP
seem
DP T′
he
T VP
to be keen to be promoted
The pronoun he is assigned nominative case by the null finite comple-
mentiser immediately above it. The present tense affix in the main clause
lowers from T onto the verb seem by Affix Hopping in the PF component,
with the result that the verb is spelled out as seems. The derivation satisfies
122 A‑Movement

the Theta Criterion, since each argument carries a single theta role: PRO
is the t h e e argument of promoted, and he the e x pe r i e nc e r argument of
be keen. Both PRO and he originate internally within VP, in accordance
with the VP-Internal Subject Hypothesis. None of the movements arrowed
in (132–134) violates the Impenetrability Condition, since none involves
movement across the edge of an intervening CP. But for present purposes,
what is more important is that our example sentence (127) He seems to be
keen to be promoted has a mixed structure which involves Passivisation,
Control and Raising.

2.6.2 A mixed Unaccusative and Raising sentence

Another type of mixed structure can be found in the sentence below:

(135) Several details have started to emerge of the proposals

In this sentence, the KP/case phrase of the proposals originates as the com-
plement of the noun details. We can see this from the observation that
the relevant KP can be positioned immediately after the noun details (as
expected if it is the complement of details) in a sentence like:

(136) Several details of the proposals have started to emerge

Consequently, it is plausible to take several details of the proposals to be a

QP. This QP originates as the complement of the unaccusative verb emerge,
as we see from the observation that it can follow emerge in an expletive
sentence such as:

(137) There have started to emerge several details of the proposals

But if the QP several details of the proposals originates as the complement

of the verb emerge, how does it come to end up as the subject of the started
clause?
A relevant point to note in this regard is that start can function as a
Raising predicate in not theta-marking its subject, and hence allowing an
unrestricted choice of subjects (including an expletive subject as in 138a
below, and an idiomatic subject as in 138b), and also allowing cross-clausal
Passivisation (as in 138c):

(138) a. There are starting to be problems

b. All hell is starting to break loose
c. A plan is starting [to be developed — ]

These considerations (and others) suggest that the QP several details of the
proposals originates in comp‑VP as the complement of the unaccusative
verb emerge in (135), and then raises up (in a successive-cyclic fashion) first
to become the specifier of infinitival to, and then to become the specifier of
 2.7 Summary 123

T‑have (in each case, attracted by an EPP feature on T‑to/T‑have), so result-

ing in the A‑Movement operations shown by the arrows below:

(139) CP

C TP2
ø
QP T′
several details
of the proposals T2 VP2
have
V TP1
started
QP T′
several details
of the proposals T1 VP1
to
V QP
emerge several details
of the proposals

When the structure in (139) is handed over to the PF component, spellout

rules determine what is overtly spelled out, and what is silent. One possi-
bility is that (via the Low Spellout Rule 67) the lowest copy of the KP of the
proposals and the highest copy of the string several details are spelled out
overtly, so giving rise to the discontinuous spellout pattern that we find in
(135) ‘Several details have started to emerge of the proposals’. However, for
the purposes of this module, the more important point to underline is that
(135/139) is a mixed Raising+unaccusative structure, serving to underline
the key point made in this module that sentences can be built up out of a
number of different types of structure.
You should now be able to tackle Exercise 2.6.

2.7 Summary

This chapter has been concerned with the syntax of subjects. In Module 2.1,
I argued that Belfast English sentences such as There should some students
get distinctions provide us with evidence that subjects originate internally
within VP, and I noted that this claim is known as the VP-Internal Subject
Hypothesis/VPISH. I also argued that standard English sentences such as
Some students should get distinctions involve movement of some students
from the specifier position within VP to the specifier position within TP, and
we saw that the relevant movement operation is known as A‑Movement.
I went on to argue that idiomatic structures like All hell will break loose
provide empirical support for the VPISH.
124 A‑Movement

In Module 2.2, I noted that the VP-Internal Subject Hypothesis allows

us to posit a uniform mapping (i.e. relationship) between thematic argu-
ment structure and syntactic structure, if we suppose that all arguments of
a predicate originate (and are theta-marked) internally within a projection
of the predicate.
In Module 2.3, we looked at the syntax of unaccusative predicates like
arise/remain/occur etc. and we saw that the subject of an unaccusative
verb originates as its complement but differs from the complement of a
transitive verb in that (in a finite clause) it receives nominative rather than
accusative case. We examined a number of further differences between
unaccusative predicates and other types of predicate (e.g. in relation to the
position of subjects in Belfast English imperatives, and auxiliary selection
in earlier varieties of English).
In Module 2.4, we began by looking at the structure of simple passive
clauses, and I argued that a passive subject originates as the thematic com-
plement of a subjectless passive participle, and is raised into spec‑TP (via
A‑Movement) in order to satisfy the EPP feature of T. We also saw that
Passivisation can be a long-distance operation involving cross-clausal
movement of an argument contained within an infinitival TP which is
the complement of a passive participle. I observed that (in consequence of
every T carrying an EPP feature requiring it to have a subject), the passi-
vised nominal moves in a successive-cyclic fashion, first into the closest
spec‑TP position above it, then into the next closest spec‑TP position above
it (and so on), in conformity with the Minimality Condition.
In Module 2.5, I argued that predicates like seem/appear function as
Raising predicates in the sense that their subjects originate internally with-
in their infinitive complement, and from there are raised into the spec‑TP
position within the seem/appear-clause: hence, in a sentence such as All
hell would appear to have broken loose, the idiomatic expression all hell
originates as the subject of broken loose and from there is raised up (one
TP at a time) first to become the specifier/subject of the lower TP headed
by infinitival to, and then to become the specifier/subject of the higher TP
headed by the auxiliary would. I contrasted Raising predicates with Control
predicates, noting that they differ in that Control predicates theta-mark
their subjects (and hence require an animate subject) and have a CP com-
plement, whereas Raising predicates do not theta-mark their subjects (and
hence freely allow inanimate, expletive and idiomatic subjects) and have a
TP complement. We also saw that (unlike Control predicates), Raising predi-
cates allow long/cross-clausal Passivisation.
In Module 2.6, I argued that complex sentences can contain a mixture of
different types of (e.g. unaccusative, passive, raising and/or control) predi-
cates, so giving rise to what I termed ‘mixed structures’.
 2.8 Bibliographical notes 125

Key constructs which our discussion in this chapter made use of in-
clude the following (where below = ‘c‑commanded by’, and above = ‘c‑
commanding’):

(32) Theta Criterion/θ‑criterion

Each argument bears one and only one θ‑role, and each θ‑role is
assigned to one and only one argument (Chomsky 1981: 36)

(36) Predicate-Internal Theta-Marking Hypothesis

An argument is theta-marked internally within a projection of a
predicate, either directly by being merged with the predicate, or
indirectly by being merged with an intermediate projection of the
predicate

(56) Uniform Theta Assignment Hypothesis/UTAH

Constituents which fulfil the same thematic role with respect to a given
predicate occupy the same initial position in the syntax

(61) Minimality Condition/MC

A constituent can only enter into a syntactic relation with the minimal
(= closest) constituent of the relevant type above/below it

(91) Impenetrability Condition

Anything below the head C of a CP is impenetrable to anything above
the CP

(94) COMP‑Trace Filter/CTF

Any structure at the end of a syntactic derivation in which an overt
COMP/complementiser is immediately adjacent to and c‑commands a
trace (i.e. a gap/null copy left behind by a moved constituent) is filtered
out as ill-formed

2.8 Bibliographical notes

Evidence adduced in support of the VP‑Internal Subject Hypothesis out-

lined in Module 2.1 can be found in a variety of works dating back to the
mid-1980s, including Kitagawa (1986, 1994), Speas (1986), Sportiche (1986,
1988), Contreras (1987), Zagona (1987), Kuroda (1988), Rosen (1990), Ernst
(1991), Koopman & Sportiche (1991), Woolford (1991), Burton & Grimshaw
(1992), McNally (1992), Guilfoyle et al. (1992), Huang (1993), and McCloskey
(1997). On in-situ subjects (i.e. structures in which subjects remain in
situ internally within VP), see Haider (2005) and Wurmbrandt (2006) for
German; Alexiadou & Anagnastopoulou (2001) for Greek; Tateishi (1995)
for Japanese; Öztürk (2005) for Turkish; and Berger (2015) for Spanish.
126 A‑Movement

For discussion of quantifier scope in A‑Movement structures, see Leb-

eaux (1995), Hornstein (1995), Romero (1997), Sauerland (1998), Lasnik
(1998, 1999), Fox (2000), Boeckx (2000, 2001) and Fox & Nissenbaum
(2004). In relation to the EPP feature, it should be noted that some lin-
guists have attempted to eliminate EPP as a factor in driving movement of
a subject out of VP into spec‑TP (e.g. Grohmann et al. 2000, Epstein et al.
2005, Bošković 2007, Castillo et al. 2009, Cable 2012, and Chomsky 2015).
On expletive there structures in Belfast English, see Borsley (2010); and on
the semantics of expletive there, see Moro (1997, 2006), and Kayne (2016,
2019). The ideas on theta roles discussed in Module 2.2 date back to pion-
eering work by Gruber (1965, 1976), Fillmore (1968, 1970), and Jackendoff
(1972). On the Theta Criterion, see Chomsky (1981, 1986a, 1995, 2021),
and Collins (2021); for a critical perspective, see Heim & Kratzer (1998),
Bruening (2013), Legate (2014), and Alexiadou et. al. (2015). The discussion
of unaccusative predicates in Module 2.3 and the idea that unaccusative
subjects originate as the complements of their associated verbs dates back
to work by Burzio (1986) on Italian, and Contreras (1986) on Spanish.
On the Constraint on Extraction Domains/CED, see Nunes & Uriagereka
(2000), Sabel (2002), Rackowski & Richards (2005), Chomsky (2005, 2008),
Stepanov (2007), Müller (2010), Jurka (2010), Jurka et al. (2011), Sheehan
(2013a, 2013b), and Sprouse et al. (2013). On potential problems with CED,
see Starke (2001: 40, fn.10), Stepanov (2001, 2007), Truswell (2007, 2009,
2011), Chomsky (2008), Chaves (2012, 2013), Haegeman et al. (2014), and
Muñoz Pérez et al. (2021).
For a range of views on the nature of the COMP‑Trace Filter, see Per-
lmutter (1968, 1971), Chomsky & Lasnik (1977), Pesetsky (1982b, 2016),
Bayer (1984a), Sobin (1987, 2002, 2009, 2016), Browning (1996), Szczegiel-
niak (1999), Roussou (2000), Pesetsky & Torrego (2001, 2004), Ackema &
Neeleman (2003), Kandybowicz (2006), Lohndal (2009), Ackema (2010), Ha
(2010), Bošković (2011a), Llinàs-Grau & Fernández Sánchez (2013), Sato &
Dobashi (2013), Erlewine (2014, 2017a), Bošković (2016), Brillman & Hirsch
(2016), Douglas (2017), and Pesetsky (2021).
The use of (the counterpart of) be rather than have as a perfect aspect (or past
tense) auxiliary with unaccusative verbs is found in a number of languages –
for example, Italian and French (Burzio 1986), Sardinian (Jones 1994), German
and Dutch (Haegeman 1994), and Danish (Spencer 1991): see Sorace (2000) for
further discussion of auxilary selection. On unaccusative structures like The
stockmarket dropped 250 points today, see Nakajima (2006).
On the syntax of passives (and the use of by‑phrases) in structures like
those discussed in Module 2.4, see Jaeggli (1986), Roberts (1987), Baker et al.
(1989), Åfarli (1989), Mahajan (1994), Goodall (1997), Embick (1998), Bow-
ers (2002), Collins (2005, 2018, 2021), Meltzer-Asscher (2012), Alexiadou &
 2.9 Workbook 127

Schäfer (2013), Bruening (2013), Legate (2014), Alexiadou et al. (2015, 2018),
and Angelopoulos et al. (2020). On the syntax of get‑passives like John got
arrested, see Haegeman (1985), Fox & Grodzinsky (1998), Authier & Reed
(2009), Brownlow (2011), and Reed (2011). The Uniform Theta Assignment
Hypothesis was devised by Baker (1988: 46; 1997: 74), but it is rejected (in-
ter alia) by proponents of the Movement Theory of Control (see below). For
evidence that long/cross-clausal A‑Movement applies in a successive-cy-
clic fashion, see Bošković (2002), Lasnik (2006), and Boeckx (2007).
For more detail on the properties of Raising predicates discussed in Mod-
ule 2.5 and on the differences between Control and Raising predicates, see
Davies & Dubinsky (2004). For evidence that Control infinitives are CPs,
see Rizzi (1982), van Riemsdijk (1985), Kayne (1989, 1991), Wilder (1989),
Branigan (1992) and Sabel (1996). Note that some verbs (e.g. begin, con
tinue, fail, and start) are ambiguous between being Raising or Control verbs
(in the sense that they can be used as either): see Perlmutter (1970), and (on
the learnability problem which this poses for children) Becker (2006).
Finally, it should be noted that this chapter presents the classical theory
of Control assumed by Chomsky in his own work. An alternative Movement
Theory of Control has been developed by Bowers (1973), O’Neil (1995),
Hornstein (1999, 2001, 2003), Boeckx (2000, 2007), Manzini & Roussou
(2000), Boeckx & Hornstein (2003, 2004, 2006a, 2006b), and Grohmann
(2003): the key assumption of the movement analysis is that in a sentence
such as Jim will try to open it, the nominal Jim originates as the subject of
the embedded clause verb open and then raises up to become the subject of
(and acquire an additional theta role from) the main clause verb try. For a
critique of the movement analysis, see Landau (2006). For a very different
account of Control which posits that PRO does not exist and that Control
clauses are subjectless, see Jackendoff & Culicover (2003), and Culicover &
Jackendoff (2001, 2005, 2006).

2.9 Workbook

In the exercises below, an [M] after an example sentence indicates that a

model answer is provided in the free-to-download Students’ Answerbook
(and in the Teachers’ Answerbook as well). You will find it helpful to look at
this model answer before attempting to tackle other examples in the exer-
cise. Some other exercise examples have an [S] after them, indicating that
they are for self-study: this means that after reading the relevant module
(and those preceding it), and looking at the helpful hints and model answer
associated with the exercise, students should be able to analyse these self-
study examples on their own, and then compare their answers with those
128 A‑Movement

provided in the Students’ Answerbook. The remaining exercise examples

(i.e. those not marked with [M] or [S] after them) are intended for teachers
to use as the basis for hands-on problem-solving work in seminars, classes,
or assignments: answers to all of these non-self-study examples are pro-
vided in the Teachers’ Answerbook. Note that Exercise 2.1 tests you on
Module 2.1, Exercise 2.2 on Module 2.2, and so on; and that each exercise
in a given chapter presupposes familiarity with material covered in earlier
chapters and in earlier modules of the same chapter. Unparenthesised num-
bers like 4 refer to material in the exercises, whereas parenthesised numbers
like (22) refer to examples and conditions in the main text of the book.

EXERCISE 2.1

This exercise is designed to be attempted after reading Module 2.1 and the
chapter preceding it.

In §2.1.2, it was suggested that evidence from floating quantifiers (i.e.

quantifiers which ‘float’ in a position below – and are separate from – the
subject which they are construed as modifying) provides an argument
in support of the VP-Internal Subject Hypothesis, if subjects originate
internally within VP and from there move to spec‑TP by A‑Movement.
Bearing in mind the helpful hints below and the two alternative analyses
sketched in the model answer in the Answerbook, discuss whether this
claim is compatible with the syntax of the universal quantifiers all/both in
sentences such as the following:

1a Both the detectives have examined the body [M]

1b The detectives have both examined the body [M]
2a Both the boys were proud of her
2b The boys were both proud of her
3a Both the boys must have been playing golf [S]
3b The boys must have been both playing golf [S]
3c The boys must have both been playing golf [S]
3d The boys must both have been playing golf [S]
4a All the men were arrested
4b The men were all arrested
4c *The men were arrested all

Helpful hints
Consider the possibility that floating quantifiers occupy positions which
the phrases they head originate in (or transit through) on the way to
spec‑TP. Compare the two different accounts of floating quantifiers
given in the model answer (in the Answerbook), and see how (or indeed
 2.9 Workbook 129

whether) either analysis can account for each of the example sentences.
For the purposes of this exercise, assume that have in 1a and 1b is a finite
auxiliary directly generated in T (setting aside the possibility that it may
originate as the head of a PERFP/perfect aspect phrase projection, and
from there move to T). In 2a and 2b, bear in mind the traditional view
that the copula b e (when finite) originates in the head V position of VP
and (by virtue of being auxiliary-like) from there moves to the head T
position of TP in finite clauses (adjoining to a null tense affix in T). In
3a–3d, assume that been occupies the head PROG position of a PROGP/
progressive aspect phrase, that have occupies the head PERF position of a
PERFP/perfect aspect phrase, and that must occupies the head T position
of TP (ignoring the possibility that must may originate as the head of a
lower MP/modal projection and from there move to T). In relation to 4a–
4c, assume that the passivised subject the prisoners originates in comp‑VP
(as the complement of the verb arrested) and from there moves to
spec‑TP, via A‑Movement/Passivisation. Bear in mind that the Minimality
Condition means that A‑Movement is successive-cyclic, and consider the
possibility of the subject moving first to the closest non-thematic specifier
position above it (e.g. as the specifier of the closest auxiliary), then to the
next closest … and so on. Bear in mind the spellout rules discussed in the
model answer in the Answerbook, and take into account the following
conditions on extraction/movement:

5 Constraint on Extraction Domains/CED (= 42)

Extraction is only possible out of a complement, not out of a specifier or
adjunct

6 Freezing Principle (from §1.5.1)

The constituents of a moved phrase are frozen internally within (and so
cannot be extracted out of) the moved phrase

EXERCISE 2.2

This exercise is designed to be attempted after reading Module 2.2 and the
module and chapter preceding it.

Discuss the role that thematic structure can play in accounting for
similarities/differences between the two or three sentences in each
numbered set below:

1a John handed Mary the phone [M]

1b John handed the phone to Mary [M]
2a John smashed the window with a brick [S]
130 A‑Movement

2b A brick smashed the window [S]

2c The window smashed [S]
3a John got an invitation from Mary [S]
3b Mary got John an invitation [S]
4a The police deliberately killed the prisoner
4b *The prisoner deliberately died
5a Personally, syntax bores me
5b *Personally, my wife beats me
6a The bombing of the airport by the terrorists was tragic [S]
6b The disappearance of/* by the terrorists was mysterious [S]
7a Mary brought John a lot of happiness last summer
7b Last summer brought John a lot of happiness
7c !Mary and last summer brought John a lot of happiness

In addition, discuss any problems posed for the claim that thematic rela-
tions play a key role in accounting for the syntax of adverbial/prepositional
constituents and coordination by sentences such as the following (where !
denotes semantic/pragmatic anomaly):

8a Personally, I have never sent a Christmas card to Boris

8b Personally, I have never been sent a Christmas card by Boris
8c Personally, I have never been sent to Boris
9a Gradual economic decline by Britain has led to worsening unemployment [S]
9b Any further fall by stocks and shares could undermine the economy [S]
10a !John and a tin of paint fell off the ladder
10b !I hit John and the ground with my stick
10c !I saw John and a daffodil in the park

Helpful hints
Do not attempt to draw tree diagrams representing the structure of the
relevant phrases and sentences. Instead, identify the thematic roles played by
key constituents, and say what part theta roles play in accounting for (i) how
constituents occupying different positions can nonetheless have essentially
the same semantic function, (ii) the distribution of adverbs and prepositional
phrases, and (iii) the types of constituent that can be coordinated.

EXERCISE 2.3

This exercise is designed to be attempted after reading Module 2.3 and the
modules and chapter preceding it.

Discuss the derivation of the sentences below: those in 6 and 7 are taken
from Shakespeare’s play Measure for Measure; and those in 8 and 9 were
 2.9 Workbook 131

produced by the children whose names and ages in years; months are
shown in parentheses (the relevant child data being from Pierce 1992).

1 On the wall, a portrait of Karl Marx had hung [M]

2 At which window did a ghostly figure appear? [S]
3 An opportunity for us to go to Rome has arisen [S]
4 Allegations of police corruption have emerged
5 Several of the prisoners have escaped
6 Now are they come (Duke, Measure for Measure, [Link]) [S]
7 Is the duke gone? (Duke, Measure for Measure, V.i)
8 Fall the cradle (Peter 2;2 = ‘The cradle fell’)
9 Come Cromer? (Adam 2;3 = ‘Has Mr Cromer come?’)

Helpful hints
Consider whether the italicised verbs in the above sentences are
unaccusative predicates, and if so what evidence there is for such an
assumption. For the purposes of this exercise, take the PPs on the wall/at
which window in 1/2 and the ADVP now in 6 to be generated in spec‑CP
(ignoring the possibility that they may originate in some lower position
below C and from there move to spec‑CP). Treat a/an-phrases as having
the status of ARTP/article projections, and of-phrases as having the status
of KP/case projections, but do not concern yourself with the internal
structure of these ARTP/KP constituents. In relation to 6 and 7, assume
that Shakespearean English is a verb-second/V2 language in which a null
C with a specifier in a root/main clause attracts an (auxiliary or non-
auxiliary) verb in T to adjoin to C. In 7, assume that yes–no questions
contain an ADVP comprising a null yes–no operator/OpYNQ in spec‑CP.
Compare two different accounts of child utterances like those in 8 and
9. On one (structural deficit) view, two-year-old children’s sentence
structures are smaller than their adult counterparts in that they can lack
functional projections like CP and TP found in adult sentences, and are
simply VPs; on an alternative (continuity) view, two-year-old’s sentences
have the same CP+TP+VP structure as their adult counterparts, but the
children may have not yet mastered the morphophonology and syntax of
functional heads (e.g. how tense/agreement morphology on T is spelled
out, and how/when T and C trigger movement).
In devising your answers, you might want to bear in mind the
following:

10 Low Spellout Rule (= 67)

A KP/PP/CP which is the lowest/rightmost constituent of a larger phrase
that undergoes movement can be spelled out on a lower copy of the
moved phrase
132 A‑Movement

11 Default Spellout Rule/DSR (= 7)

For a constituent whose spellout is not determined in some other way,
the highest copy of the constituent is overtly pronounced, and any lower
copies are silent

12 Clause Typing Conditions (= 5)

(i) A clause is interpreted as being of a given type (interrogative or
exclamative etc.) if it contains a peripheral clause-typing specifier of
the relevant type
(ii) An indicative clause is interpreted as declarative in type by default if
it contains no peripheral clause-typing specifier

EXERCISE 2.4

This exercise is designed to be attempted after reading Module 2.4 and the
modules and chapter preceding it.

Discuss the derivation of the passive sentences below, accounting for why
they are (un)grammatical:

1 They are said to have made allegations of corporate corruption [M]

2 *They are said that have made allegations of corporate corruption [M]
3 Allegations of corporate corruption were said to have been made [S]
4 Allegations were said to have been made of corporate corruption [S]
5 There were said to have been made allegations of corporate corruption
6 *There were said allegations of corporate corruption to have been made
7 *Allegations of corporate corruption were said that had been made
8 Allegations of corporate corruption are expected (*for) to be made

Helpful hints
For the purposes of this exercise, make the simplifying assumption
that finite auxiliaries occupy the head T position of TP, but that
nonfinite auxiliaries are the heads of auxiliary projections below T
(e.g. have is the head PERF constituent of a PERFP/perfect aspect
phrase, and been in passive uses is the head VOICE constituent of a
passive VOICEP/voice projection). Assume that definite nominal or
pronominal arguments are DPs, and that indefinite nominal arguments
are QPs headed by an overt or null Q. Bear in mind the role of the
Impenetrability Condition in 9 below in blocking movement across
the edge of an intervening CP, and also bear in mind the role of the
 2.9 Workbook 133

COMP‑Trace Filter in 10 in filtering out certain types of superficial

syntactic structure as illicit:

9 Impenetrability Condition (= 91)

Anything below the head C of a CP is impenetrable to anything above
the CP (where below = c‑commanded by, and above = c‑commanding)

10 COMP‑Trace Filter/CTF (= 94)

Any structure at the end of a syntactic derivation in which an overt
COMP/complementiser is immediately adjacent to and c‑commands a
trace (i.e. a gap/null copy left behind by a moved constituent) is filtered
out as ill-formed

Consider also the possibility that some of the sentences may involve split/
discontinuous spellout, and in this connection, bear in mind the following
two PF spellout rules:

11 Low Spellout Rule/LSR (= 67)

A KP/PP/CP which is the lowest/rightmost constituent of a larger phrase
that undergoes movement can be spelled out on a lower copy of the
moved phrase

12 Default Spellout Rule/DSR (= 7)

For a constituent whose spellout is not determined in some other way,
the highest copy of the constituent is overtly pronounced, and any lower
copies are silent

Finally, note that nothing in the account of Passivisation given in the book
accounts for the ungrammaticality of sentences like 6. See if you can sug-
gest possible factors that might account for the ungrammaticality of such
sentences.

EXERCISE 2.5

This exercise is designed to be attempted after reading Module 2.5 and the
modules and chapter preceding it.
Discuss the derivation of each of the sentences below, saying whether
the italicised verb in each sentence functions as a Raising or Control
predicate (or is potentially ambiguous between the two), and adducing
evidence in support of your analysis.

1 Employers tend to exploit employees [M]

2 He has decided to do it [M]
134 A‑Movement

3 He’s beginning to irritate me [M]

4 We came to appreciate the classes (in the sense of ‘We ended up appreciating
the classes’) [S]
5 They will have to do something
6 They have failed to score enough goals [S]
7 He tried to sabotage the car [S]
8 He refused to sign the petition
9 They attempted to pervert the course of justice
10 I happened to be passing the house
11 He is going to help me
12 He is bound to win the election
13 He needs to have a shave
14 We are hoping to get a visa [S]
15 He is threatening to sue the CIA [S]

Helpful hints
Bear in mind that because Raising predicates don’t theta-mark their
subjects, they allow an unrestricted choice of subjects (including
inanimate, expletive, and idiomatic subjects); by contrast, because Control
predicates theta-mark their subjects, they impose selection restrictions
on their choice of subject (typically allowing animate subjects, but
not expletive, idiomatic or inanimate subjects). For similar reasons,
Raising predicates (if they allow a finite clause paraphrase) typically
allow a finite clause paraphrase with an impersonal/expletive main-
clause subject, whereas Control predicates (if they allow a finite clause
paraphrase) typically allow a finite clause paraphrase with a personal/
thematic main-clause subject. Bear in mind that Raising predicates
select an infinitival TP complement, whereas Control predicates select
an infinitival CP complement: so (like other CPs), the complement of a
Control predicate can generally appear in focus position in a pseudo-cleft
(but the complement of a Raising predicate cannot). Another difference
is that Raising predicates allow long/cross-clausal Passivisation across
their TP complement, but Control predicates do not (because they have a
CP complement, and the Impenetrability Condition blocks A‑Movement
across the edge of an intervening CP). A complication to bear in mind
is that some predicates allow a dual use, and can function either as
Raising or Control predicates. Consider whether your derivation satisfies
requirements imposed by the conditions below (where above means
‘c‑commanding’, and below means ‘c‑commanded by’):

16 Theta Criterion/θ‑criterion (= 32)

Each argument bears one and only one θ‑role, and each θ‑role is
assigned to one and only one argument
 2.9 Workbook 135

17 Minimality Condition/MC (= 61)

A constituent can only enter into a syntactic relation with the minimal
(= closest) constituent of the relevant type above/below it

18 Impenetrability Condition (= 91)

Anything below the head C of a CP is impenetrable to anything above
the CP

For the purposes of the exercise, assume that the infinitive complement of
each of the italicised predicates has the thematic role of being the t h e e
argument of the relevant italicised predicate; and assume, too, that only
T constituents (containing a finite auxiliary, a tense affix or the infini-
tive particle to) carry an EPP feature. In addition (again for the purposes
of this exercise) make the simplifying assumption that finite auxiliaries
are directly generated in T, rather than originating as the head of a lower
auxiliary projection and from there raising to adjoin to a tense affix in T.
In 6, take enough goals to be a QP/quantifier phrase, and in 13–14, take a
shave/a visa to be ARTP/article phrase constituents.

EXERCISE 2.6

This exercise is designed to be attempted after reading Module 2.6 and the
modules and chapter preceding it.

Discuss the derivation of each of the examples in 1–5 below, presenting

evidence in support of your analysis:

1 The negotiations do seem likely to prove difficult [M]

2 No traces would appear to have been found of any dinosaurs
3 He is thought to be intending to sell it [S]
4 No soldier expects to die
5 Evidence of impropriety is starting to emerge

In addition, discuss the derivation of sentences 6–10 below (recorded by

me from live, unscripted radio broadcasts), saying what is unusual/non-
standard about them:

6 Six venues are planned to be used (Journalist, BBC Radio 5)

7 Next of kin are still trying to be identified (Reporter, BBC News TV) [S]
8 Some of these are thought will have medicinal values ([Link]) [S]
9 These works are hoped will start during February 2013 ([Link])
10 They are deemed that they are too expensive (Spokeswoman for the National
Health Service, BBC Radio 5; both instances of they refer to drugs for treating
cancer)
136 A‑Movement

Helpful hints
Note that the examples in this exercise are mixed structures involving a
combination of clauses containing Raising/Control/passive/unaccusative
predicates: you might find it useful to bear in mind the helpful hints
given for Exercises 2D and 2E. Bear in mind too the following additional
hints in relation to the sentences in 6–10:

• Some speakers may allow an item which normally selects a CP comple-

ment to select a TP complement
• Some speakers may allow some Control predicates to be used without a
thematic subject as Raising predicates
• Conditions like the Impenetrability Condition and the COMP‑Trace Fil-
ter mean that A‑Movement is not possible across an intervening CP
(though is possible across an intervening TP)
• Structures in which constraints on movement prevent a gap inside some
island from having an antecedent outside the island can be repaired if
an overt copy is used in place of a gap/null copy
In relation to 2, bear in mind the possibility of split spellout arising from
applying the Low Spellout Rule in 11 below, rather than the Default
Spellout Rule in 12:

11 Low Spellout Rule/LSR (= 67)

A KP/PP/CP which is the lowest/rightmost constituent of a larger phrase
that undergoes movement can be spelled out on a lower copy of the
moved phrase

12 Default Spellout Rule/DSR (= 7)

For a constituent whose spellout is not determined in some other way,
the highest copy of the constituent is overtly pronounced, and any lower
copies are silent

Where relevant, bear in mind the following filter:

13 COMP‑Trace Filter (= 94)

Any structure at the end of a syntactic derivation in which an overt
COMP/complementiser is immediately adjacent to and c‑commands a
trace (i.e. a gap/null copy left behind by a moved constituent) is filtered
out as ill-formed

In 6, take six venues to be a NUMP/numeral projection, but don’t concern

yourself with its internal structure. In 7, take still to be an ADVP which
adjoins to the VP headed by trying to form an even larger VP; and treat
next of kin as a QP headed by a null quantifier ø meaning ‘an unspecified
number of’ (but don’t try to analyse the internal structure of this QP). In 8,
take some of these and medicinal values to be QPs (in the latter case, headed
 2.9 Workbook 137

by a null quantifier ø meaning ‘an unspecified number of’), but don’t con-
cern yourself with their internal structure. In 9 take during February 2013
to be a PP which is an adjunct to the VP headed by start, but do not con-
cern yourself with the internal structure of this PP. In 10 take too expensive
to be an AP/adjectival phrase, but don’t concern yourself with its internal
structure; and in addition, take are (in are too expensive) to be a copular
verb which originates as the head V of VP, and from there moves to adjoin
to a present tense affix in the head T position of TP.
3 Agreement

3.0 Overview

In this chapter, we take a look at the syntax of agreement. I begin by argu-

ing that agreement involves a relation between a probe and a goal (though
it should be noted that the term ‘goal’ in this chapter is used in an entirely
different way from the term g oa – written in capital letters – which was
used to denote the thematic role played by a particular kind of argument
in relation to its predicate in the Chapter 2). I then go on to explore how
agreement works in a range of structures, and the role which agreement
plays in A‑Movement on the one hand and Case-marking on the other.

3.1 Agreement, A‑Movement and Case-marking

This module begins by considering the nature of Agreement, and goes on

to explore its relation to A‑Movement and Case-marking.

3.1.1 Agreement and A‑Movement

In traditional grammars, finite auxiliaries are said to agree with their sub-
jects in person and number. Since (in the classic CP+TP+VP model outlined
in Chapter 1) finite auxiliaries occupy the head T position of TP and their
subjects are in spec‑TP, in earlier work agreement was taken to involve
a specifier-head relationship between T and its specifier – as illustrated
below:

(1) a. Numerous complications have arisen

b. The responsibility of taking decisions has lain on senators
c. His rivals have amassed a much smaller fortune than Senator
Swindle
d. Inflation has gone up, and standards of living have come down
e. The negotiations have been so protracted that progress is slow

In each of the above examples, the bold-printed T auxiliary h av e agrees

with its italicised superficial subject, making it seem plausible to suppose
that agreement involves a spec(ifier)-head relation between an auxiliary in
T and its specifier in spec‑TP.
 3.1 Agreement, A‑Movement and Case marking 139

However, a spec–head account of agreement proves problematic in that it

fails to account for agreement between the auxiliary and its subject in cases
where the subject is postverbal (perhaps remaining in situ inside VP) – as
is the case with agreement between the bold-printed auxiliary h av e and its
italicised postverbal subject in structures like those below:

(2) a. There have arisen numerous complications

b. On senators has lain the responsibility of taking decisions
c. Senator Swindle has acquired a larger fortune than have his rivals
d. Up has gone inflation, and down have come standards of living
e. So protracted have been the negotiations that progress is slow

Since the agreeing subject is postverbal in (2a–2e) above, it is clear that any
account of agreement as involving a relation between T and its specifier
would be potentially problematic in such cases.
So let’s explore an alternative account which makes use of the VP-
Internal Subject Hypothesis, and let’s suppose that agreement involves a
relation between a T constituent (e.g. a finite auxiliary) and a VP‑internal
constituent which it c‑commands. To see how this works, consider the der-
ivation of the sentence produced by speaker B below:

(3) s p ea k e r a : What happened to the protestors?

s p ea k e r b : They were arrested

In particular, let’s look at the features carried by the pronoun they on the
one hand, and the auxiliary were on the other, and how these are central to
agreement and movement. In (3b ), we can suppose that a third person plural
pronoun is required in order to refer back to the third person plural con-
stituent the protestors, and that a past tense auxiliary is required because the
event described took place in the past. So (as it were) the person/number fea-
tures of they and the past tense feature of were are determined in advance,
before the items enter the derivation. By contrast the case feature assigned
to they and the person/number features assigned to were are determined in
the course of the derivation: for example, if the subject had been the third
person singular pronoun one, the auxiliary would have been third person
singular via agreement with one (as in ‘One was arrested’); and if they had
been used as the object of a transitive verb, it would have been spelled out
in the accusative form them (as in ‘The police arrested them’).
Generalising at this point, let’s suppose that noun and pronoun expres-
sions like they enter the syntax with their person and number features
intrinsically valued, but their case feature unvalued. (The notation t h e y is
used here to provide a case-independent characterisation of the word which
is variously spelled out as they/them/their depending on the case assigned
to it in the syntax.) Using a transparent feature notation, let’s say that t h e y
140 Agreement

enters the derivation carrying the features [3‑Pers, Pl‑Num, u‑Case], where
3 = third, Pers = person, Pl = plural, Num = number, and u = unvalued.
Similarly, let’s suppose that finite T constituents enter the derivation with
their tense feature already valued, but their person and number features as
yet unvalued (because they will be valued in the course of the derivation
via agreement with an appropriate constituent). This means that the aux-
iliary b e (as used in 3 b ) enters the derivation with the features [Past-Tns,
u-Pers, u-Num], where Tns = tense. Let’s also make the traditional assump-
tion that T constituents carry an EPP feature. And finally, since (3 b ) is a
passive sentence, let’s take the pronoun t h e y to originate as the thematic
complement of the passive participle arrested. Given these assumptions, the
derivation will proceed as follows.
The verb arrested merges with its complement t h e y to form the VP ar-
rested t h e y . This VP is then merged with the passive auxiliary b e , forming
the T‑bar shown in simplified form in (4) below (where the auxiliary and
pronoun are shown in their dictionary citation forms b e / t h e y in order to
indicate that their morphological spellout can’t yet be determined):

(4) T′

T VP
BE
[Past-Tns] V DP
[u-Pers] arrested THEY
[u-Num] [3-Pers]
[EPP] [Pl-Num]
[u-Case]

The tense auxiliary T‑b e needs to agree with an appropriate nominal within
the structure containing it. Given the Structure Dependence Principle
of §1.3.2 (requiring syntactic operations to be sensitive to hierarchical
relations – and to the relation c‑command in particular), let’s suppose that
T agrees with a nominal which it c‑commands. Accordingly, once the struc-
ture in (4) is formed, T‑b e searches for a nominal which it c‑commands to
agree with in person and number.
Introducing some new technical terminology at this point, let us say that
by virtue of being the highest head in the overall structure (4) at this point
in the derivation, b e serves as a probe which searches for a c‑commanded
(pro)nominal goal which can value its unvalued person and number features
(where a probe is a head/X and a goal is a maximal projection/XP). It is
conventional to refer to person and number features together as φ-features
(where φ is the Greek letter phi, pronounced as [fai] in English): using this
terminology, we can say that the probe T‑b e searches for a goal which can
value its φ-features.
 3.1 Agreement, A‑Movement and Case marking 141

In addition, let’s also assume that operations like Agreement, A‑Movement

and Case-marking can only apply if both the probe and the goal are active,
and that the following conditions determine this:

(5) Activity Conditions

(i) A head can only be an active probe for Agreement or A‑Movement
or Case-marking if it carries some unvalued φ-feature/s, and becomes
inactive once its φ-feature/s have been valued (Probe Condition)
(ii) An XP can only be an active goal for Agreement, A‑Movement,
or Case-marking if it has an unvalued case feature, and becomes
inactive once its case feature is valued (Goal Condition)
(iii) If a constituent is part of a movement chain at a given point in
a derivation, only the highest copy in the chain can be active for
operations like Agreement, A‑Movement or Case-marking (Highest
Copy Condition)

Furthermore, let’s suppose that agreement and case-marking are subject to

the conditions in (6) below:

(6) Feature Valuation Conditions

(i) When an active probe like T c‑commands an active XP goal, each
unvalued φ-feature on the probe will be valued by the goal – that
is, assigned a value which is a copy of the corresponding feature
value on the goal (Agreement)
(ii) Only a φ-complete XP (i.e. one with person and number) can serve
as a goal valuing unvalued features on a probe, and conversely
only a φ-complete probe (i.e. one with person and number) can
value unvalued features on a goal (Completeness Condition)
(iii) Each feature on a constituent must be valued (i.e. assigned a single,
appropriate value) in order for the constituent to be legible at
the PF/LF interfaces; if not, the derivation will crash (Legibility
Condition)

(A structure is legible at the PF interface if it can be assigned an appro-

priate phonetic/PF representation, and legible at the LF interface if it can
be assigned an appropriate semantic/LF representation.) Finally, let’s also
suppose that the EPP feature on T works in conjunction with person agree-
ment, in accordance with the following condition:

(7) EPP Condition

An EPP feature on a probe T attracts the closest XP goal that T agrees
with in person to undergo A‑Movement to spec‑TP, and the EPP feature
is stripped from T (and removed from the derivation) once the movement
has taken place
142 Agreement

The reason for supposing that the EPP feature on T attracts an XP that it
agrees with in person is in order to ensure that T attracts a (pro)nominal
goal (and not e.g. an ADVP or PP goal) to move to spec‑TP. Of course, it fol-
lows from the Goal Condition (5ii) that the XP goal can only be attracted by
the EPP feature on T if the goal is active (by virtue of carrying an unvalued
case feature). The reason for positing that the EPP feature is stripped from
T (and thereby removed from the derivation entirely) once it has done its
work is that the EPP feature is illegible in the PF and LF components (since
EPP has no phonetic spellout and no semantic interpretation), and hence
will cause the derivation to crash if it is not removed.
The reason that I use the term stripped (meaning ‘removed from the
derivation entirely’) in (7) rather than deleted, is that deletion is classically
taken to involve giving a null phonological spellout in the PF component
to a constituent which nonetheless remains in the syntax and is assigned
a semantic interpretation in the LF/semantic component. To illustrate how
deletion works, consider the NP Deletion operation which deletes the brack-
eted NP in (8) below:

(8) He drank three cans of beer, and I drank four [NP cans of beer]

In this structure, the second (bracketed) occurrence of the NP cans of beer

can be deleted for economy reasons (to avoid redundant repetition of the
NP cans of beer). What this means is that the bracketed NP is present in the
syntax, and is inputted into both the LF component (where it is assigned a
meaning paraphrasable as ‘cans of beer’) and the PF component (where it is
given a silent spellout). However, when (7) specifies that the EPP feature is
stripped from T (once it has done its work), this means that it is eliminated
from the derivation entirely, and thus is no longer present in the syntax,
and hence plays no further role in the syntax and is not inputted into the
LF or PF components.
A further point to note is that the EPP feature on T is different in kind
from other features. For example, a feature like [Past-Tns] comprises an
attribute (= the property tense) and a value (= past). By contrast, the EPP
feature on T is an instruction specifying ‘Attract an XP goal that I agree
with in person to become my specifier’. It is arguably because it does not
have the canonical attribute-value format of other features that an EPP fea-
ture is illegible at the PF and LF interfaces, so has to be removed from the
derivation once it has triggered movement. Another reason for assuming
that EPP is removed once it has moved some constituent to spec‑TP is to
prevent it triggering a second movement which would result in T (illicitly)
ending up with two specifiers. One way of accounting for EPP not having
the attribute-value structure of other features would be to suppose that the
 3.1 Agreement, A‑Movement and Case marking 143

person feature on T has the property of attracting a goal carrying a person

feature: this could be marked (e.g.) by eliminating the EPP feature altogeth-
er, and instead asterisking the person feature on T in (4), so that T‑b e has
the features [Past-Tns, u-Pers*, u-Num], with the asterisk interpreted as an
instruction to ‘attract an XP that can value my person feature to become
my specifier’. However, I will retain the traditional EPP notation here (be-
cause it will be familiar to many readers), and take it to operate as specified
in (7) above.
We can take the conditions in (5–7) to be principles of UG/Universal
Grammar, in keeping with the goal of linguistic theory being to estab-
lish general principles governing how linguistic operations apply. In the
light of these conditions, let’s return to consider what happens once we
have formed the structure in (4) above. At this point, the T auxiliary b e
serves as a probe by virtue of being the highest head in the structure;
T‑b e is rendered active for Agreement and A‑Movement by its unval-
ued (person/number) φ-features. T‑b e searches for the closest active goal
which it c‑commands that can value its φ-features, locating t h e y (which
is active by virtue of its unvalued case feature). Let’s see how these two
operations apply.
Consider first Agreement. Since t h e y is φ-complete (i.e. it carries both
person and number), it can value the unvalued φ-features on the T‑probe
be , in conformity with the Completeness Condition (6ii). Via the Agreement
operation (6i), the values of the (third) person and (plural) number features
on the goal they are copied onto the probe b e , so that b e is thereby given
the feature values shown in italics below:

(9) T′

T VP
BE
[Past-Tns] V DP
[3-Pers] arrested THEY
[Pl-Num] [3-Pers]
[EPP] [Pl-Num]
[u-Case]
In accordance with the EPP Condition in (7), the EPP feature on the
probe b e attracts the closest XP goal c‑commanded by b e which it agrees
with in person – namely t h e y (which is active through its unvalued case
feature). This means that b e attracts (a copy of) t h e y to move to spec‑TP
in the manner shown by the arrow below (with the EPP feature thereafter
being stripped from T and removed from the derivation), resulting in
the following structure (simplified by showing the position out of which
144 Agreement

they moves as a gap, though in reality the gap position is occupied by a

null copy of t h e y ):

(10) TP

DP T′
THEY
[3-Pers] T VP
[Pl-Num] BE
[u-Case] [Past-Tns] V DP
[3-Pers] arrested —
[Pl-Num]

However, there is unfinished business in (10), because the italicised case

feature on the subject t h e y has not yet been valued – and if it doesn’t get
valued, the derivation will fall foul of the Legibility Condition (6iii) and
crash (because the PF component will be unable to determine whether to
spell out t h e y as nominative they, accusative them or genitive their). So
how does t h e y get case-marked here? This is a question addressed in the
next section.

3.1.2 Case-marking
Let’s suppose that the derivation continues by merging the TP in (10) above
with a null finite complementiser which carries an indicative mood feature
[Ind-Mood]. On the assumption that a finite (i.e. mood-specified) comple-
mentiser with a TP complement assigns nominative case to a c‑commanded
goal with an unvalued case feature, the unvalued case feature on t h e y will
be valued as nominative by the finite complementiser. However, since the
Completeness Condition (6ii) specifies that only a φ-complete probe can
value an unvalued feature on a goal, this means that Case-marking must
involve agreement in person and number between the complementiser and
the goal that it case-marks. Given this assumption, Case-marking can be
given an agreement-based account along the lines outlined informally in
(11) below:

(11) Case Conditions

A φ-complete case-assigning probe values an unvalued case feature on
an XP goal that it agrees with in respect of one or more φ-features, and
assigns the goal a case value that depends on the nature of the probe –
namely:
(i) nominative if the probe is a finite C with a TP complement (where a
finite C is one that is indicative, subjunctive or imperative in mood)
(ii) null if the probe is an intransitive infinitival C with a TP complement
(iii) accusative if the probe is transitive
 3.1 Agreement, A‑Movement and Case marking 145

(I use the term transitive to denote a probe which is an accusative

case-assigner, and intransitive to denote a probe which is not an accusative
case-assigner. Note that this is a minor departure from the traditional
assumption that an intransitive item is one with no complement: you can
use the term non-transitive instead if you prefer.) Given these assumptions,
the complementiser will enter the derivation carrying an indicative mood
feature [Ind-Mood], as well as unvalued person and number φ-features
(with unvalued features shown in italics below):

(12) CP

C TP
ø
[Ind-Mood] DP T′
[u-Pers] THEY
[u-Num] [3-Pers] T VP
[Pl-Num] BE
[u-Case] [Past-Tns] V DP
[3-Pers] arrested —
[Pl-Num]

Agreement between the complementiser probe and the DP goal t h e y will

result in the φ-features on the complementiser being valued as third per-
son plural via Agreement (6i), and the unvalued case-feature on the goal
they being valued as nominative via the Nominative Case Condition (11i),
so deriving the structure below (where the features on the complementiser
and subject which have been valued via Agreement and Case-marking are
shown in italics):

(13) CP

C TP
ø
[Ind-Mood] DP T′
[3-Pers] THEY
[Pl-Num] [3-Pers] T VP
[Pl-Num] BE
[Nom-Case] [Past-Tns] V DP
[3-Pers] arrested —
[Pl-Num]

The resulting structure (13) satisfies the requirement imposed by the

Legibility Condition (6iii) for every feature to be valued at the end of a
syntactic derivation. The structure (13) is subsequently transferred to the LF
and PF components, and PF spellout rules determine that the pronoun they
is spelled out in its nominative form they, and the auxiliary be in its third
146 Agreement

person plural past tense form were. The overall sentence is spelled out at PF
as They were arrested. (A point of detail to note is that the T auxiliary be will
also carry an indicative mood feature, perhaps copied from the head C of CP:
but I will set aside this detail here, for the sake of presentational simplicity.)
An aspect of the analysis outlined above which might at first sight
seem implausible is the postulation of an (abstract/invisible) agreement
relation between complementiser and subject. However, there are both
theoretical and empirical arguments in support of this assumption. The
theoretical argument is that the Completeness Condition (6ii) requires a
probe which values a feature on a goal to be φ-complete (i.e. to carry both
person and number). The empirical argument is that there are languages
in which a finite complementiser agrees overtly in person and number
with the subject of its clause, including varieties of Flemish (Haegeman
1992), and of German and Dutch (van Koppen 2017). This phenomenon
can be illustrated in terms of the following West Flemish data (from van
Koppen 2017):

(14) a. K peinzen da/*dan dienen student nen buot gekocht eet

I think [Link]/ that student a boat bought has
that [Link]
‘I think that that student has bought a boat’
b. K peinzen dan/*da die studenten nen buot gekocht een
I think [Link]/ those students a boat bought have
that [Link]
‘I think that those students have bought a boat’
The bold-printed complementiser in these examples (= the West Flemish
counterpart of English that) agrees in person and number with the italicised
subject of the clause it introduces, and thus is third person singular ([Link])
by agreement with the third person singular subject dienenthat studentstudent
in (14a), and third person plural ([Link]) by agreement with the third person
plural subject diethose studentenstudents in (14b). Data such as (14) thus provide
cross-linguistic evidence of the existence of complementiser-subject agree-
ment. Given the impoverished nature of English morphology, however,
complementiser-subject agreement is abstract in English, in the sense that
it has no overt phonetic reflex and so is invisible/inaudible. Still, abstract
agreement is an idea that is widely utilised in English: for example, if
we make the traditional assumption that indicative verbs/auxilaries agree
with their subjects in person and number, it follows that in the same way
as helps agrees overtly with its subject he in ‘He helps her’, so too helped
agrees (albeit abstractly/invisibly/inaudibly) with its subject in ‘He helped
her’; and in the same way as the auxiliary is agrees overtly with its subject
 3.1 Agreement, A‑Movement and Case marking 147

in ‘He is helping her’, so too the auxiliary will agrees abstractly with its
subject in ‘He will help her’.
The discussion of case-marking above is simplified insofar as the constit-
uent assigned nominative case is a single word (the pronoun/pronominal
DP they). But what if in place of they we had phrases like those italicised in
the clauses bracketed below?

(15) a. Everyone knows [that we foreigners are exploited]

b. Some people think [that all the foreign workers are exploited]
c. We feel [that some of us are exploited]

On the assumptions made here, the DP we foreigners in (15a), the QP all

the foreign workers in (15b), and the QP some of us in (15c) are all assigned
nominative case by the finite complementiser that. But precisely which
words in these phrases receive nominative case? Let’s look at each example
in turn.
If we take we foreigners in (15a) to be a DP headed by the first person
plural D/determiner we, it is clear that the head D of this DP receives nom-
inative case, since it is spelled out in the nominative form we. Since we
is the head D of the DP, and phrases are projections of their heads, it fol-
lows from theoretical considerations that the head D of a nominative DP
will carry nominative case. But does its complement (foreigners) also carry
nominative case? This is not straightforward, because nouns don’t overt-
ly inflect for nominative/accusative case in present-day English – though
nouns did inflect for case in Old English, and still do in many other lan-
guages like German and Russian. So, one possible approach would be to
suppose that nouns always carry case, but that nominative and accusative
case have no visible reflex on nouns in contemporary English.
Now consider the case of the QP all the foreign workers in (15b). Since
this QP is assigned nominative case (as we see from the possibility of
substituting the QP by the nominative pronoun they), it follows from the-
oretical considerations that its head quantifier all must carry nominative
case (albeit this is invisible). But does the determiner the also carry (in-
visible) nominative case? Since the determiner we in (15a) carries nomi-
native case, the most consistent assumption would be that all determiners
carry case in English, even though (because of the impoverished nature
of English case morphology) this is nearly always invisible. But what
of the noun workers? Since I argued for treating the noun foreigners in
(15a) as carrying nominative case, the most consistent assumption would
be to suppose that all nouns carry case – and hence that workers carries
an invisible nominative case feature in (15b). And what of the adjective
foreign? In a language like Russian (with a richer case morphology than
148 Agreement

English), both adjectives and nouns overtly inflect for case – as illustrated
by the following example:

(16) Krasivaya dyevushka vsunula chornuyu koshku v pustuyu korobku

Beautiful girl put black cat in empty box
‘The beautiful girl put the black cat in the empty box’

The nouns and adjectives in (16) carry (italicised) case endings (-a is a
nominative suffix, and -u an accusative suffix). So (on one view), the
broader conclusion that we might draw from our overall discussion here
is that crosslinguistic evidence makes it plausible to suppose that nouns
and their modifiers (determiners, quantifiers, adjectives) carry case, albeit
the relevant case features are no longer directly visible in present-day
English.
Now consider the QP some of us in (15c). This QP is assigned nomina-
tive case (and so can be substituted by we), and hence its head quantifier
some will be nominative. However, it is clearly not the case that all three
words in the QP some of us are nominative: after all, the pronoun us has
accusative case, and the case particle of (like prepositions) is not the kind
of word which inflects for case (even in a language like Russian with a
richer case morphology – as we see from the Russian preposition vin being
uninflected in 16 above). How can a nominative QP like some of us in (15)
contain an accusative pronoun? To try and understand what’s going on
here, consider the structure of the bracketed clause ‘that some of us are
exploited’ in (15c), which is as below (simplified by not showing the in-
ternal structure of the T‑bar are exploited, because this is not relevant to
our discussion of case here):

(17) CP

C TP
that
QP T′
are exploited
Q KP
some
K DP
of us
Let’s suppose that (in consequence of the Minimality Condition introduced in
§1.3.2), constituents are case-marked by the minimal/closest case-assigner
above (i.e. c‑commanding) them. What this means is that the pronoun/
pronominal DP us will be assigned accusative case by the transitive case
particle of in accordance with the case condition (11iii) above, since of is
the closest case-assigner c‑commanding us. By contrast, the closest case
 3.1 Agreement, A‑Movement and Case marking 149

assigner c‑commanding the quantifier some in (17) is the complementiser

that, and since finite complementisers assign nominative case via (11i), this
means that some will be assigned nominative case. Since case assigners
typically don’t themselves carry case, this means that both the case particle
of and the complementiser that will be caseless. Thus, in the nominative QP
some of us, the head quantifier some carries (invisible) nominative case, the
case particle of is caseless, and the pronoun us is accusative. So, an import-
ant conclusion to draw from our discussion here is that when a complex DP
or QP is assigned (say) nominative case, this does not mean that every one
of the subconstituents in that DP or QP carries nominative case. In order
to simplify our subsequent discussion of case-marking, I shall henceforth
simply talk about the case assigned to a given DP or QP, and not be con-
cerned with the case assigned to individual subconstituents of those DPs or
QPs which have a complex internal structure.

3.1.3 Valued and unvalued features

Our discussion of how agreement and case-marking work in a sentence
such as (3 b ) They were arrested has wider implications. One of these is
that items enter the derivation with some of their features already valued
and others unvalued: for example, t h e y enters the derivation in (4) with
its (third) person and (plural) number features valued, but its case feature
unvalued. This raises the question of which features are initially valued
when they first enter the derivation, which are initially unvalued –
and why.
Chomsky (1998, 1999) argued that the difference between valued and
unvalued grammatical features correlates with a related distinction be-
tween those grammatical features which are interpretable (in the sense that
they play a role in semantic interpretation) and those which are uninter-
pretable (and hence play no role in semantic interpretation). For example, it
seems clear that the case feature of a pronoun like t h e y is uninterpretable,
since a subject pronoun surfaces as nominative, accusative or genitive de-
pending on the type of [bracketed] clause it occurs in, without any effect
on meaning – as the examples below illustrate:

(18) a. It seems [they were arrested]

b. He expected [them to be arrested]
c. He was shocked at [their being arrested]

By contrast, the (person/number/gender) φ-features of pronouns are inter-

pretable, since for example, a first person singular pronoun like I clearly
differs in meaning from a third person plural pronoun like they, and a
masculine pronoun like he differs in meaning from a feminine pronoun
like she.
150 Agreement

Now consider the interpretable and uninterpretable features carried by

finite auxiliaries like those italicised below:

(19) a. She is working/She was working

b. He is writing the assignment/He has written the assignment
c. He insists that she is respected/He insists that she be respected
d. They are working/*They am working

The tense feature on the present/past tense auxiliary is/was in (19a) deter-
mines whether the sentence is interpreted as describing a present or past
state of affairs, and so tense is clearly an interpretable feature. Likewise, the
choice between the progressive aspect auxiliary is and the perfect aspect
auxiliary has in (19b) determines whether the sentence is interpreted as
describing an action which is in progress or one which is completed, and
so aspect is also an interpretable feature of the auxiliary. In much the same
way, the choice between the indicative mood auxiliary is and its subjunct-
ive mood counterpart be in (19c) determines whether the sentence is inter-
preted as describing a realis (i.e. existing) state of affairs or an irrealis one
(i.e. one which does not exist at present but which may exist in the future),
and so mood must also be an interpretable feature of auxiliaries.
By contrast, the person/number φ-features of auxiliaries are uninterpret-
able, in that they serve purely to mark agreement with a particular (pro)-
nominal: consequently, if we replace the third person plural auxiliary are
by the first person singular form am in a sentence like (19d) which has a
third person plural subject, we do not change the meaning of the sentence
but rather simply make the resulting sentence ungrammatical (*They am
working).
We can summarise the picture which emerges from the discussion above
in terms of the schema in (20) below:

(20) Interpretable and uninterpretable features

constituent interpretable features include uninterpretable
features include
auxi iary tense, aspect and mood person and number
( pro ) no ina person, number and gender case

Each structure generated by the syntactic component of the grammar is

subsequently transferred to the PF component of the grammar to be spelled
out (i.e. assigned a PF representation which provides a representation of its
phonetic form). Since unvalued features are illegible to (i.e. cannot be pro-
cessed by) the PF and LF components, it follows that every feature which
enters the derivation unvalued must be valued in the syntax, or else the
derivation will crash (i.e. fail) at the PF interface because the PF compo-
nent is unable to spell out unvalued features: in the words of Chomsky
 3.1 Agreement, A‑Movement and Case marking 151

(2007: 13) ‘If transferred to the interface unvalued, uninterpretable features

will cause the derivation to crash’ – an idea embodied in the Legibility
Condition (6iii). In more concrete terms, this amounts to saying that unless
the syntax specifies whether we require (for example) a first person singu-
lar or third person plural present tense form of b e , the derivation will crash
at the phonetics interface because the PF component cannot determine
whether to spell out b e as am or are.
In addition to being sent to the PF component, each structure generated
by the syntactic component of the grammar is simultaneously transferred
to the semantic component, where it is assigned an appropriate seman-
tic representation. Clearly, interpretable features play an important role
in computing semantic representations. Equally clearly, however, uninter-
pretable features play no role in this process; and since they are illegible
to the semantic component, it is clear that uninterpretable features are not
processed in the semantic component. Nonetheless, we need to ensure that
uninterpretable features are processed in the PF component (since e.g. PF
needs to know whether a third person masculine singular pronoun has been
assigned nominative, accusative or genitive case in order to know whether
to spell it out as he, him or his). Thus, the overall conclusion to draw from
this is: all features (both interpretable and uninterpretable) are processed
in the PF/phonological component, but only interpretable features are pro-
cessed in the LF/semantic component.
However, Chomsky’s idea that features entering the derivation valued are
interpretable and those entering the derivation unvalued are uninterpreta-
ble proves over-simplistic in certain respects. For example, this assumption
is problematic for languages which have so-called ‘arbitrary gender’ – that
is, languages in which nouns may carry a gender feature that is unpredict-
able (e.g. not correlated with their meaning). For example, the word for sea
in Italian (mare) is masculine in gender, but there is no obvious rationale
for this, since its counterpart in French (mer) is feminine. There are two
words for table in Italian, one of which is masculine in gender (tavolo)
and the other feminine (tavola). The Italian word paio ‘pair’ is masculine
in gender in its singular form, but feminine in its plural form (paia). The
German word Mädchen ‘girl’ is neuter in gender, in spite of denoting a
female human being. Such examples suggest that (rather than assuming
that gender is an interpretable feature), it is more appropriate to suppose
that gender is an inherent lexical feature of nouns/pronouns (specified in
their lexical/dictionary entry), and that inherent features enter the deriva-
tion valued. On this view the person/number/gender features of a personal
pronoun like he are inherent features of the pronoun (in the sense that the
pronoun carries these features irrespective of the type of structure in which
it occurs), whereas its case feature (which determines whether the pronoun
152 Agreement

is spelled out in the nominative form he, the accusative form him or the
genitive form his) is a contextual feature – that is, one determined by the
syntactic context (i.e. the type of syntactic structure) in which the word
occurs. Contextual features enter the derivation unvalued, and are assigned
a value via operations like Agreement and Case-marking in the syntax.
(For a theoretical discussion of grammatical features, and views on the re-
lation between valued/unvalued features and interpretable/uninterpretable
features, see Pesetsky & Torrego 2007, Willim 2012, and Svenonius 2017.)
To summarise: we have seen in Module 3.1 how (subject-verb) Agreement
involves a relation between a finite T‑probe and a nominal or pronominal
XP goal which it c‑commands: the T‑probe is made active by unvalued
(person/number) φ-features which it carries when it enters the derivation,
and its (pro)nominal goal is made active by its unvalued case feature. More
generally, the probe and goal enter the derivation carrying a set of valued
inherent features, and a set of unvalued contextual features which are val-
ued by operations like Agreement and Case-marking in the course of the
derivation. We have also seen that Case-marking is contingent on a person/
number agreement relation between probe and goal, and that A‑Movement
involves T attracting the closest goal it agrees with in person. At the end
of the syntactic derivation, both interpretable and uninterpretable features
are processed in the PF component, but only interpretable features are pro-
cessed in the LF/semantic component.
You should now be able to tackle Exercise 3.1.

3.2 Agreement and case in expletive clauses

So far, all the sentences we have looked at have involved agreement between
a (T or C) probe and a thematic nominal or pronominal DP or QP. However,
English has two expletive pronouns (it and there) which differ from other
pronouns in that they are non-thematic (in the sense that they cannot
occupy a position to which a theta role is assigned) and non-referential (in
that they cannot refer back to some other expression, they cannot be sub-
stituted by referential pronouns like this/that, and they cannot have their
reference questioned by interrogative pronouns like what/where). In this
module, we look at how agreement works in expletive clauses.

3.2.1 Expletive it clauses

Typical examples of sentences with an expletive it subject are given below:

(21) a. It has transpired that he was cheating

b. It does not matter that the economy is shrinking
 3.2 Agreement and case in expletive clauses 153

c. It can be difficult to cope with long-term illness

d. It was unfortunate that she couldn’t attend the meeting
e. It was claimed that Boris knew nothing about Partygate

The pronoun it in sentences like these appears to be an expletive, since it

cannot be substituted by a referential pronoun like this or that, and cannot
be questioned by what. Let’s take a closer look at the syntax of expletive it,
beginning by looking at the features that it carries.
Since expletive it is intrinsically a third person singular pronoun, let
us suppose that it enters the derivation carrying the inherent φ-features
[3‑Pers, Sg‑Num]. An interesting question which arises in this regard is
whether the person/number features on expletive it are interpretable. Some
evidence that they may be comes from the observation that expletive it can
serve as the controller/antecedent for PRO in sentences such as:

(22) a. It can sometimes rain after [PRO snowing] (cf. Chomsky 1981: 324)
b. It can be difficult to overcome depression without [PRO being
impossible]

Since the semantic component needs to identify expletive it as the antecedent

of PRO in such cases, it needs to be able to ‘see’ the person/number features of
it in order to determine whether they match the features of PRO. This means
that the person/number features of expletive it must be visible at the seman-
tics interface, and hence they must be interpretable. Chomsky (1981: 325) sug-
gests that expletive it is a quasi-argument (in that it exhibits some but not all
of the properties of true arguments). For example, it has the argumental prop-
erty of being able to occupy an A‑position like spec‑TP (a position which only
arguments can occupy), and being able to serve as the antecedent of PRO.
Now consider the issue of whether expletive it carries a gender feature.
It is plausible to suppose that the pronoun it carries an interpretable (neu-
ter/inanimate) gender feature when used as a referential pronoun (e.g. in a
sentence like The book has lots of exercises in it, where it refers back to the
book). However, I shall assume that in its use as an expletive pronoun, it
is genderless and so carries no interpretable gender feature (which may be
why it can’t serve as the antecedent of the floating neuter anaphor itself in
*‘It can itself be hard to master syntax’).
Finally, consider whether expletive it has case. An interesting point to
note in this connection is that expletive it can undergo Raising/A‑Move-
ment in sentences such as:

(23) It does not seem to matter that the economy is shrinking

Given the VP-Internal Subject Hypothesis, expletive it will originate as

the subject of the verb matter and from there raise up to become first the
154 Agreement

subject of infinitival to and then the subject of does. However, the Goal
Condition (5ii) specifies that a pronoun can only be an active goal for
A‑Movement if it has an unvalued case feature, so it follows that expletive
it must enter the derivation with an unvalued case feature. The case feature
on expletive it will be valued as nominative in a sentence like (24a) below,
and accusative in a sentence like (24b):

(24) a. It is irresponsible to behave like that

b. I consider it to be irresponsible to behave like that

In addition, we find expletive it used in the genitive form its as the subject
of a gerund (like being) in sentences such as the following:

(25) a. In the event of its being impossible to complete the work of removal
and neutralization in time, the party concerned shall mark the
spot by placing visible signs there (Agreement on the Cessation of
Hostilities in Laos, 1954)
b. But it was understood that such preparations would now be made
as would render it possible to give effect to the blockade measures
proposed, in the event of its being necessary to take such action
(Papers from the Paris Peace Conference, 1919)
c. In 1856 a Committee of the other House was appointed in
consequence of its being necessary to renew the Act … (hansard.
[Link])
d. In 1936 Edward VIII abdicated, and spoke on the radio of its being
impossible ‘to discharge my duties as King … without the help and
support of the woman I love’ ([Link])

This suggests that expletive it carries a case feature; on the assumptions

made here, it will enter the derivation unvalued for case.
If the reasoning in the paragraphs above is along the right lines, exple-
tive it enters the derivation carrying the features [3‑Pers, Sg‑Num, u‑Case].
Given the VP-Internal Subject Hypothesis, we can suppose that (like other
subects) expletive it originates internally within VP (e.g. as the complement
of an unaccusative or passive verb, but as the specifier of other types of
verb). To see what all of this means in more concrete terms, let’s look at the
derivation of the it-clause in (21a) It has transpired that he was cheating.
This proceeds as follows.
The verb transpired merges with its CP complement that he was cheat-
ing to form the V‑bar transpired that he was cheating. This V‑bar is then
merged with the quasi-argument it to form the VP it transpired that he was
cheating. The resulting VP is in turn merged with the T auxiliary h av e , so
forming the T‑bar below (where only features on constituents of immediate
 3.2 Agreement and case in expletive clauses 155

interest are shown, with uninterpretable features being in italics, and inter-
pretable ones in non-italic print):

(26) T′

T VP
HAVE
[u-Pers] DP V′
[u-Num] IT
[EPP] [3-Pers] V CP
[Sg-Num] transpired that he was
[u-Case] cheating
For the reasons given in §3.1.3, I have assumed in (26) that the finite T
auxiliary have enters the derivation carrying an interpretable present-tense
feature, unvalued (and uninterpretable) person/number φ-features, and an
EPP feature which is illegible at the PF and LF interfaces because it has no
form or meaning.
The auxiliary h av e is a probe which is active for Agreement and
A‑Movement by virtue of its unvalued person and number features. Ac-
cordingly, have searches for a c‑commanded goal which can value its per-
son and number features, locating the expletive pronoun i t (which is active
for Agreement and A‑Movement by virtue of its unvalued case feature).
Via Agreement (6i), the goal i t values the unvalued person/number fea-
tures on the probe h av e as third person singular. Since the goal i t carries a
complete set of (person and number) φ-features, there is no violation of the
Completeness Condition (6ii).
In addition, in accordance with the EPP Condition (7) above, the EPP
feature on the T auxiliary h av e in (26) attracts the closest XP that it agrees
with in person (namely the pronominal DP i t ) to move to spec‑TP, and
the EPP feature (being illegible at both PF and LF interfaces) is thereafter
stripped from T (and thereby removed from the derivation), resulting in the
structure below (with the arrow showing A‑Movement of i t ):

(27) TP

DP T′
IT
[3-Pers] T VP
[Sg-Num] HAVE
[u-Case] [Pres-Tns] DP V′
[3-Pers] — transpired that
[Sg-Num] he was cheating
156 Agreement

This TP is then merged with a null complementiser carrying an interpret-

able indicative-mood feature, and uninterpretable (and unvalued) person
and number φ-features, so forming the CP below (with uninterpretable fea-
tures italicised):

(28) CP

C TP
ø
[Ind-Mood] DP T′
[u-Pers] IT
[u-Num] [3-Pers] T VP
[Sg-Num] HAVE
[u-Case] [Pres-Tns] DP V′
[3-Pers] — transpired that
[Sg-Num] he was cheating
The null complementiser in (28) serves as a probe (active by virtue of its
unvalued person and number φ-features) and searches for an XP goal which
can value these, locating the pronominal DP i t (which is active by virtue of
its unvalued case feature, and φ-complete by virtue of its person and num-
ber features). Via the Agreement operation (6i), the unvalued φ-features on
the null complementiser are valued as third person singular by agreement
with it . At the same time, the complementiser (being finite and φ-complete)
values the unvalued case feature on i t as nominative via the Nominative
Case Condition (11i). Application of Agreement and Case-marking map (28)
above into (29) below:

(29) CP

C TP
ø
[Ind-Mood] DP T′
[3-Pers] IT
[Sg-Num] [3-Pers] T VP
[Sg-Num] HAVE
[Nom-Case] [Pres-Tns] DP V′
[3-Pers] — transpired that
[Sg-Num] he was cheating
The tree (29) meets the requirement imposed by the Legibility Condition
(6iii) for all features to be valued at the end of a syntactic derivation.
The structure (29) then undergoes transfer to the PF and LF components.
The PF component processes all the features in (29), and thus spells out
have as the third person singular present tense form has, and spells out
expletive i t in its nominative form it. The italicised uninterpretable features
are (by their nature) not processed in the semantic component, so that only
 3.2 Agreement and case in expletive clauses 157

the non-italic interpretable features are assigned an interpretation at the

semantics interface.
I will briefly conclude this section by addressing four issues raised by an
anonymous reviewer in relation to the analyses sketched here. One is why
expletive it should be taken to originate inside VP, rather than being direct-
ly merged in spec‑TP in order to satisfy the EPP feature on T. One answer
lies in Chomsky’s treatment of expletive it as a quasi-argument: given that
(under the VP‑internal Subject Hypothesis) arguments originate in a pos-
ition where they are first- or second-merged with a predicate, it is plausible
to suppose that the same is true of quasi-arguments. A second answer is
that, given the account of agreement outlined here under which a finite T
constituent agrees with a (pro)nominal that it c‑commands, it follows that
in a structure like (26) above, expletive it must originate in a (VP‑internal)
position below T‑have in order for have to agree with it.
A second (related) question asked by the reviewer is what theta role ex-
pletive it is assigned if it originates in an argument position within VP.
One possibility is that it has the thematic role of e x p e t i ve argument of the
predicate it merges with (though of course, this is merely an ad hoc stipula-
tion). Another is to suppose that a predicate like transpire has the property
that (when used with a CP complement) it assigns the theta role t h e e to
its complement, but assigns no theta role to its highest argument (i.e. to its
subject/specifier); and we will explore another type of structure in which a
verb does not theta-mark its subject in §3.4.3.
A third question asked by the reviewer is whether, if expletive it lacks
gender it won’t be φ-complete, and the Completeness Condition (6ii) will
prevent expletive it from triggering agreement on an auxiliary like has
in a sentence like (21a) ‘It has transpired that he was cheating’. However,
note that (6ii) defines a φ-complete constituent as ‘one with person and
number’, and does not mention gender at all: this is because finite verbs/
auxiliaries agree with their subjects in person and number only in English,
not in gender. So, being φ-complete in a given language means possessing
all the features which are involved in agreement in the relevant language:
and for English, this means person and number (but not gender). Thus, if
(as claimed here) expletive it has person and number (but not gender), this
will not prevent it from participating in agreement operations.
A fourth question raised by the reviewer is why (under the formulation
of the EPP Condition in 7), EPP should be taken to involve T attracting
the closest active goal that it agrees with in person. One reason is that this
accounts for why T attracts (pro)nominal constituents to become its spe-
cifier, since these have person (whereas APs, PPs, ADVPs etc. do not carry
person, and so cannot be attracted by T). Another answer is that EPP can
attract expletive there to move to spec‑TP, and this is a defective pronoun
158 Agreement

which carries only person (not number or gender) – as we will now see in
the next section.

3.2.2 Expletive there clauses

Typical examples of sentences with expletive there subjects are given below:
(30) a. There have been several clashes with the police
b. There exists a passion for comprehension, just as there exists a
passion for music ([Link])
c. There remain several unresolved issues
d. There persists a widespread assumption that all Africans are
‘indigenous’ (Huffington Post)
e. There endures a disparity in budget allocation between offline and
online worlds ([Link])
f. Deep in the heart of the Derbyshire dales, there survives a medieval
sport where almost anything goes except murder ([Link])

In such sentences, expletive there is said to be existential in use, since it

occurs as the subject of an existential predicate like be/exist/remain/persist/
endure/survive (remain/persist/endure/survive being existential predicates
in that they have a meaning loosely paraphrasable as ‘continue to exist’).
There are other uses of there (e.g. locative and presentational), but I will
set these aside here and concentrate on existential uses of expletive there.
Consider first the features carried by expletive there. An interesting prop-
erty illustrated below is that expletive there can only occur with a third
person verb (like is in 31a below), not with a first person verb (like am in
31b), nor with a second person verb (like are in 31c):
(31) a. There is only me/you taking the Fantasy Syntax course
b. *There am only I taking the Fantasy Syntax course
c. *There are only you taking the Fantasy Syntax course

This suggests that expletive there is intrinsically third person and so carries
an inherent [3‑Pers] feature. By contrast, expletive there can freely occur
both with a singular verb (like is in 32a below) and with a plural verb (like
are in 32b):
(32) a. There is only one student taking the Fantasy Syntax course
b. There are only two students taking the Fantasy Syntax course

This suggests that expletive there does not have an inherent number feature.
Some support for the claim that expletive there is numberless comes from
the observation that (unlike expletive it: see the examples in 22 above), it
cannot serve as the antecedent of PRO, for example, in a sentence like:
(33) There can’t be a divorce without [there/*PRO being a marriage first]
 3.2 Agreement and case in expletive clauses 159

If PRO requires an antecedent with person and number features, we can

account for why expletive there cannot serve as the antecedent of PRO by
supposing that there carries person but not number.
But what about gender? Since expletive there never has an antecedent,
it is reasonable to conclude that it does not carry a gender feature either.
By contrast, expletive there seems to enter the derivation with an unvalued
case feature which makes it active for undergoing Raising/A‑Movement in
a structure such as the following (where there originates as the subject of
been and raises up to become the subject of T‑to and then of T‑do):

(34) There do seem [there to have there been several clashes with the police]

The unvalued case feature on there will be valued as nominative in a sen-

tence like (35a) below, and as accusative in a sentence like (35b):

(35) a. There is no good excuse for bad behaviour

b. I consider there to be no good excuse for bad behaviour

And indeed for some speakers, expletive there has the genitive form there’s –
as shown by the internet-sourced examples below, where the genitive exple-
tive there’s is used as the subject of the gerund being:

(36) a. Squidward was ranting about there’s being no culture in Bikini

Bottom … ([Link])
b. People always complain about there’s being too many swordmen on
smash, and Sothis also said this ([Link])
c. I’d been singing (in my horrible voice) the song about there’s being
‘no way to stop it’, a song not used in the movie ([Link])
d. There’s being no other business, the meeting was adjourned by the
chairman at 18:30 ([Link])
e. Whites how do you feel about there’s being a WHITE terrorist
that has killed lot of people because he’s simply a bigoted racist?
([Link])
f. You started off very aggressively in multiple posts in the beginning
of the thread about there’s being nothing OP can do ([Link])

This suggests that (like expletive it), expletive there carries case. (People
who don’t use the genitive form there’s may well treat there as defective in
the same way as other pronouns which have no genitive form – e.g. this/
that/what.)
The considerations outlined above suggest that expletive there enters the
derivation with an inherently valued [3‑Pers] feature, and with an unval-
ued case feature [u-Case]. In the light of this assumption, let’s consider the
derivation of our earlier sentence (30a) There have been several clashes with
the police. This will proceed as follows.
160 Agreement

The verb been merges with the QP complement several clashes with the
police to form the V‑bar been several clashes with the police. This V‑bar is
then merged with the expletive pronoun there to form the VP there been
several clashes with the police. Given our conclusion in the previous para-
graph that expletive there enters the derivation with the features [3‑Pers,
u‑Case], the VP headed by the verb been will have the following structure
(if we set aside the features on been):

(37) VP

DP V′
there
[3-Pers] V QP
[u-Case] been several clashes
with the police
[3-Pers]
[Pl-Num]
[u-Case]
The VP in (37) is merged with the T auxiliary h av e (which carries an inter-
pretable present-tense feature, along with uninterpretable and unvalued
person and number features, and an EPP feature), so deriving (38) below:

(38) T′

T VP
HAVE
[Pres-Tns] DP V′
[u-Pers] there
[u-Num] [3-Pers] V QP
[EPP] [u-Case] been several clashes
with the police
[3-Pers]
[Pl-Num]
[u-Case]
The T auxiliary have is active as a probe for Agreement and A‑Movement at
this point (by virtue of its unvalued φ-features) and so have searches for goals
which can value its unvalued person and number features. Let us suppose that
(in consequence of the Minimality Condition) an agreement probe searches for
the minimal (i.e. closest and smallest) set of one or more active goals that can
value all its unvalued agreement features. So, a probe first locates the closest
goal that it c‑commands, and then (if that goal is unable to value all of the
probe’s unvalued features), the probe then locates the next closest goal … and
so on, until the point where all unvalued features of the probe have been val-
ued (at which point the probe stops searching for further goals).
 3.2 Agreement and case in expletive clauses 161

Accordingly, when T‑h av e probes in (38) above, it locates there as the

closest goal (with there being rendered active by its unvalued case feature).
However, there (being numberless) cannot value the unvalued number fea-
ture on be ; indeed, there cannot value the person feature on h av e either,
since there is a defective goal (by virtue of lacking number) and it follows
from the Completeness Condition (6ii) that only a complete goal (i.e. one
carrying both person and number) can value a feature on a probe.
Consequently, T‑h av e continues to probe and locates the QP several
clashes with the police as a second goal. Thus, T‑h av e locates two active
goals: one is there (which is active by virtue of its unvalued case feature);
the other is the QP several clashes with the police (which is likewise active
by virtue of its unvalued case feature). When a probe locates more than one
active goal, it undergoes simultaneous multiple agreement with all active
goals accessible to it. Accordingly, via the Agreement operation (6i), the
T‑probe have agrees in person with both goals (the third-person constitu-
ents there and several clashes with the police) and the unvalued person fea-
ture on have is thereby valued as third person (i.e. h av e is assigned a value
for its person feature which matches that of both goals). However, h av e also
agrees in number with the plural goal several clashes with the police, and
so the unvalued number feature on h av e is valued as plural.
In addition, in accordance with the EPP Condition (7), T‑h av e attracts
the closest goal that it agrees with in person (namely expletive there, active
through its unvalued case feature) to move to spec‑TP (and the EPP feature
on T‑have is thereafter stripped/removed from the derivation). Consequent-
ly, application of Agreement and A‑Movement (arrowed below) derives the
following structure:

(39) TP

DP T′
there
[3-Pers] T VP
[u-Case] HAVE
[Pres-Tns] DP V′
[3-Pers] —
[Pl-Num] V QP
been several clashes
with the police
[3-Pers]
[Pl-Num]
[u-Case]

The resulting TP is then merged with a null complementiser which enters

the derivation carrying an interpretable indicative-mood feature, and
162 Agreement

uninterpretable (and unvalued) person and number features, so forming

the CP in (40) below:

(40) CP

C TP
ø
[Ind-Mood] DP T′
[u-Pers] there
[u-Num] [3-Pers] T VP
[u-Case] HAVE
[Pres-Tns] DP V′
[3-Pers] —
[Pl-Num] V QP
been several clashes
with the police
[3-Pers]
[Pl-Num]
[u-Case]

The indicative complementiser in (40) serves as a probe (active by virtue of

its unvalued person and number φ-features) and searches for active goals
which can value these φ-features, locating the pronominal DP t h e re and
the QP several clashes with the police. The unvalued φ-features on the
null complementiser are valued as third person plural via multiple agree-
ment with t h e re and several clashes with the police. At the same time,
the unvalued case features on DP t h e re and QP several clashes with the
police are both valued as nominative via the Nominative Case Condition
(11i). Accordingly, application of Agreement and Case-assignment maps
(40) above into the structure in (41) below:

(41) CP

C TP
ø
[Ind-Mood] DP T′
[3-Pers] there
[Pl-Num] [3-Pers] T VP
[Nom-Case] HAVE
[Pres-Tns] DP V′
[3-Pers] —
V QP
[Pl-Num]
been several clashes
with the police
[3-Pers]
[Pl-Num]
[Nom-Case]
 3.2 Agreement and case in expletive clauses 163

Since all features are valued at the end of the syntactic derivation, the
Legibility Condition (6iii) is satisfied.
The structure (41) then undergoes transfer to the PF and LF components.
The auxiliary hav e is spelled out at PF in its third person plural present
tense form have, and expletive t h e re in its nominative form there (rather
than e.g. in its genitive form there’s). Likewise, the QP is spelled out in its
nominative form several clashes with the police. Accordingly, the overall
structure in (41) is ultimately spelled out as the PF string (30a) There have
been several clashes with the police. When the structure (41) is transferred
to the semantic component, only the non-italic interpretable features are
processed there.
The discussion above highlights a further interesting property of exple-
tive there which differentiates it from expletive it. This is that expletive
there must always be associated with another goal (like the QP several
clashes with the police above) which can value the number feature on T
(and on C). This additional goal is referred to as the associate of the ex-
pletive (e.g. several clashes with the police is the associate of there in 41).
By contrast, expletive it does not require an associate (as we see from sen-
tences like It is raining) precisely because it has a full set of person and
number features of its own.
Our discussion of agreement in (41) above claims that the null com-
plementiser assigns nominative case to the QP goal several clashes with
the police. But an important question of detail which this claim raises is
which of the words in this QP carries nominative case – an issue discussed
earlier in relation to the examples in (15) above. Let’s suppose that (as in
languages like Latin with a richer system of overt case marking), nouns
and the determiners and quantifiers which modify them carry case, and
hence all enter the derivation carrying an unvalued case feature. Let’s also
suppose that (in consequence of the Minimality Condition), a constituent
with an unvalued case feature is case-marked by the closest case assigner
above (i.e. c‑commanding) it. For the words several and clashes, the closest
case-assigner above them in (41) is the null complementiser C‑ø at the top
of the tree; hence (since C‑ø is indicative in mood and thus a nominative
case assigner), several and clashes are assigned nominative case. By con-
trast, the closest case assigner above the and police is the transitive prep-
osition with; hence the and police are assigned accusative case by with.
Thus, in saying that the QP goal several clashes with the police is assigned
nominative case, what we really mean is that case features on several and
clashes are valued as nominative: the preposition with assigns (but does
not itself have) case.
Our assumption that T in (41) undergoes multiple agreement with both
the expletive goal there and the non-expletive goal several prizes and that
164 Agreement

there is intrinsically third person carries with it the tacit implication that
the associate of the expletive must likewise be a third person expression,
since a probe cannot agree with multiple goals which have different per-
son properties (because this will mean that the probe cannot be assigned a
person value which matches that carried by each of its goals – that is, there
will be feature mismatch). Such an assumption accounts for contrasts like:

(42) a. Only I am representing the club

b. *There am/is only I representing the club
c. There is only me representing the club

In (42a), the only goal for the T‑probe b e is the (italicised) first person sin-
gular expression only I, and consequently b e agrees with this (single) goal
and is marked as first person singular. But in (42b), b e has two (italicised)
goals which it must agree with – namely the third person expletive there
and its first person singular associate only I. If b e agrees in person and
number with the associate only I, it will be marked as first person singular
and ultimately be spelled out as am: but this will mean that am does not
agree in person with the third person expletive there, causing the deriv-
ation to crash. Conversely, if b e agrees in person with there and in number
with the associate only I, it will be marked as third person singular and
ultimately be spelled out as is: but this will mean that is does not agree
in person with only I. Because of the resulting feature mismatch, (42b) is
ungrammatical.
What is at first sight puzzling is the pattern of agreement in (42c), where
use of the third person form is suggests that both of its goals (there and
only me) must be third person. How can this be? The answer would appear
to be that only me has the property of being able to function as a third per-
son singular expression, as we see from third person singular agreement
on the bold-printed verb/auxiliary in (internet-sourced) sentences such as
the following:

(43) a. Now only me is working on it ([Link])

b. Only me is capable of breaking a spoon on my second spoonful of
sonic ice cream! (Kylie Morton on Twitter)
c. So at the moment, only me has access to normally blocked sites
([Link])
d. Only me knows why (Title of a song by Harold Walden)
e. Only me doesn’t have kids ([Link])

How can we account for this? A plausible answer is that only me has a more
abstract structure akin to that of a QP/quantifier phrase like nobody but me
(where but functions as a preposition like except, according to Householder
1987). This would account for third person agreement, because nobody
 3.2 Agreement and case in expletive clauses 165

requires third person singular agreement (e.g. ‘Nobody is that stupid’). It

would also account for the use of the accusative pronoun me, if me is the
complement of an invisible transitive preposition like but/except (i.e. if only
me can have a structure akin to that of ‘nobody but/except me’).
Support for this analysis of only-structures comes from the singular
agreement found in sentences such as the following where only has a plu-
ral pronoun after it:

(44) a. I thought there was only us (L. Lowry, The Giver, Bantam, New York,
1993)
b. And then there was only us (Title of book by K. Buff, Polar Bear
Books, 2015)
c. When there was only us (Title of song by Aether, on Spotify)
d. I have no friends, there is only you two (M. Clark, 2000, Terrorists,
Google Books)
e. As you watch, there is only them, and you are transfixed ([Link])

If we analyse only us/only you two/only them in (44) as having a more abstract
structure paraphrasable as ‘nobody except us/you two/them’, we can account
for singular agreement on the verbs was and is in structures like these by
positing that the head of the structure is a singular pronoun like nobody. Of
course, as an anonymous reviewer rightly points out, such an analysis raises
the question of how ‘nobody except’ comes to be spelled out as only.

3.2.3 Conditions on the use of expletives

An important question to ask in the context of our discussion of expletive
it in §3.2.1 and expletive there in §3.2.2 is what conditions govern the use
of expletives. What I shall suggest here is that expletives are subject to the
following conditions:

(45) Expletive Conditions

(i) External Argument Condition
An expletive can only be merged as the highest argument of a verb with
no external argument

(ii) Indefiniteness Condition

Existential there can only be merged as the specifier of an existential
verb with an indefinite (pro)nominal complement

(iii) Inactivity Condition

Expletive it can only be merged with a constituent which does not
contain an active nominal or pronominal (i.e. one carrying an unvalued
case feature)

Let’s briefly illustrate how these conditions work.

166 Agreement

Consider first the External Argument Condition (45i). If a verb phrase

can only have one external argument/specifier (either a thematic argument,
or an expletive quasi-argument) and if the italicised external argument of
the verb in a transitive structure like (46a) below originates as the specifier
of the verb, then it follows that the verb cannot also have an expletive ex-
ternal (quasi-) argument like that bold-printed in (46b):

(46) a. A spokeman for the president has denied allegations of impropriety

b. *There has a spokesman for the president denied allegations of
impropriety

If an expletive is ‘the last argument to be added’ in a derivation (Felser & Rupp

2001: 314), it follows that (in a sentence like There occurred several unfortunate
incidents), the QP several unfortunate incidents will be the first argument to be
merged with the verb occur (as its complement), and expletive there the sec-
ond/last argument (as the specifier of the VP headed by the verb occur).
Note that the condition that an expletive is merged as the highest argu-
ment of a predicate does not imply that it should always be merged as the
specifier of the predicate. This is because, in cases where the expletive is the
only argument of a predicate it will initially be merged as the complement
of the predicate. A case in point is ‘weather it’ in sentences such as:

(47) It has been raining/snowing all night long

Here, expletive it is the only argument of the unaccusative predicate rain/

snow, and thus it will initially be merged in comp‑VP as the complement
of the verb rain/snow, before raising up to spec‑TP to become the subject/
specifier of the T auxiliary has. Cross-linguistic evidence that rain/snow
are unaccusative verbs comes from the observation that the Italian coun-
terparts of the verbs rain/snow can be used with the Italian counterpart of
the perfect auxiliary b e – as is typical of unaccusative verbs:

(48) È piovuto/nevicato tutta la notte

is rained/snowed all the night
‘It has been raining/snowing all night long’

Since Italian is a null subject language, the Italian counterpart of it has a

null spellout in (48).
Now consider the Indefiniteness Condition (45ii). This is illustrated by
Kayne (2016) in terms of contrasts such as the following:

(49) a. Three/many/several/no/some books were on the table

b. There were three/many/several/no/some books on the table

(50) a. The treasure/it definitely exists, so keep looking

b. *There definitely exists the treasure/it, so keep looking
 3.2 Agreement and case in expletive clauses 167

Kayne (2016: 1) formulates the condition as follows: ‘When co-occurring

with expletive there (or a counterpart of it in other languages), existential
verbs only combine with indefinite noun phrases’ – hence with indefinite
(pro)nominal expressions like those italicised in (49), but not with definites
like those italicised in (50).
However, this raises the question of why an expletive should be used at
all in sentences like (49b). Use of an expletive in association with an indef-
inite internal argument may well be motivated by semantic considerations.
Thus, indefinites which move to spec‑TP are ambiguous between a specific
and a non-specific reading, whereas indefinites which remain in situ within
VP in expletive structures allow only a non-specific reading – as illustrated
by the contrast below:

(51) a. A man is in the room

b. There is a man in the room

So, while a man in (51a) can have either a specific or a non-specific inter-
pretation, in (51b) it can only have a non-specific interpretation. This sug-
gests that the use of an expletive pronoun in sentences like (51b) is a device
for ensuring that the associated indefinite expression does not receive a
specific interpretation.
Finally, consider the Inactivity Condition (45iii). In this respect, consider
the following contrast:

(52) There/*It exists no tangible proof of his involvement

Here, expletive there can be used in conjunction with the indefinite QP no

tangible proof of his involvement (in conformity with the Indefiniteness
Condition in 45ii), but expletive it cannot be used in place of there. How
come? The answer is that expletive it would be merged as the specifier of
the V‑bar exists no tangible proof of his involvement, but this V‑bar con-
tains the QP no tangible proof of his involvement which is active (when
it enters the derivation) by virtue of its unvalued case feature: hence, the
Inactivity Condition blocks the use of it here. The condition ensures that
expletive it is only used in a structure which would otherwise lack a goal
active for agreement, with it serving to provide an active goal which can
value unvalued φ-features on a T‑probe.
To summarise: in Module 3.2 we have looked at the syntax of clauses with
an expletive subject. I argued that expletive it enters the derivation with in-
herently valued third-person and singular-number features and an unvalued
case feature, and that its person/number features value those of the auxil-
iary have in sentences such as It has transpired that he was cheating, while
the unvalued case feature on it is valued by (abstract/invisible) agreement
with the null complementiser introducing the main clause. I suggested that
168 Agreement

expletive subjects (being quasi-arguments) originate within VP (like other

arguments), and from there are attracted by the EPP feature on T to move to
spec‑TP. I went on to argue that expletive there enters the derivation with
an inherently valued third person feature and an unvalued case feature,
and that the expletive and its associate jointly value the unvalued features
on the auxiliary (or affix) in T via multiple agreement. The unvalued case
features on expletive there and its associate are in turn valued via (abstract)
agreement with the complementiser introducing the clause.
You should now be able to tackle Exercise 3.2.

3.3 Agreement, A‑Movement and case in infinitives

So far, our discussion of agreement has focused on finite clauses. I have

argued that T and C in such clauses enter the derivation carrying valued
interpretable features (e.g. indicative-mood in the case of C, present/past-
tense in the case of T), and unvalued uninterpretable (person and number)
φ-features. The φ-features on T make (an auxiliary or tense-affix in) T
active as a probe for Agreement and A‑Movement, while the φ-features on
C make the complementiser active as a probe for case assignment. However,
a question which has not been addressed so far concerns what role (if any)
agreement plays in infinitival clauses; and this is the topic tackled in this
module.
English has a range of different types of infinitive clause, including those
bracketed below:

(53) a. There do seem [to have been some problems]

b. We believe [him to be telling the truth]
c. He is believed [to be telling the truth]

(54) a. We are anxious [for the agreements to be ratified]

b. They want [the president to be impeached]
c. It will be essential [to prepare myself]

For reasons outlined in Radford (2020: §4.4–§4.5), the infinitive clauses

bracketed in (53) are TPs, whereas those bracketed in (54) are CPs. More
specifically, the bracketed infinitival TP in (53a) is the complement of the
Raising verb seem, that in (53b) is the complement of the ECM (i.e. excep-
tional case-marking) verb believe, and that in (53c) is the complement of
the passive participle believed. By contrast, the bracketed CP in (54a) is a
for-infinitive, that in (54b) is a for‑deletion infinitive (i.e. a CP headed by
a null/deleted counterpart of for), and that in (54c) is a Control infinitive
headed by a null complementiser. The question which concerns us here is
 3.3 Agreement, A‑Movement and case in infinitives 169

how Agreement, A‑Movement and Case-marking work in infinitive clauses.

In order to answer this, we’ll take a look at each of the sentences in (53, 54)
in turn – starting with the Raising structure in (53a).

3.3.1 Raising infinitives

In (53a), the expletive pronominal DP there is superficially in spec‑TP, but
given the VP-Internal Subject Hypothesis, the expletive will originate as
the specifier of the VP headed by the verb been. It is clear that the auxiliary
do carries agreement properties, since it overtly inflects for agreement (e.g.
here it is in the third person plural form do, but we require the third person
singular form does in ‘There does seem to have been a problem’). By con-
trast, infinitival to doesn’t overtly inflect for agreement at all, though there
are theoretical reasons for thinking that infinitival to must carry a person
feature (albeit this is abstract/invisible).
The reason is that infinitival to attracts expletive there to become its spe-
cifier at the intermediate stage of derivation represented by the bold-print-
ed copy of there in (55) below, when there moves from being the specifier
of the bracketed VP to becoming the specifier of the bracketed TP within
the infinitive clause:

(55) There do seem [TP there to have [VP there been some problems]]

Under the person agreement account of A‑Movement outlined in Module

3.1, a T‑constituent carrying an EPP feature triggers movement in accord-
ance with the condition in (56) below (repeated from 7 above):

(56) EPP Condition

An EPP feature on a probe T attracts the closest XP goal that T agrees
with in person to undergo A‑Movement to spec‑TP, and the EPP feature
is stripped from T (and removed from the derivation) once the movement
has taken place

What this means is that infinitival to in (55) must carry an (abstract) person
feature in order to be able to attract the goal there to become its specifier.
Let us suppose that (for economy reasons), person is the only φ-feature
carried by infinitival to (since there is no theoretical or empirical rationale
for positing that infinitival to carries number as well: e.g. unlike the T aux-
iliary do, infinitival to can’t carry third person singular -s in ‘He does seem
to/*tos be unwell’).
Given the assumptions made above (and others), sentence (53a) There
do seem to have been some problems will be derived as follows. The verb
been is first-merged with the QP some problems and second-merged with
expletive there, forming the VP there been some problems. The resulting VP
170 Agreement

is merged with the perfect aspect auxiliary have to form the PERFP have
there been some problems. The resulting PERFP is then merged with T‑to,
forming the T‑bar below:

(57) T′

T PERFP
to
[Nf-Tns] PERF VP
[u-Pers] have
[EPP] DP V′
there
[3-Pers] V QP
[u-Case] been some problems
[3-Pers]
[Pl-Num]
[u-Case]

Let us suppose that expletive there is a pronominal DP which enters the

derivation carrying a third person feature and an unvalued case feature.
Let’s also suppose that infinitival to enters the derivation carrying an inter-
pretable feature [Nf-Tns] denoting nonfinite tense, an uninterpretable (and
unvalued) person feature, and an EPP feature which is illegible at the PF
and LF interfaces, by virtue of having neither form nor meaning. (I note in
passing that the idea that infinitives carry an abstract person-agreement
feature gains potential cross-linguistic support from the observation that
European Portuguese has infinitives which overtly inflect for agreement in
both person and number with their subject: see Raposo 1987, Ambar 2007,
and Gonçalves et al. 2014.)
In accordance with the Probe Condition (5i), the infinitive particle T‑to
is active as a probe (for Agreement and A‑Movement) by virtue of its un-
valued person feature, and searches for an active goal which can value this
feature. The closest such goal is expletive there, which is active by virtue of
its unvalued case feature; however, there is not φ-complete (since it carries
person but not number), and so cannot value the unvalued person feature
on T‑to by itself. Consequently, T‑to probes deeper into the structure and
locates the QP some problems as the next closest goal, and this is φ-com-
plete (since it carries both person and number), and active (by virtue of its
unvalued case-feature). The two active goals there and some problems can
thus jointly value the unvalued person feature on T‑to as third person via
Agreement (on the assumption that the Completeness Condition 6ii is sat-
isfied as long as one of the agreeing goals is φ-complete).
In addition, the EPP feature on T‑to attracts the closest goal that it agrees
with in person to move to spec‑TP: the closest such goal c‑commanded by
 3.3 Agreement, A‑Movement and case in infinitives 171

T‑to is the expletive pronoun there, which is active by virtue of its unval-
ued case feature. Agreement and A‑Movement result in the structure below
(with the arrow showing A‑Movement, and the position out of which there
moves being shown as a gap —, and the EPP feature being stripped from
T‑to once it has done its job in triggering A‑Movement):

(58) TP

DP T′
there
[3-Pers] T PERFP
[u-Case] to
[Nf-Tns] PERF VP
[3-Pers] have
DP V′
—
V QP
been some problems
[3-Pers]
[Pl-Num]
[u-Case]
The resulting TP is then merged as the complement of the Raising verb
seem, forming the VP seem there to have — been some problems. This VP
is in turn merged as the complement of the T auxiliary do (which enters
the derivation carrying an interpretable present-tense feature, uninterpret-
able and unvalued person and number features, and an EPP feature which
is illegible at both interfaces), forming the T‑bar shown below (simplified
by showing only features on those constituents which remain active for
A‑operations at this point, in order to avoid distracting details):

(59) T′

T VP
DO
[Pres-Tns] V TP
[u-Pers] seem
DP T′
[u-Num] there
[EPP] [3-Pers] T PERFP
[u-Case] to
PERF VP
have
DP V′
—
V QP
been some problems
[3-Pers]
[Pl-Num]
[u-Case]
172 Agreement

At this point, the T auxiliary d o is active as a probe (for Agreement and

A‑Movement) by virtue of its unvalued person and number features, and
searches for an active goal which can value these. The closest goal c‑com-
manded by d o is expletive there, which is active by virtue of its unvalued
case feature. However, there is a defective goal (since it carries no num-
ber feature), and hence the Completeness Condition (6ii) prevents there
from valuing any of the φ-features on d o on its own. Consequently, d o
probes further down into the structure, locating the QP some problems as
the next closest goal (active by virtue of its unvalued case feature). Since
this QP carries both person and number and is therefore φ-complete, mul-
tiple agreement can take place at this point between the active probe d o ,
and its two active goals (there and some problems). Via agreement, the third
person features on there and some problems jointly value the unvalued per-
son feature on d o as third-person, and the plural number feature on some
problems values the unvalued number feature on do as plural.
The EPP feature on d o then attracts the closest goal it agrees with in per-
son (expletive there) to move to spec‑TP, and the EPP feature is thereafter
stripped from T, and thereby removed entirely from the derivation because
it is illegible at the PF and LF interfaces. The resulting TP is merged with
a null complementiser, deriving the CP below (with the arrow showing
A‑Movement):

(60) CP
C TP
ø
[Ind-Mood] DP T′
[u-Pers] there
[3-Pers] T VP
[u-Num] DO
[u-Case]
[Pres-Tns] V TP
[3-Pers] seem T′
DP
[Pl-Num] —
T PERFP
to
PERF VP
have
DP V′
—
V QP
been some
problems
[3-Pers]
[Pl-Num]
[u-Case]

The null complementiser at the top of the tree is active as a probe (for
Agreement and Case-marking) by virtue of its unvalued person and number
 3.3 Agreement, A‑Movement and case in infinitives 173

features, and searches for an XP goal which can value these. The closest
goal c‑commanded by C‑ø is the expletive DP there, which is active by
virtue of its unvalued case feature. However, there is a defective goal (lack-
ing a number feature), and so cannot value any of the φ-features on ‑ø.
Consequently, the complementiser probes further down into the structure,
locating the QP some problems as the next closest potential goal (active by
virtue of its unvalued case feature). Since this QP carries both person and
number and is therefore φ-complete, multiple agreement can take place
at this point between the complementiser and its two active goals (there
and some problems). Via multiple agreement, the third person features on
there and some problems jointly value the unvalued person feature on the
complementiser as third-person, and the plural number feature on some
problems values the unvalued number feature on the complementiser as
plural. Since the complementiser is finite (in that it carries an indicative
mood feature) and φ-complete (in that it carries both person and number),
it can value the unvalued case features on the two goals (there and some
problems) as nominative, so resulting in the final derived syntactic struc-
ture below:

(61) CP
C TP
ø
[Ind-Mood] DP T′
[3-Pers] there
T VP
[Pl-Num] [3-Pers]
[Nom-Case] DO
V TP
[Pres-Tns] seem
[3-Pers] DP T′
[Pl-Num] —
T PERFP
to
PERF VP
have
DP V′
—
V QP
been some
problems
[3-Pers]
[Pl-Num]
[Nom-Case]

The structure in (61) satisfies the requirement imposed by the Legibility

Condition (6iii) for all features to be valued at the end of the syntactic
derivation.
When the tree in (61) undergoes transfer to the PF component, the T
auxiliary do is spelled out in its third person plural present tense form do,
174 Agreement

the expletive is spelled out in its nominative form there, and the QP some
problems is likewise spelled out in its nominative form some problems. The
overall structure is spelled out as the sentence (53a) There do seem to have
been some problems.

3.3.2 ECM infinitives

Now let’s turn to consider the derivation of the ECM infinitive structure
(53b) We believe him to be telling the truth. The infinitive clause can be
argued to be a TP on the grounds that its subject can passivise (as in 53c
‘He is believed to be telling the truth’), and Passivisation is possible across
a TP boundary, but not across a CP boundary (because the Impenetrability
Condition bars movement across the edge of an intervening CP). Given
the VP-Internal Subject Hypothesis, the infinitive subject h e (used as a
case-independent designation of the pronoun variously spelled out as he/
him/his) will originate as the specifier of telling the truth, and then (via
A‑Movement) will be raised to become the specifier of the infinitive particle
T‑to, where it is assigned accusative case by the transitive verb believe in
the higher clause. Let’s take a closer look at the mechanics of this.
The verb telling first-merges with the DP the truth and second-merges
with the pronominal DP h e to form the VP h e telling the truth. The result-
ing VP is merged with the progressive aspect auxiliary/PROG be to form the
PROGP be h e telling the truth, and this in turn is merged with the infinitival
particle T‑to, thereby forming the T‑bar below (simplified by showing only
features on constituents relevant to our discussion of agreement/move-
ment/case-marking here – hence for example, omitting the masculine-gen-
der feature on h e , because this plays no part in person/number agreement):

(62) T′

T PROGP
to
[Nf-Tns] PROG VP
[u-Pers] be
[EPP] DP V′
HE
[3-Pers] V DP
[Sg-Num] telling the truth
[u-Case]

At this point, the infinitive particle T‑to is active as a probe for Agreement
and A‑Movement by virtue of its unvalued person feature, and searches
for an XP goal which can value this feature. The closest such goal c‑com-
manded by T‑to is the pronominal DP h e , which is active by virtue of its
unvalued case-feature. The pronoun h e is φ-complete (since it carries both
 3.3 Agreement, A‑Movement and case in infinitives 175

person and number), and so can value the unvalued person feature on T‑to
as third person via Agreement (6i). In addition, the EPP feature on T‑to
attracts the closest goal that T‑to agrees with in person to move to spec‑TP:
the closest such goal is the third-person pronoun h e , which is active by
virtue of its unvalued case feature. After A‑Movement takes place, T‑to is
stripped of its EPP feature (because EPP has neither form nor meaning, and
so is illegible at the PF and LF interfaces). Agreement and A‑Movement
thus give rise to the structure below (where the A‑Movement operation is
arrowed):

(63) TP

DP T′
HE
[3-Pers] T PROGP
[Sg-Num] to
[u-Case] [Nf-Tns] PROG VP
[3-Pers] be
DP VP
—
V DP
telling the truth

The resulting VP is then merged with the transitive verb believe, which is
responsible for assigning accusative case to the pronoun h e . Since Case-
marking (under the account in 11 above) is contingent on the case assigner
being φ-complete (i.e. carrying person and number) and agreeing with the
case assignee in respect of one or more φ-features, this means that the verb
believe will carry an abstract/invisible set of agreement features. Merging
the verb believe with the TP in (63) above will form the V‑bar shown below
(simplified by not showing the internal structure of the VP telling the truth):

(64) V′

V TP
believe
[u-Pers] DP T′
[u-Num] HE
[3-Pers] T PROGP
[Sg-Num] to
[u-Case] [Nf-Tns] PROG VP
[3-Pers] be telling the truth

The verb believe serves as an active probe for Agreement and Case-marking
at this point, by virtue of its unvalued person and number features. V‑believe
searches for the closest goal that can value its unvalued φ-features, locating
176 Agreement

the pronoun h e (which is active by virtue of its unvalued case feature).

Being φ-complete, h e can value the unvalued φ-features on V‑believe as
third person singular. Conversely, the φ-complete probe V‑believe (being
transitive) can value the unvalued case feature on h e as accusative, result-
ing in the structure below:

(65) V′

V TP
believe
[3-Pers] DP T′
[Sg-Num] HE
[3-Pers] T PROGP
[Sg-Num] to
[Acc-Case] [Nf-Tns] PROG VP
[3-Pers] be telling the truth

Consequently, the pronoun h e is ultimately spelled out at PF in its accusa-

tive form him. Since the agreement features on believe and to are abstract
(i.e. invisible), they have no visible reflex on the spellout of these items.
The derivation in (65) will then continue until it eventually results in the
structure associated with (53b) We believe him to be telling the truth, but I
will not be concerned with the relevant details here, since my main focus is
on agreement and case-marking in the infinitive clause.
An interesting issue raised by the analysis in (65) is how the syntax
‘knows’ that the verb believe is transitive (i.e. is an accusative case as-
signer). One possibility would be to assume that this is because believe
has an external argument (its thematic subject we, not shown in 65);
however, this would clearly not account for why a transitive comple-
mentiser like infinitival for or a transitive preposition like about are also
accusative case assigners (since neither has an external argument/the-
matic subject).
An alternative possibility would be that case assigners carry a fea-
ture which specifies the case that they assign. On this view, a transitive
verb such as believe in a structure like (64) above would carry a feature
like [Ass-Acc] ‘Assign accusative case to any agreeing goal(s) with an
unvalued case feature’. We can suppose that (like EPP features) case-as-
signment features on probes are stripped (and thereby removed from the
derivation entirely) once they have done their work in assigning case to
one or more appropriate goals, since they have no form (e.g. a prepos-
ition like about doesn’t inflect for its ability to assign accusative case),
and no meaning (e.g. the fact that about assigns accusative case to its
complement doesn’t affect its meaning), and thus they are illegible at the
PF and LF interfaces. Indeed, we might similarly treat the EPP feature as
 3.3 Agreement, A‑Movement and case in infinitives 177

being of the form [Attr-Pers] ‘Attract the closest XP with person to be-
come my specifier’. I will not speculate on these issues any further here,
however (though Exercises 3.3 and 3.4 will give you an opportunity to
explore this idea).
A further issue raised by the analysis outlined above concerns the ra-
tionale for positing that transitive verbs agree abstractly with constituents
they assign accusative case to. The rationale is partly theoretical, in that it
follows from the Completeness Condition (6ii) that only a φ-complete probe
can value features on a goal. In addition, there is cross-linguistic empirical
evidence in support of this assumption, from languages where verbs can
agree overtly with constituents to which they assign accusative case – as in
the following Hungarian example from Bárány (2015), where o b denotes
an object agreement suffix:

(66) Lát-om a sajtburger-t

See-1 s g . ob the cheeseburger-acc
‘I see the cheeseburger’

Here, the verb látsee overtly agrees with the DP athe sajtburger-tcheeseburger-ACC
that it assigns accusative case to, and this lends cross-linguistic plausibility
to the idea that accusative case assignment involves agreement (albeit this
type of agreement is abstract/invisible in English).

3.3.3 Passive infinitives

Now let’s turn to consider the derivation of the passive infinitive struc-
ture (53c) He is believed to be telling the truth. Let’s suppose that this
proceeds as for (53b) We believe him to be telling the truth until we have
formed the infinitival TP h e to be — telling the truth in (63) above. At this
point, this TP is merged with the passive verb believed to form the VP
believed h e to be — telling the truth. The resulting VP is then merged with
the T auxiliary b e , forming the structure below (simplified, inter alia, by
showing only features playing a part in Agreement, Case-marking, and
A‑Movement):

(67) T′

T VP
BE
[Pres-Tns] V TP
[u-Pers] believed
[u-Num] DP T′
HE
[EPP]
[3-Pers] T PROGP
[Sg-Num] to
[u-Case] [Nf-Tns] PROG VP
[3-Pers] be — telling the truth
178 Agreement

The finite T auxiliary be (at the top of the tree) enters the derivation carry-
ing an interpretable present-tense feature, uninterpretable (and unvalued)
person and number features, and an EPP feature; the verb believed is a pas-
sive participle here, so is intransitive and therefore does not have the ability
to assign accusative case (and consequently, let us suppose, believed does
not carry agreement features either).
The T auxiliary b e is active as a probe by virtue of its unvalued person
and number features, and searches for the closest XP goal that can value
these, locating the pronominal DP h e (whose unvalued case feature makes
it active). Since h e is φ-complete (through having both person and num-
ber), it can value the unvalued person and number features on T‑b e as third
person singular. At the same time, the EPP feature on T‑b e searches for
the closest (and only) goal that it agrees with in person (namely the third
person pronoun h e , which is active through its unvalued case feature), and
attracts h e to move to spec‑TP. The resulting TP is merged with a null com-
plementiser carrying an interpretable indicative mood feature and uninter-
pretable person and number features, forming the CP shown below (where
the internal structure of the T‑bar to be telling the truth is not shown, and
arrows mark the two A‑Movement operations undergone by the pronoun
he ):

(68) CP
C TP
ø
[Ind-Mood] DP T′
[u-Pers] HE
[3-Pers] T VP
[u-Num] [Sg-Num] BE
[u-Case] [Pres-Tns] V TP
[3-Pers] believed DP T′
[Sg-Num] — to be — telling the truth

The null complementiser in (68) is active as a probe for Agreement and

Case-marking by virtue of its unvalued person and number features,
and searches for an active XP goal which can value these. The closest
goal c‑commanded by the complementiser is the pronominal DP h e ,
which is active by virtue of its unvalued case feature. Since this DP car-
ries both person and number and is therefore φ-complete, it can value
the unvalued person and number features on the complementiser as
third person singular. Since the complementiser is finite (in that it car-
ries an indicative-mood feature) and φ-complete (in that it carries both
person and number), it can in turn value the unvalued case feature on
 3.3 Agreement, A‑Movement and case in infinitives 179

the pronoun h e as nominative, so resulting in the final derived syntactic

structure below:

(69) CP
C TP
ø
[Ind-Mood] DP T′
[3-Pers] HE
[3-Pers] T VP
[Sg-Num] BE
[Sg-Num] V TP
[Pres-Tns]
[Nom-Case]
[3-Pers] believed DP T′
[Sg-Num] — to be — telling the truth

The structure in (69) satisfies the requirement imposed by the Legibility

Condition (6iii) for all features to be valued at the end of the syntactic derivation.
When (69) undergoes transfer to the PF component, the T auxiliary b e is
spelled out in its third person singular present tense form is, and the pro-
noun he is spelled out in its nominative form he. The overall sentence is
spelled out as (53c) He is believed to be telling the truth.

3.3.4 Infinitival CPs

Having looked at the case/agreement properties of infinitival TPs, let’s now
turn to look at those of infinitival CPs, like the italicised complement clause
introduced by the complementiser for in (54a) above ‘We are anxious for
the agreements to be ratified.’ Since the for-clause is passive, its superficial
subject (the DP the agreements) originates as the complement of the passive
participle ratified, and the derivation proceeds as follows.
The (passive) verb ratified merges with the DP the agreements to form the
VP ratified the agreements. This VP is then merged with the passive voice
auxiliary be, forming the VOICEP be ratified the agreements. The resulting
VOICEP is in turn merged with the infinitival T‑particle to, forming the
T‑bar below:

(70) T′

T VOICEP
to
[Nf-Tns] VOICE VP
[u-Pers] be
[EPP] V DP
ratified the agreements
[3-Pers]
[Pl-Num]
[u-Case]
180 Agreement

In accordance with our earlier assumptions, the infinitive particle T‑to

enters the derivation carrying an interpretable nonfinite-tense feature, an
uninterpretable (and unvalued) person feature, and an EPP feature.
At this point, T‑to is active as a probe for Agreement and A‑Movement
by virtue of its unvalued person feature, and searches for an XP goal which
can value this feature. The closest such goal c‑commanded by to is the DP
the agreements, which is active by virtue of its unvalued case feature. This
DP is φ-complete (since it carries both person and number), and so can val-
ue the unvalued person feature on T‑to as third person via Agreement (6i).
In addition, the EPP feature on T‑to attracts the closest goal that it agrees
with in person to move to spec‑TP: the closest (and only) such goal is the
DP the agreements, which is active by virtue of its unvalued case feature.
After A‑Movement takes place, T‑to is stripped of its EPP feature, which
has neither form nor meaning and so is illegible at the PF and LF interfaces.
The resulting TP is merged with the infinitival complementiser for to form
the CP below, with the arrow showing Passivisation/A‑Movement applying
to the DP the agreements:

(71) CP

C TP
for
[Nf-Mood] DP T′
[u-Pers] the agreements
[u-Num] [3-Pers] T VOICEP
[Pl-Num] to
[u-Case] [Nf-Tns] VOICE VP
[3-Pers] be
V DP
ratified —

Because it is a case assigner, and case-assigners are φ-complete in conse-

quence of the Completeness Condition (6ii), the complementiser for enters
the derivation carrying a complete set of uninterpretable (and unval-
ued) φ-features, together with an interpretable nonfinite-mood feature
[Nf-Mood]. The nonfinite-mood feature carried by infinitival complement-
isers generally has much the same irrealis interpretation as we find in the
corresponding subjunctive clause italicised in ‘We are anxious that the
agreements be ratified’ – although the two types of clause differ in the case
assignment properties of their complementisers (that being a nominative
case assigner, and for an accusative case assigner).
The complementiser for in (71) is active as a probe for Agreement and
Case-marking by virtue of its unvalued person and number features,
and searches for an XP goal which can value these. The closest goal
c‑commanded by the complementiser is the DP the agreements, which is
 3.3 Agreement, A‑Movement and case in infinitives 181

active by virtue of its unvalued case feature. Since this DP carries both
person and number and is therefore φ-complete, it can value the unvalued
person and number features on the complementiser as third person plural.
Since the complementiser for is φ-complete (in that it carries both person
and number) and transitive (like its prepositional counterpart in ‘Don’t feel
sorry for me’), it can in turn value the unvalued case feature on the DP the
agreements as accusative in accordance with the Case Condition (11iii), so
resulting in the final derived syntactic structure below:

(72) CP

C TP
for
[Nf-Mood] DP T′
[3-Pers] the agreements
[Pl-Num] [3-Pers] T VOICEP
[Pl-Num] to
[Acc-Case] [Nf-Tns] VOICE VP
[3-Pers] be
V DP
ratified —

Although the accusative case assigned to the DP the agreements is not dir-
ectly visible, it is indirectly visible in the sense that if we replace the DP
by a personal pronoun, we require the accusative form them (as in ‘We are
anxious for them to be ratified’). The structure in (72) satisfies the require-
ment imposed by the Legibility Condition (6iii) for all features to be valued
at the end of the syntactic derivation. Because φ-features on complement-
isers and on nonfinite T constituents are abstract/invisible in English, the
φ-features on the complementiser for and the infinitive particle to have
no overt spellout in the PF component. The overall sentence is ultimately
spelled out as (54a) ‘We are anxious for the agreements to be ratified.’
Now consider the derivation of the italicised infinitive clause in (54b)
‘They want the president to be impeached.’ If we make the traditional as-
sumption that this is a for‑deletion structure (i.e. one in which the italicised
complement clause is headed by a null counterpart of the transitive com-
plementiser for), it follows that the derivation will be essentially parallel to
that outlined in (70–72) above, save for minor lexical differences (i.e. dif-
ferences in the choice of words). Hence, I will not go into the details here,
to avoid unnecessary repetition.
Finally, consider the derivation of a Control infinitive clause like that itali-
cised in (54c) ‘It will be essential to prepare myself.’ The classic analysis of such
clauses assumes that infinitival to has a PRO subject with null case. In this use,
PRO is interpreted as referring to the speaker: this means that PRO (as used
here) is a first person singular pronoun (and hence can bind the first person
182 Agreement

singular anaphor myself). Under the VP-Internal Subject Hypothesis, PRO will
originate as the specifier of the VP headed by the verb prepare. Given this as-
sumption (and others), the complement clause will be derived as follows.
The verb prepare first-merges with the reflexive anaphor myself and sec-
ond-merges with PRO to form the VP PRO prepare myself. The resulting VP
is merged with the infinitive particle T‑to, forming the T‑bar below (where
only features on the constituents of direct relevance to our discussion here
are shown, and where PRO and myself are pronominal DPs):

(73) T′

T VP
to
[Nf-Tns] DP V′
[u-Pers] PRO
[EPP] [1-Pers] V DP
[Sg-Num] prepare myself
[u-Case]

T‑to enters the derivation carrying an interpretable nonfinite-tense feature,

an uninterpretable (and unvalued) person feature, and an EPP feature.
At this point, the infinitive particle T‑to is active as a probe for Agree-
ment and A‑Movement by virtue of its unvalued person feature, and
searches for an active XP goal which can value this feature. The closest
such goal c‑commanded by to is PRO, which is active by virtue of its un-
valued case-feature. PRO is φ-complete (since it carries both person and
number), and so can value the unvalued person feature on T‑to as first
person via Agreement (6i). In addition, the EPP feature on T‑to attracts the
closest goal that it agrees with in person to move to spec‑TP: the closest
such goal is PRO, which is active by virtue of its unvalued case feature.
After A‑Movement takes place, T‑to is stripped of its EPP feature, which
has neither form nor meaning, and so is illegible at the PF and LF inter-
faces. The resulting TP is merged with a null complementiser/C‑ø carrying
a nonfinite mood feature [Nf-Mood] to form the CP below (with the arrow
showing A‑Movement of PRO):

(74) CP

C TP
ø
[Nf-Mood] DP T′
[u-Pers] PRO
[u-Num] [1-Pers] T VP
[Sg-Num] to
[u-Case] [Nf-Tns] DP V′
[1-Pers] —
V DP
prepare myself
 3.3 Agreement, A‑Movement and case in infinitives 183

Because it is a case assigner, and the Completeness Condition (6ii) requires

case-assigners to be φ-complete, the null complementiser enters the deriv-
ation carrying a complete set of uninterpretable (and unvalued) person and
number φ-features, together with an interpretable nonfinite mood feature.
The null complementiser in (74) is active as a probe for Agreement and
Case-marking by virtue of its unvalued person and number features, and
searches for an active XP goal which can value these. The closest such
goal c‑commanded by the complementiser is PRO, which is active by virtue
of its unvalued case feature. Since PRO carries both person and number
and is therefore φ-complete, it can value the unvalued person and number
features on the complementiser as first person singular. Since the comple-
mentiser is nonfinite, intransitive and φ-complete (in that it carries both
person and number), it can in turn value the unvalued case feature on PRO
as null in accordance with the Case Condition (11ii), so resulting in the final
derived syntactic structure below:

(75) CP

C TP
ø
[Nf-Mood] DP T′
[1-Pers] PRO
[Sg-Num] [1-Pers] T VP
[Sg-Num] to
[null-Case] [Nf-Tns] DP V′
[1-Pers] —
V DP
prepare myself

The structure in (75) satisfies the requirement imposed by the Legibility

Condition (6iii) for all features to be valued at the end of the syntactic der-
ivation. Since the infinitive subject has been assigned null case, it receives
a null spellout as PRO in the PF component.
To summarise: I have argued in Module 3.3 that infinitival to (although
not overtly inflecting for agreement) enters the derivation carrying an
unvalued person feature which (in conjunction with the EPP feature on in-
finitival to) serves to attract the closest XP goal carrying person to undergo
A‑Movement to spec‑TP and thereby become the subject of infinitival to
(provided that the XP goal is active). If the resulting to‑infinitive TP is
subsequently embedded as the complement of an item which is not a case-
assigner (like the Raising verb seem in 53a ‘There do seem to have been
some problems’ or the passive participle believed in 53c ‘He is believed to
be telling the truth’), the subject will continue to move until it ends up in
a position within the domain of a case-assigner. However, if the infinitival
TP is embedded as the complement of a case-assigner (e.g. an ECM verb like
184 Agreement

believe in 53b ‘We believe him to be telling the truth’, or a complementiser

like for in 54a ‘We are anxious for the agreements to be ratified’), the sub-
ject in spec‑TP will be case-marked via agreement with the relevant case
assigner, and will thereafter become inactive for further A‑operations.
You should now be able to tackle Exercise 3.3.

3.4 Cross-clausal agreement

In the previous module, we saw that A‑Movement and Agreement can

extend across an infinitive clause boundary. In this module, we will look at
an interesting class of non-standard structures in which agreement applies
across a finite clause boundary.

3.4.1 Agreement across a finite TP

As we see from (76) below, cross-clausal agreement is possible across a
(bracketed) infinitive clause boundary, so allowing agreement between the
bold-printed T auxiliary do and the italicised QP some problems:

(76) There do seem [to have been some problems]

By contrast, agreement is not possible across a finite clause boundary like

that bracketed in (77) below:

(77) *There do seem [that have been some problems]

Why should this be? A plausible answer is that the bracketed infinitive
clause in (76) is a TP (as shown in 60 above), whereas the bracketed finite
clause in (77) is a CP (headed by the complementiser that), and syntactic
operations (including movement and agreement) are barred from applying
across the edge of an intervening CP by the Impenetrability Condition,
which was given the following informal characterisation in §1.5.1 (where
above/below mean ‘c‑commanding/c‑commanded by’):

(78) Impenetrability Condition/IC

Anything below the head C of a CP is impenetrable to anything above
the CP

Since the QP some problems is positioned below the complementiser that

in (77) and the T auxiliary do is positioned above the bracketed CP headed
by that, it follows that the Impenetrability Condition blocks cross-clausal
agreement here (and it likewise blocks expletive there from raising out
of the that-clause to become the subject of do). By contrast, IC does not
block cross-clausal agreement in the corresponding infinitive structure
(76), because the bracketed infinitive complement clause in (76) is a TP
 3.4 Cross clausal agreement 185

(as in 60 above), not a CP, and the Impenetrability Condition only pre-
vents operations from applying across the edge of a CP. If (as claimed
in Radford 2020: 184) ‘All finite clauses have the status of CPs’, then we
should expect that (because of the Impenetrability Condition) syntactic
operations like movement and agreement will never apply across the edge
of a finite clause.
However, the claim that agreement and movement cannot apply across
a finite clause boundary is potentially called into question by Raising of
the (italicised) subject out of a finite clause which is the complement of a
(bold-printed) Raising predicate such as likely/seem in sentences like (79)
below, and by Passivisation of the (italicised) subject of a finite clause
which is the complement of a (bold-printed) passive participle in sentences
like (80):

(79) a. China is likely will field a ground-based laser weapon (WION, on

Facebook)
b. An advanced rider is likely will opt for a variety of bearings
([Link])
c. The additional risk of further borrowing is likely will be reflected in
the interest rate offer you are likely to receive ([Link])
d. He seems might be the honey I need now (lyrics to a song by Grace,
[Link])
e. He seems is very active and shows no sign of illness (labradorforums.
[Link])
f. Another thing to like about Kizer is that he seems has that ‘it’ factor
([Link])
g. Juan Mata appears just has to pass a medical (Alan Brazil,
talkSPORT Radio)

(80) a. He was believed had lured Olga from her apartment … (Medicine Hat
News, August 7th 1962, p.1)
b. He is expected will become a prominent face in showbiz in the future
([Link])
c. He was considered was a hero in Turkey … ([Link])
d. He was deemed was no longer acceptable for naval service …
([Link])
e. Approximately 30% of those were thought were students (warwick.
[Link])
f. He was said was advised ‘by his Turkish community’ that the ‘best
way to pass his theory was to arrange a Bluetooth cheat’ (dailymail.
[Link])

Such structures (although widely attested in both spoken and written English)
tend to be stigmatised as non-standard by prescriptive grammarians (and
186 Agreement

are characterised as ‘bizarre’ by one anonymous reviewer) – and indeed,

they would be expected to be ungrammatical on our existing assumptions.
To see why, consider the derivation of (79a) China is likely will field a
ground-based laser weapon.
Under the VP-Internal Subject Hypothesis, China will originate as the
thematic subject of the verb field, and the derivation of the sentence will
proceed as follows. The verb field first-merges with its complement a
ground-based laser weapon and second-merges with its thematic subject
China to form the VP China field a ground-based laser weapon. The re-
sulting VP is in turn merged with the auxiliary T‑will to form the T‑bar
shown in simplified form below (where I have assumed that T‑will is mor-
phologically a present-tense form carrying unvalued person/number agree-
ment features and an EPP feature, and that China is a DP headed by a null
determiner, with a meaning paraphrasable as ‘the entity China’):

(81) T′

T VP
will
[Pres-Tns] DP V′
[u-Pers] ø China field a ground-based
[u-Num] [3-Pers] laser weapon
[EPP] [Sg-Num]
[u-Case]
The unvalued person/number features on T‑will make it active as a probe
for Agreement and A‑Movement, so it searches for an active XP goal that
can value these. The closest goal c‑commanded by T‑will is the DP ø China,
which is active by virtue of its unvalued case feature. Being φ-complete,
the DP ø China can value the unvalued φ-features on T‑will as third person
singular (although these φ-features have no overt spellout on a modal like
will). In addition, the EPP feature on T‑will attracts the closest goal that
T‑will agrees with in person (= the DP ø China) to move to become the spe-
cifier of will (as shown by the arrow below), and the EPP feature is there-
after stripped from T (because EPP is illegible at the PF and LF interfaces),
giving rise to the following structure:

(82) TP

DP T′
ø China
[3-Pers] T VP
[Sg-Num] will
[u-Case] [Pres-Tns] DP V′
[3-Pers] — field a ground-based
[Sg-Num] laser weapon
 3.4 Cross clausal agreement 187

On the assumption made earlier that all finite clauses are CPs, the resulting
TP will be merged with a null finite complementiser carrying an interpreta-
ble indicative mood feature, and uninterpretable (and unvalued) person and
number features. The complementiser will agree with (and assign nomina-
tive case to) the DP ø China, so deriving the CP below:

(83) CP

C TP
ø
[Ind-Mood] DP T′
[3-Pers] ø China
[Sg-Num] [3-Pers] T VP
[Sg-Num] will
[Nom-Case] [Pres-Tns] DP V′
[3-Pers] — field a ground-
[Sg-Num] based laser weapon

However, given the impoverished nature of English morphology, the case/

agreement features italicised in (83) will have no overt phonetic spellout
at PF (although those on the subject and auxiliary are visible in a sentence
like ‘She is fielding a laser weapon’).
The derivation then continues by merging the CP in (83) as the comple-
ment of the adjective likely, forming an AP which is subsequently merged
as the complement of the verb b e , forming a VP. This VP is then merged
with a null present tense affix (= Af), forming the T‑bar shown in simplified
form below:

(84) T′

T VP
Af
[Pres-Tns] V AP
[u-Pers] BE
[u-Num] A CP
[EPP] likely
C TP
ø
[Ind-Mood] DP T′
[3-Pers] ø China will field a ground-
[Sg-Num] [3-Pers] based laser weapon
[Sg-Num]
[Nom-Case]
At the stage of derivation in (84), the tense affix (T‑Af) is active as a probe
by virtue of its unvalued person and number features, and searches for an
active goal to value these; likewise, the EPP feature on T‑Af searches for
188 Agreement

the closest XP goal that it agrees with in person to attract to spec‑TP. So,
what we might expect to happen at this point is for T‑Af to agree with
(and attract) the DP goal ø China. However, neither operation (Agreement
or Attraction) is possible, for two reasons. Firstly, the DP ø China occu-
pies a position below the null complementiser, and so the Impenetrability
Condition (78) prevents this DP from agreeing with or being attracted by
the tense affix (because the affix is positioned above the CP headed by
the null complementiser). And secondly, the DP ø China is inactive (for
either Agreement or A‑Movement) when we reach the stage of derivation
in (84), because its case feature was valued as nominative at the earlier
stage of derivation in (83). Since a goal can only be active for Agreement or
A‑Movement if it has an unvalued case feature, it follows that the nomina-
tive DP ø China is unable to agree with the tense affix in (84), and unable
to move to become its specifier either. This means that the derivation in (84)
crashes because the person/number features on the tense affix can’t be val-
ued, and because the EPP feature on T is unable to do its work and so can’t
be deleted – and both are illegible at the PF and LF interfaces. So, on the
assumptions made here, (79a) China is likely will field a ground-based laser
weapon is underivable, and hence predicted to be ungrammatical. However,
the problem this poses is that (as data like 79 and 80 above show), such
sentences do actually occur in some varieties of English; and this raises
the question of how cross-clausal agreement (also termed long-distance
agreement) is licensed in those varieties. Below, I’ll suggest an answer to
this question.
The constituent which effectively blocks the tense affix in (84) from
agreeing with and attracting the DP ø China is the intervening null com-
plementiser, because on the one hand this induces an impenetrability vio-
lation, and on the other hand it inactivates the DP ø China by assigning it
nominative case. This raises the possibility that (in varieties which allow
sentences like 79 and 80 above) the complement clause will field a ground-
based laser weapon is a defective clause which projects only as far as TP,
and contains no CP projection at all. Let’s take a closer look at the effect
that this would have on the derivation.
Suppose that the derivation proceeds as before, until we have formed the
TP in (82) above. Suppose, too, that this TP is then merged with the adjec-
tive likely to form an AP, and that the resulting AP is subsequently merged
with the verb b e to form a VP, and that this VP is in turn merged with a
tense-affix to form the T‑bar shown in simplified form in (85) below:
 3.4 Cross clausal agreement 189

(85) T′

T VP
Af
[Pres-Tns] V AP
[u-Pers] BE
[u-Num] A TP
[EPP] likely
DP T′
ø China will field a ground-
[3-Pers] based laser weapon
[Sg-Num]
[u-Case]

The unvalued person/number features on the tense affix/T‑Af make it

active as a probe for Agreement and A‑Movement, so T‑Af searches for an
XP goal that can value these. The closest such goal is the DP ø China, which
is active by virtue of its unvalued case feature. Being φ-complete, the DP ø
China can value the unvalued φ-features on T‑Af as third person singular.
The EPP feature on T‑Af attracts the closest goal that it agrees with in per-
son (= the DP ø China) to move to spec‑TP (as shown by the arrow below),
and the EPP feature is thereafter stripped from T‑Af, because it is illegible
at the PF and LF interfaces. This gives rise to the TP in (86) below:

(86) TP

DP T′
ø China
[3-Pers] T VP
[Sg-Num] Af
[u-Case] [Pres-Tns] V AP
[3-Pers] BE
[Sg-Num] A TP
likely
DP T′
— will field a laser-
based weapon

In addition, V‑be adjoins to T‑Af via Head Movement (arrowed in 87 below):

190 Agreement

(87) TP

DP T′
ø China
[3-Pers] T VP
[Sg-Num]
[u-Case] V Af V AP
BE [Pres-Tns] —
[3-Pers] A TP
[Sg-Num] likely
DP T′
— will field a laser-
based weapon
The resulting TP is merged with a null complementiser to form the CP in
(88) below:

(88) CP
C TP
ø
[Ind-Mood] DP T′
[u-Pers] ø China
[3-Pers] T VP
[u-Num] BE+Af
[Sg-Num] AP
[Pres-Tns] V
[u-Case] —
[3-Pers]
A TP
[Sg-Num] likely
DP T′
— will field a laser-
based weapon …
The null complementiser in (88) is active as a probe for Agreement and
Case-marking by virtue of its unvalued person and number features, and
searches for an XP goal which can value these. The closest such goal c‑com-
manded by the complementiser is the DP ø China, which is active by virtue
of its unvalued case feature. Since this DP carries both person and number
features and is therefore φ-complete, it can value the unvalued person and
number features on the complementiser as third person singular. Since the
null complementiser is finite (by virtue of being indicative in mood) and
φ-complete (in that it carries both person and number), it can in turn value
the unvalued case feature on the DP ø China as nominative in accordance
with the Nominative Case Condition (11i), so resulting in the final derived
syntactic structure in (89) below:
 3.4 Cross clausal agreement 191

(89) CP
C TP
ø
[Ind-Mood] DP T′
[3-Pers] ø China
[Sg-Num] [3-Pers] T VP
BE+Af
[Sg-Num] AP
V
[Nom-Case] [Pres-Tns] —
[3-Pers] A TP
[Sg-Num] likely
DP T′
— will field a laser-
based weapon
The resulting structure satisfies the requirement imposed by the Legibility
Condition for all features to be valued at the end of a derivation, and is
ultimately spelled out at PF as (79a) China is likely will field a ground-based
laser weapon.
The overall conclusion to come out of our discussion here is that
cross-clausal Subject Raising/A‑Movement in sentences like China is likely
will … is ungrammatical in standard varieties of English, and this follows
from the assumption that the complement clause (= the will-clause) is a
CP (as in 84 above), since the head C of the relevant CP inactivates the
complement clause subject (by assigning it nominative case), and makes
the subject impenetrable to a probe in the main clause. By contrast, for
speakers who accept sentences like (79a), the complement will-clause is a
defective clause which has the status of a TP that does not project further
into a CP (as in 85 above), so the complement clause subject is not inac-
tivated, and is not impenetrable to a higher probe. The difference between
the two varieties reduces to the following: in standard varieties, all finite
clauses are CPs, and consequently finite complement clauses do not allow
their subjects to agree with or be attracted by a higher T‑probe; but in some
(non-standard) varieties, the complement of a Raising predicate (or passive
participle) can be a finite TP, and this allows a T probe in a higher clause
to agree with and trigger A‑Movement of the complement clause subject in
sentences like (79, 80) above.
The analysis of cross-clausal A‑Movement outlined in (86) above has
interesting implications for case assignment. Under the analysis of nomina-
tive case-marking in (11i), a finite C assigns nominative case to (and there-
by inactivates) the subject of its clause: the assumption that the will-clause
is a defective clause which projects only up to TP and hence contains no C
192 Agreement

to case-mark its subject allows the subject of will to remain active for the
A‑Movement operation arrowed in (86) above.

3.4.2 Agreement across a finite CP

The key assumption made in the previous section (§3.4.1) is that cross-
clausal agreement and A‑Movement are only possible out of a finite
complement clause that is defective (and has the status of TP), not out a
complete clause that projects into CP. However, the empirical basis of this
claim is cast in doubt by sentences such as the following (which are used
by a minority of speakers only):

(90) a. There appear that there are no major differences … (researcherslinks.

com)
b. There appear that there are many users having this issue … (support.
[Link])
c. There appear that there were 17 articles published or broadcast from
July through November ([Link])
d. We must push boundaries to succeed and continue pushing
boundaries even where there seem that there are no more to push
([Link])
e. There seem that there are many people that are having problems with
TI requirements ([Link])
f. There is likely that there will be less suffering in those areas where
humans are the direct cause or recipient of suffering … (forum.
[Link])
g. There was felt that there was not as tight a control as there should
be … ([Link])
h. There was rumoured that there would be a 6th single called ‘Higher
Ground’ … ([Link])

In these expletive there sentences, a (bold-printed) verb in a higher clause

agrees with an italicised associate in a complement clause introduced by the
underlined complementiser that. Moreover, we find two copies of expletive
there – one as the subject of the main clause, and another as the subject of
the that-clause. What is going on here? A potential answer can be offered
in relation to (90a) above, if we suppose that this is derived as follows.
The verb b e first-merges with the (negative) QP complement no major
differences and second-merges with expletive there to form the VP there
be no major differences. This VP is then merged with a present-tense affix,
forming the T‑bar below:
 3.4 Cross clausal agreement 193

(91) T′

T VP
Af
[Nf-Tns] DP V′
[u-Pers] there
[u-Num] [3-Pers] V QP
[EPP] [u-Case] BE no major differences
[3-Pers]
[Pl-Num]
[u-Case]
The tense affix/T‑Af is active as a probe at this point (because of its unvalued
person and number features), and it searches for the minimal goal set that
can value these. The closest goal c‑commanded by T‑Af is expletive there
(active by virtue of its unvalued case feature), but this is a defective goal by
virtue of lacking number (so not being φ-complete), and thus it cannot value
any φ-features on T‑Af on its own. Accordingly, T‑Af probes deeper into the
structure to find the next closest goal, locating the QP no major differences.
Since this QP has both person and number and so is φ-complete, and since
it is active through its unvalued case feature, it can enter into a multiple
agreement relation with T‑Af. As a result, the third-person features on the
QP and DP goals (no major differences and there) jointly value the unvalued
person feature on T‑Af as third person, and the plural number feature on the
QP goal values the unvalued number feature on the T‑Af probe as plural. At
the same time, the EPP feature on T‑Af attracts the closest goal that it agrees
with in person (= there) to move to spec‑TP (with the EPP feature thereafter
being stripped/removed from the derivation). In addition, the (auxiliary-like)
verb be adjoins to the tense affix in T, via Head Movement. The resulting TP
is then merged with the complementiser that, forming the CP below (with
the gaps — representing the positions out of which there and be move):

(92) CP

C TP
that
[Ind-Mood] DP T′
there
[3-Pers] T VP
[u-Case] BE+Af
[Nf-Tns] DP V′
[3-Pers] —
[Pl-Num] V QP
— no major differences
[3-Pers]
[Pl-Num]
[u-Case]
194 Agreement

Let’s suppose that the complementiser that in (92) carries indicative mood,
but is defective in respect of lacking agreement features (and hence lacking
the ability to assign case, since the Completeness Condition 6ii requires case
assigners to be φ-complete); this means that expletive there and its asso-
ciate remain active at this point, since both have unvalued case features.
The derivation then continues by merging the that-clause structure in
(92) as the complement of the verb appear, forming a VP which is in turn
merged as the complement of a tense affix/T‑Af, so generating the T‑bar
below:

(93) T′

T VP
Af
[Pres-Tns] V CP
[u-Pers] appear
C TP
[u-Num] that
[EPP] [Ind-Mood] DP T′
there
[3-Pers] T VP
BE+Af
[u-Case]
[Nf-Tns] DP V′
—
[3-Pers] V QP
[Pl-Num] — no major
differences
[3-Pers]
[Pl-Num]
[u-Case]
The tense affix/T‑Af in the higher clause (at the top of the tree) is active as a
probe by virtue of its unvalued person and number features, and it searches
for any goal(s) that can value these. The closest goal c‑commanded by T‑Af
is expletive there, whose unvalued case feature makes it active; however,
there is a defective goal (lacking number), so cannot (on its own) value any
of the φ-features on the tense affix. Consequently, T‑Af probes deeper into
the structure, locating the QP no major differences as the next closest goal
(active by virtue of its unvalued case feature); since this QP is φ-complete,
multiple agreement can take place at this point between the probe T‑Af and
its two goals (expletive there and its QP associate no major differences).
Accordingly, the third-person features on the expletive and its associate
jointly value the unvalued person feature on T‑Af as third-person, and the
plural-number feature on the no-QP values the unvalued number feature
on T‑Af as plural. In addition, the EPP feature on T‑Af attracts the closest
 3.4 Cross clausal agreement 195

goal that it agrees with in person (= expletive there) to move to spec‑TP

(and the EPP feature is thereafter stripped from T‑Af), and the resulting TP
is merged with a null complementiser to form the CP in (94) below (with
the arrow showing A‑Movement of there from spec‑TP in the complement
clause to spec‑TP in the main clause):

(94) CP
C TP
ø
[Ind-Mood] DP T′
there
[u-Pers] [3-Pers] T VP
[u-Num] [u-Case] Af
[Pres-Tns] V CP
appear
[3-Pers] C TP
[Pl-Num] that
[Ind- DP T′
—
Mood] T VP
BE+Af
[Nf-Tns] DP V′
[3-Pers] —
[Pl-Num] V QP
— no major
differences
[3-Pers]
[Pl-Num]
[u-Case]

At this point, the null complementiser (C‑ø) at the top of the tree is active
as a probe (by virtue of its unvalued person and number features), and
searches for a goal set that can value them. The closest goal c‑com-
manded by C‑ø is expletive there, whose unvalued case feature makes it
active; but there is a defective goal (lacking number), so cannot value
the unvalued φ-features on C‑ø. Accordingly, C‑ø probes deeper into the
structure, locating the QP no major differences as the next closest goal
(active by virtue of its unvalued case feature). Since this no-QP is φ-com-
plete, it can also serve as a controller for agreement. Consequently, mul-
tiple agreement takes place, and the third-person feature carried by there
and by the no-QP jointly value the unvalued person feature on C‑ø as
third-person, and the plural-number feature on the no-QP values the
unvalued number feature on C‑ø as plural. At the same time, the (finite
and φ-complete) complementiser C‑ø values the unvalued case features
on expletive there and its no-QP associate as nominative, so deriving the
structure below:
196 Agreement

(95) CP
C TP
ø
[Ind-Mood] DP T′
there
[3-Pers]
[3-Pers] T VP
[Pl-Num] Af
[Nom-
[Pres-Tns] V CP
Case]
[3-Pers] appear C TP
[Pl-Num] that
[Ind- DP T′
—
Mood] T VP
BE+Af
[Nf-Tns] DP V′
[3-Pers] —
V QP
[Pl-Num] — no major
differences
[3-Pers]
[Pl-Num]
[Nom-Case]

When the structure in (95) is transferred to the PF component, the lower

T‑head in the complement clause (comprising b e +Af) is spelled out as the
third person plural present tense form are; in addition, the T‑affix in the
main clause is lowered onto the verb appear (by Affix Hopping in the PF
component), and the resulting complex V‑head (of the form V+Af) is spelled
out as the third person plural present tense form appear. The overall struc-
ture would be expected to be spelled out as in (96a) below; but instead, we
find the double-there structure in (96b):

(96) a. *There appear that — are no major differences

b. There appear that there are no major differences

It would seem that when expletive there undergoes the Raising/A‑Movement

operation arrowed in (94), instead of leaving behind a null copy of there
as in (96a), it leaves behind an overt copy of there immediately following
the complementiser that in (96b). For obvious reasons, the type of Raising
found in (96b) has become known as Copy Raising (because the raised
constituent leaves an overt copy of itself behind). Why should there leave
a copy behind when it raises?
A plausible answer is this. Following the idea put forward by Ross
(1967) that constraints on extraction apply only to gap-creating movement
operations (i.e. to movement operations which leave a gap/null copy be-
hind), let’s suppose that what rules out (96a) is that it creates the structure
shown in simplified form below:

(97) *There appear [CP [C that] [TP there [T are] no major differences]]
 3.4 Cross clausal agreement 197

It would then follow that a structure like (97) is ruled out because it vio-
lates the Impenetrability Condition, in that a null copy of there (= there)
is positioned below the complementiser that, and its antecedent (= there)
is positioned above the CP headed by that – leading to an impenetrability
violation. Additionally, we can suppose that a structure like (97) violates the
Minimality Condition, if this requires a trace/gap/null copy in an A‑position
to be bound by (i.e. have as its antecedent) a constituent on the edge of the
minimal/closest clausal functional projection above it; and since constitu-
ents in an A‑position can either move to the edge of TP via A‑Movement
or to the edge of CP by A‑bar Movement, this means that a trace in an
A‑position must be bound within the closest TP or CP above it. If so, there
would be a minimality violation in (97) because there is not bound by a con-
stituent on the edge of the CP immediately above the TP containing there.
Furthermore, (97) would also violate a filter (devised by Chomsky &
Lasnik 1977: 451) which was given the following formulation in §1.5.2:

(98) COMP‑Trace Filter/CTF

Any structure at the end of a syntactic derivation in which an overt
COMP/complementiser is immediately adjacent to and c‑commands a
trace (i.e. a gap/null copy left behind by a moved constituent) is filtered
out as ill-formed

The CTF violation arises because the overt complementiser that c‑commands
and immediately precedes a null copy of the raised subject there. We could
then hypothesise that use of an overt (italicised) copy of there after that in
(96b) is a rescue strategy designed to avoid the Impenetrability/Minimality/
COMP‑Trace violations which would ensue if a gap/null copy were used as
in (96a). On this view, spelling out the copy of there following that overtly
provides a way to repair the structure (i.e. to ensure that it does not violate
the Impenetrability/Minimality/COMP‑Trace Conditions). By hypothesis,
repair mechanisms are only used as a last resort, where required in order to
avoid a derivation crashing.
The Copy Raising analysis sketched above for there appear(s) that there …
structures can arguably be extended to cases like the following:

(99) a. There look like there might be some interesting changes …

([Link])
b. There look like there have been some problems (Radford 2004b: 317)
c. There look like there were a few that didn’t make it
([Link])
d. There looks like there is a massive leak … ([Link])
e. There looks like there has been a high turnout ([Link])
f. There looks like there was a crowd of young people …
([Link])
198 Agreement

(100) a. There sound like there are some interesting developments …

([Link])
b. There sound like there were some significant differences too
([Link])
c. There sound like there have been quite a lot of changes …
([Link])
d. There sounds like there is going to be a fun event … (salisburyjournal.
[Link])
e. There sounds like there has been a lot of sadness … (unexplained-
[Link])
f. There sounds like there was a game last night ([Link])

Rogers (1974: 53), Rooryck (2000: 48), López-Couso & Méndez-Naya (2012),
and Kanda & Honda (2018: 141–42) treat like as a complementiser in this
use (and they propose a similar treatment for as if/as though). If so, using
an overt copy of there following like could be a repair strategy designed to
avoid Impenetrability/Minimality/COMP‑Trace violations.

3.4.3 Copy Raising with non-expletive subjects

It might seem as if the Copy Raising analysis outlined in §3.4.2 for sen-
tences with expletive subjects can be extended in a relatively straightfor-
ward fashion to handle sentences like those in (101) below with italicised
non-expletive personal pronouns as their subjects:

(101) a. I appear that I may have accidentally mis-represented the 802.3

position ([Link])
b. We KNOW that we need at least 6 players but we appear that we
MIGHT get one ([Link])
c. Do you appear that you’re lying, even when you’re telling the truth?
([Link])
d. He appears that he is unable to deal with disappointment ([Link].
uk)
e. She appears that she doesn’t want to budge at all ([Link])
f. They appear that they are planning to bring mischief to DCI
([Link])

However, a complication that arises with a Copy Raising analysis of sentences

like (101) above is that when the (italicised) main clause subject is nominal,
the (bold-printed) complement clause subject is a personal pronoun:

(102) a. The president seems that he’s having fun ([Link])

b. Boris Johnson looks like he’ll win big ([Link])
c. Your other trouble sounds like it may be a one-time glitch ([Link].
[Link])
 3.4 Cross clausal agreement 199

d. When faced with the task, many students appear as if they do not
know where to begin (Gasper 2014: 1)
e. Boris looks as though he’s survived the Cummings lockdown road
trip ([Link])
f. Baroness Thatcher was announced that she had died (Danny Savage,
BBC Radio 5 Live)
g. The police have been found that they can be held accountable
(Listener, BBC Radio 5 Live)

In other words, if sentences like (102) involve Copy Raising, we have to say
that the copy left behind by raising the italicised subject out of the that/
like/as if/as though-clause is not a full copy of the raised DP/QP, but rather
a pronominal copy. Sentences like (102) have been argued to involve move-
ment, because idiomatic nominals like those italicised below can occur in
this type of structure:

(103) a. The shit looks like it’s gonna hit the fan (Rogers 1974: 154)
b. All hell sounds like it’s breaking loose under the hood
([Link])
c. Some headway looks like it could be made in the near future
([Link])
d. Close tabs sound as if they are being kept on the CIA
e. The fur seems like it’s gonna fly
f. Dirty tricks look like they are being played on the opposition
g. The wool looks like it’s been pulled over his eyes

The premise of the idiom argument for movement is that (e.g. in 103g) the
DP the wool originates as the complement of the verb pull in the idiom pull
the wool over someone’s eyes, and hence must undergo Copy Raising in
order to become the subject of the looks clause. However, if sentences like
those in (103) do indeed involve Copy Raising, the question arises as to why
the lower copy is spelled out as a pronoun (rather than as a full copy of the
raised italicised subject).
One answer which could be given to this question is in terms of econ-
omy of spellout (i.e. a desire to minimise the amount of overt material
pronounced). Reasoning along these lines, we might suppose that only the
person/number/gender/case features of the copy of the raised subject fol-
lowing the underlined complementiser are spelled out/lexicalised (not its
other features), and this can most economically be done by using a person-
al pronoun to spell out the relevant bundle of person/number/gender/case
features. An alternative (syntactic) approach is to suppose that the italicised
raised nominal is underlyingly contained within an XP shell headed by the
pronoun (so that the nominal and its pronominal copy underlyingly form
200 Agreement

a single phrase/XP), and that raising moves only the nominal, leaving the
pronoun behind. We can provide an informal illustration of the difference
between these two approaches in relation to (102a) The president seems
that he’s having fun.
Under the spellout approach, the DP the president undergoes Copy Rais-
ing, resulting in the superficial syntactic structure shown in highly simpli-
fied form below:

(104) The president seems that the president is having fun

In the PF component, only the person/number/gender/case features on

the bold-printed copy following that are spelled out, with the result that
the bold-printed copy is lexicalised as the third person masculine singular
nominative pronoun he.
By contrast, under the shell approach, the DP the president originates in-
side an XP headed by the pronoun he, and the DP the president undergoes
movement on its own, resulting in the syntactic structure in (105) below,
where the pronoun he is stranded as the subject of the complement clause:

(105) The president seems that [he the president] is having fun

Both approaches are problematic in a number of ways: for example, the

spellout solution in (104) in effect allows the PF component to introduce
a new word (the pronoun he) into the derivation; and the shell analysis in
(105) (in extracting the president out of a larger XP he the president which
has moved from being the specifier of having to becoming the specifier
of is) leads to violation of constraints on movement (e.g. the Constraint
on Extraction Domains which bars movement out of a specifier, and the
Freezing Principle which bars extraction out of a moved constituent).
One way of avoiding technical problems in accounting for how a raised
constituent can leave behind a pronominal copy is to adopt an alternative
binding approach developed in work dating back to Lappin (1983, 1984).
The key assumption underlying the binding analysis is to suppose that sen-
tences like those in (102) involve two separate (i.e. independently generat-
ed) subjects, with the main clause subject binding (i.e. serving as the ante-
cedent of) the pronoun which functions as the complement clause subject.
Let’s look briefly at how such an analysis (if adapted to the framework used
here) would handle a sentence like (102b) Boris Johnson looks like he’ll
win big. Let’s suppose that like is a finite complementiser which merges
with the TP he’ll win big and assigns nominative case to the pronoun he.
Suppose too that the CP like he’ll win big is merged as the complement of
look, forming the V‑bar look like he’ll win big. In addition, suppose that
this V‑bar is then merged with Boris Johnson (which we can take to be a
DP headed by a null determiner, with the overall DP being paraphrasable
 3.4 Cross clausal agreement 201

informally as ‘the specific individual Boris Johnson’), to form a VP which is

in turn merged with a present-tense affix to form the T‑bar below:

(106) T′

T VP
Af
[Pres-Tns] DP V′
[u-Pers] Boris Johnson
[u-Num] [3-Pers] V CP
[EPP] [Sg-Num] look like he’ll win big
[u-Case]
The tense affix/Af is active as a probe by virtue of its unvalued person
and number features, and it searches for any active goal(s) that can value
these. The only active goal c‑commanded by T‑Af is the DP Boris Johnson
(whose unvalued case feature makes it active), since all constituents inside
the complement clause CP are inactive at this point. Accordingly, the per-
son/number features on the DP Boris Johnson value the unvalued person/
number features on T‑Af as third person singular. In addition, the EPP
feature on T‑Af attracts (a copy of) the closest goal that it agrees with in
person (= the DP Boris Johnson) to move to spec‑TP (and the EPP feature
is thereafter stripped from T‑Af and thereby removed from the derivation).
Subsequently, the resulting TP is merged with a null complementiser to
form the CP in (107) below (with the arrow showing A‑Movement of the DP
Boris Johnson from spec‑VP to spec‑TP in the main clause):

(107) CP

C TP
ø
[Ind-Mood] DP T′
[u-Pers] Boris
[u-Num] Johnson T VP
[3-Pers] AF
[Sg-Num] [Pres-Tns] DP V′
[u-Case] [3-Pers] —
[Sg-Num] V CP
look like he’ll win big

At this point, the null complementiser at the top of the tree is active as a
probe (by virtue of its unvalued person and number features), and searches
for any active goal(s) that can value them. The only active goal c‑com-
manded by the complementiser is the DP Boris Johnson (whose unvalued
case feature makes it active); accordingly, this DP values the unvalued
φ-features on the complementiser as third person singular. At the same
time, the complementiser (being both finite and φ-complete) also values the
202 Agreement

unvalued case feature on the DP Boris Johnson as nominative, so deriving

the structure in (108) below:

(108)
CP

C TP
ø
[Ind-Mood] DP T′
[3-Pers] Boris
[Sg-Num] Johnson T VP
[3-Pers] Af
[Sg-Num] [Pres-Tns] DP V′
[Nom-Case] [3-Pers] —
[Sg-Num] V CP
look like he’ll win big

When this structure is handed over to the PF component, the present tense
affix is lowered from T onto the verb look by Affix Hopping, and the result-
ing complex V‑head look+Af is spelled out as looks (and the overall sen-
tence as Boris Johnson looks like he’ll win big).
More interesting is the question of what happens when the structure in
(108) is processed in the semantic component. On the assumption that look
(in this use) is a verb which does not theta-mark its subject (which is why
look allows an expletive there subject, as we saw earlier), it follows that the
DP Boris Johnson will not be assigned a theta role by the verb look. How
ever, since this DP is not an expletive (and occupies an A‑position), it must
be assigned a theta role somehow, or else the derivation will crash. But how?
The answer suggested by Lappin (1984: 245) is that in such structures ‘The
matrix subject inherits its theta role from the pronoun which it binds in
the complement.’ This means that if the DP Boris Johnson is interpreted as
binding the complement clause subject he, then Boris Johnson will inherit
the theta role of he (and hence will be interpreted as the agent argument of
win). This binding is possible because the pronoun he requires a third person
masculine singular antecedent, and the DP Boris Johnson satisfies this re-
quirement (although gender is not shown in the simplified structure of 108).
A potential problem might at first sight appear to be that this binding
relation applies across the edge of an intervening CP (= across the edge of
the CP headed by like), so appears to violate the Impenetrability Condition.
However, this is unproblematic, since a (bold-printed) pronominal like he
can freely be bound (across an intervening CP boundary) by an (italicised)
antecedent in a higher clause, as we see from (109) below:

(109) Boris didn’t know [CP that/whether/if he was behind in the polls]

Consequently, any such objection to the derivation in (106–108) is spurious.

 3.5 Summary 203

What is interesting about the derivations sketched in this section from a

theoretical point of view is that they suggest that a non-thematic position
(like the specifier position in a VP headed by a verb like seem/appear/look
etc.) can either be filled by an expletive (as in our discussion of sentences
like there seems/appears/looks …), or by a non-expletive constituent that
is construed as binding a pronoun which originates in a thematic position
within a complement clause (thereby enabling the relevant pro/nominal to
inherit the theta role of the pronoun that it binds).
To summarise: in Module 3.4 we have looked at cases of cross-clausal
Agreement, A‑Movement and Case-marking. I began by discussing sen-
tences like (79a) China is likely will field a ground-based laser weapon,
where China is raised from being the subject of will to becoming the subject
of is, and is thereby extracted out of a finite will-clause. I argued that in
such cases, the finite complement clause is defective (in the sense that it is
a TP, not a CP), and that the absence of a complementiser to case-mark the
subject of will means that the subject is active and can (via A‑Movement)
be raised to become the subject of is. I went on to discuss sentences like
(90a) There appear that there are no major differences, where Agreement
(between appear and no major differences) applies across an intervening CP
headed by that. I sketched a Copy Raising account on which expletive there
(via A‑Movement across the intervening complementiser that) raises from
being the subject of are to becoming the subject of appear, leaving behind
an overt copy of there, in order not to induce an Impenetrability/Minimal-
ity/COMP‑Trace violation. I went on to discuss sentences like (102b) Boris
Johnson looks like he’ll win big, arguing that these involve directly gener-
ating Boris Johnson in a non-thematic position as the subject of look, and
he in a thematic position as the subject (and ag e nt argument) of win, with
Boris Johnson binding (and thereby inheriting the theta role of) he: on this
view, such sentences do not involve cross-clausal Agreement, A‑Movement
or Case-marking.
You should now be able to tackle Exercise 3.4.

3.5 Summary

In this chapter, we have looked at the syntax of Agreement, A‑Movement,

and Case-marking. In Module 3.1 we saw that subject-verb Agreement
involves a relation between a finite T‑probe and a nominal or pronominal
XP goal which it c‑commands: the T‑probe is made active by unvalued
(person/number) φ-features which it carries when it enters the derivation,
and its (pro)nominal goal is made active by its unvalued case feature. More
generally, the probe and goal enter the derivation carrying a set of valued
204 Agreement

inherent/interpretable features, and a set of unvalued contextual/uninter-

pretable features, and the unvalued features are valued by operations like
Agreement and Case-marking in the course of the derivation. We also saw
that Case-marking is contingent on a person/number agreement relation
between probe and goal, and that A‑Movement involves a T probe attract-
ing the closest goal that it agrees with in person to become its specifier. I
argued that Agreement, A‑Movement and Case-marking are subject to the
following conditions governing their application:

(5) Activity Conditions

(i) A head can only be an active probe for Agreement or A‑Movement
or Case-marking if it carries some unvalued φ-feature/s, and becomes
inactive once its φ-feature/s have been valued (Probe Condition)
(ii) An XP can only be an active goal for Agreement, A‑Movement,
or Case-marking if it has an unvalued case feature, and becomes
inactive once its case feature is valued (Goal Condition)
(iii) If a constituent is part of a movement chain at a given point in
a derivation, only the highest copy in the chain can be active for
operations like Agreement, A‑Movement or Case-marking (Highest
Copy Condition)

(6) Feature Valuation Conditions

(i) When an active probe like T c‑commands an active XP goal, each
unvalued φ-feature on the probe will be valued by the goal – that
is, assigned a value which is a copy of the corresponding feature
value on the goal (Agreement)
(ii) Only a φ-complete XP (i.e. one with person and number) can serve
as a goal valuing unvalued features on a probe, and conversely
only a φ-complete probe (i.e. one with person and number) can
value unvalued features on a goal (Completeness Condition)
(iii) Each feature on a constituent must be valued (i.e. assigned a single,
appropriate value) in order for the constituent to be legible at the PF/
LF interfaces; if not, the derivation will crash (Legibility Condition)

(7) EPP Condition

An EPP feature on a probe T attracts the closest XP goal that T agrees
with in person to undergo A‑Movement to spec‑TP, and the EPP feature
is stripped from T (and removed from the derivation) once the movement
has taken place

(11) Case Conditions

A φ-complete case-assigning probe values an unvalued case feature
on an XP goal that it agrees with in respect of one or more φ-features,
 3.5 Summary 205

and assigns the goal a case value that depends on the nature of the
probe – namely:

(i) nominative if the probe is a finite C with a TP complement (where a

finite C is one that is indicative, subjunctive or imperative in mood)
(ii) null if the probe is an intransitive infinitival C with a TP complement
(iii) accusative if the probe is transitive

In Module 3.2, we looked at the syntax of clauses with an expletive sub-

ject. I argued that expletive it enters the derivation with inherently valued
third-person and singular-number features and an unvalued case feature,
and that its person/number features value those of the auxiliary h av e in
sentences such as It has transpired that he was cheating, while the unval-
ued case feature on it is valued by (abstract/invisible) agreement with the
(overt or null) complementiser introducing the relevant clause. I further
argued that expletive subjects (like other subjects) are quasi-arguments that
originate within VP, and from there are attracted by the EPP feature on T
to move to spec‑TP. I went on to claim that expletive there (in existential
clauses) enters the derivation with an inherently valued third person fea-
ture and an unvalued case feature, and is initially merged as the specifier
of an existential predicate which has the associate of the expletive as its
complement. Expletive there and its associate jointly value the unvalued
features on the auxiliary (or affix) in T via multiple agreement. The un-
valued case features on expletive there and its associate are in turn valued
via (abstract) agreement with the complementiser introducing the relevant
clause. I suggested that expletives are subject to the following conditions
on their use:

(45) Expletive Conditions

(i) External Argument Condition
An expletive can only be merged as the highest argument of a verb with
no external argument
(ii) Indefiniteness Condition
Existential there can only be merged as the specifier of an existential
verb with an indefinite (pro)nominal complement
(iii) Inactivity Condition
Expletive it can only be merged with a constituent which does not
contain an active nominal or pronominal (i.e. one carrying an unvalued
case feature)

In Module 3.3, we looked at the syntax of Agreement, A‑Movement and

Case-marking in infinitive clauses. I argued that infinitival to (although not
overtly inflecting for agreement) enters the derivation carrying an unvalued
person feature which (in conjunction with the EPP feature on infinitival to)
206 Agreement

serves to attract the clause subject to undergo A‑Movement from its ini-
tial position inside VP to its derived position in spec‑TP as the subject of
infinitival to. If the resulting to‑infinitive TP is subsequently embedded as
the complement of an item which is not a case-assigner (like the Raising
verb seem in ‘There do seem to have been some problems’ or the passive
participle believed in ‘He is believed to be telling the truth’), the subject
will continue to move until it ends up in a position within the domain of
a case-assigner. However, if the infinitival TP is embedded as the comple-
ment of a case-assigner (e.g. an ECM verb like believe in ‘We believe him
to be telling the truth’, or a complementiser like for in ‘We are anxious for
the agreements to be ratified’), the subject in spec‑TP will be case-marked
by the relevant case assigner (in conformity with the Case Conditions in 11
above), and will thereafter become inactive for further A‑operations.
In Module 3.4, we looked at cases of cross-clausal Agreement, A‑Move-
ment and Case-marking. I began by discussing sentences like (79a) China
is likely will field a ground-based laser weapon, where China is raised from
being the subject of will to becoming the subject of is. I argued that in such
cases, the finite complement clause is defective (in that it is a TP rather than
a CP), and that the absence of a complementiser to case-mark the subject of
will means that the complement clause subject is active and can be raised
up to become the subject of is by A‑Movement. I then went on to discuss
sentences like (90a) There appear that there are no major differences, where
Agreement (between appear and no major differences) applies across an
intervening CP headed by that. I sketched a Copy Raising account on which
expletive there (via A‑Movement across the defective intervening comple-
mentiser that) raises from being the subject of are to becoming the subject
of appear, leaving behind an overt copy of there, in order to avoid viola-
tions of the Impenetrability/Minimality/COMP‑Trace conditions. I went on
to discuss sentences like (102b) Boris Johnson looks like he’ll win big, ar-
guing that these involve base-generating Boris Johnson in a non-thematic
position as the subject of look, and he in a thematic position as the subject
(and ag e nt argument) of won, with Boris Johnson binding (and thereby in-
heriting the theta role of) he: on this view, such binding structures do not
involve Copy‑Raising (and likewise do not involve cross-clausal Agree-
ment, A‑Movement or Case-marking).

3.6 Bibliographical notes

On agreement typology (i.e. a cross-linguistic survey of agreement pat-

terns in a wide range of languages), see Corbett (2006); for an overview of
the treatment of agreement in Minimalism, see Fuß (2005), Baker (2008),
 3.6 Bibliographical notes 207

Miyagawa (2010), Preminger (2012, 2014), and Smith et al. (2020). Note that
this chapter is specifically concerned with agreement in the verbal domain,
setting aside Concord in the nominal domain (e.g. between determiners,
adjectives and the nouns they modify): on Concord, see Norris (2014). Also
set aside here are other phenomena which are taken by some to involve
agreement, such as Negative Concord (Zeijlstra 2004), or sequence of tense
restrictions (Zeijlstra 2012; Kauf & Zeijlstra 2018; see also Armenante &
Braun 2022).
The probe-goal account of Agreement sketched in Module 3.1 is
loosely based on Chomsky (1998, 1999, 2001): for a critique of this ap-
proach, see Danon (2011); and for a critique of the Activity Condition,
see Nevins (2005), and Bošković (2007). The assumption about feature
valuation made in Chomsky (1998, 1999) that features which enter the
derivation valued are interpretable and those which enter the derivation
unvalued are uninterpretable is called into question in work by Pesetsky
& Torrego (2004, 2007); they argue for the two additional possibilities
that features can also enter the derivation (i) uninterpretable but val-
ued, and (ii) interpretable but unvalued. In much the same vein, Bošk-
ović (2011b) argues that gender features on nouns with arbitrary gender
in languages like Serbo-Croation enter the derivation valued, in spite
of arbitrary gender being an uninterpretable feature. Chomsky (2007,
2008) and Miyagawa (2005, 2006a, 2010) argue that agreement features
originate on C and are ‘handed down’ to T; for a challenge to this view,
see Carstens (2003), and Haegeman & van Koppen (2012). On so-called
semantic agreement in British English structures like The government
are ruining the country, see den Dikken (2001), Sauerland & Elbourne
(2002), and Smith (2017). On complementiser agreement, see Bayer
(1984b), Haegeman (1992), Zwart (1993, 1997, 2006), Shlonsky (1993),
Bennis & Haegeman (1994), Hoekstra & Smits (1999), Carstens (2003,
2016), Fuß (2004, 2008, 2014), Barbiers et al. (2005), van Koppen (2005,
2017), Weiß (2005), Gruber (2008), Mayr (2010), Diercks (2010, 2013),
Haegeman & van Koppen (2012), Boef (2013), Lewis (2013), Deal (2015),
and Diercks et al. (2020).
The idea that nominative case is assigned by a finite C dates back to
Chomsky (2001). However, much other research follows earlier work by
Iatridou (1993) and Chomsky (1998, 1999) in taking a tensed T/INFL to
be the constituent that assigns nominative case to clause subjects under
agreement. A variant of this view is to see nominative case as the reflex of
a tense feature on T (Varlokosta 1994, Alexiadou & Anagnastopoulou 1999,
Uchibori 2000, Pesetsky & Torrego 2001, 2004, 2007). On languages which
show overt agreement between verbs and constituents to which they assign
accusative case, see Bárány (2015), and Smith (2020); however, note that
208 Agreement

overt object agreement in some languages may be restricted to objects with

a certain type of property, such as definiteness or specificity. It should also
be noted that a number of researchers have argued that case-marking can
be independent of agreement: see Uchibori (2000), Woolford (2006), Baker
(2008), Bobaljik (2009), Dalrymple & Nikolaeva (2011), Ussery (2009, 2012),
Petersen (2011), and Preminger (2012, 2014).
On the nature of the EPP feature of T which attracts the closest ap-
propriate goal to move to become its specifier (and whether it is possible
to dispense with it), see Chomsky (1982, 1995, 1998, 1999, 2001), Roth-
stein (1983), Alexiadou & Anagnastopoulou (1998), Martin (1999), Déprez
(2000), Grohmann et al, (2000), Holmberg (2000), Sabel (2000), Goodall
(2001), Kiss (2001), Lasnik (2001), Bošković (2002, 2007), Roberts & Rous-
sou (2002), Rosengren (2002), Svenonius (2002a), Haeberli (2003), Richards
(2003), Epstein et al. (2005), Goldshlag (2005), Miyagawa (2005, 2006a,
2010), van Craenenbroeck & den Dikken (2006), Epstein & Seely (2006),
Sheehan (2006), Landau (2007), Sifaki & Sitaridou (2007), Boeckx (2008),
Castillo et al. (2009), Roberts (2010), Sigurðsson (2010), Lin (2011), Cable
(2012), Chapman (2013), Chomsky (2015), Richards (2016), and Fernán-
dez-Salgueiro (2020).
On the syntax of expletive pronouns/structures, see Jenkins (1975), Bol-
inger (1977), Milsark (1977), Breivik (1983), Williams (1984, 2006), Postal
& Pullum (1988), Inoue (1991), Lasnik (1992, 1995, 2003), Abbott (1993),
den Dikken (1995), Groat (1995, 1999), Rothstein (1995), Law (1996), Moro
(1997, 2006), Sobin (1997), Chomsky (1998, 1999, 2000, 2001), Schütze
(1999), Koster & Zwart (2000), Felser & Rupp (2001), Bowers (2002), Han
(2004), Hazout (2004a, 2004b), Sobin (2004), Rupp (2005), Moro (2006),
Rezak (2006), Henry & Cottell (2007), Hornstein (2007), Biberauer (2008),
Hartmann (2008), Kallulli (2008), Deal (2009), Light (2015), Kayne (2016),
Haider (2017, 2019), Greco et al. (2017), and van Craenenbroeck (2020). On
structures such as There were three fish caught in the lake, see Chomsky
(1999), Bowers (2002), Caponigro & Schütze (2003), and Rezak (2006). On
agreement in structures like There’s lots of people in the room, see Sobin
(1997), Schütze (1999), and den Dikken (2001). On multiple agreement in
expletive structures, see Chomsky (2001, 2008), Hiraira (2001, 2005), and
Henderson (2006).
The claim made in Module 3.3 that infinitival to carries EPP and per-
son features follows the analysis proposed in Chomsky (1999) for defective
infinitive clauses; but note that (contrary to what is proposed here), much
work assumes that to in Control clauses carries both person and number
features. On infinitives which overtly inflect for person/number agreement
in European Portuguese, see Raposo (1987), Ambar (2007), and Gonçalves
 3.6 Bibliographical notes 209

et al. (2014). Not touched on in our discussion here are non-standard for-to
infinitives (as in He is likely for to get injured) in some varieties of English,
including Belfast English (Henry 1992, 1995), Ottowa Valley English (Carroll
1983), Ozark English (Elgin & Haden 1991), and Smoky Mountain English
(Montgomery & Hall 2004).
The phenomenon discussed in Module 3.4 of Raising subjects out of fi-
nite clauses in structures like China is likely will field weapons has potential
parallels with Raising out of certain types of finite complement clause in
languages like Brazilian Portuguese, Greek, Romanian, and Bantu (a phe-
nomenon sometimes termed Hyper‑Raising): see Grosu & Horvath (1984),
Harford Perez (1985), Rivero (1989), Iatridou (1993), Watanabe (1993), Ura
(1994), Varlokosta (1994), Alexiadou & Anagnastopoulou (1999), Ferrei-
ra (2004, 2009), Rodriguez (2004), Martins & Nunes (2005, 2009), Zeller
(2006), Nunes (2008), Landau (2009), Alexiadou et al. (2012), Carstens &
Diercks (2013), Petersen & Terzi (2015), Yuan (2016), Fong (2018), and Zy-
man (2018a).
Copy Raising is a phenomenon briefly noted in Postal (1971: 162–63,
1974: 200), and analysed in more detail in Rogers (1971, 1972, 1974), Jo-
seph (1976), Perlmutter & Soames (1979), Horn (1981), Lappin (1983, 1984),
Déprez (1992), Heycock (1994: 272ff), Ura (1994, 1995), Gisbourne (1996,
2010), Groat (1997), Moore (1998), Terada (2000), Potsdam & Runner (2001),
Matushansky (2002), Asudeh (2002, 2004, 2011b, 2012), Rezak (2004, 2011),
Fujii (2005, 2007), Asudeh & Toivonen (2006, 2007, 2012), Landau (2009,
2011), Mack (2010), Takano (2013), Kim (2014), Ziegler (2014), Doran (2015),
Frazier & Clifton (2015), Brook (2016, 2018), Poortvliet (2016), den Dikken
(2017, 2018), Camilleri (2018), Kanda & Honda (2018), Kobayashi (2018)
and Rudolph (2019).
On cross-clausal/long-distance agreement, see Bruening (2001), Brani-
gan & McKenzie (2001), Polinsky & Potsdam (2001, 2006), Polinsky (2003),
Hazout (2004a), Legate (2005), Frank (2006), Bošković (2007), Alexiadou et
al. (2012), Petersen & Terzi (2015), Branan (2018), Ostrove (2018), Bjorkman
& Zeijlstra (2019), Wurmbrandt (2019), and Börjeson & Müller (2020).
A technical issue not touched on in this chapter concerns the locus of
agreement, and more specifically whether (as assumed here) Agreement
is an operation taking place in the syntax and dependent on syntactic re-
lations like c‑command, or whether it is a post-syntactic operation which
takes place in the PF component (Bobaljik 2009) – or indeed whether it
involves two suboperations – one which sets up an agreement link be-
tween probe and goal in the syntax, and another that copies the relevant
φ-features in the PF component (Benmamoun et al. 2009; Arregi & Nevins
2012; Bhatt & Walkow 2013; Smith 2015, Kalin 2020).
210 Agreement

3.7 Workbook

In the exercises below, an [M] after an example sentence indicates that a

model answer is provided in the free-to-download Students’ Answerbook
(and in the Teachers’ Answerbook as well). You will find it helpful to look at
this model answer before attempting to tackle other examples in the exer-
cise. Some other exercise examples have an [S] after them, indicating that
they are for self-study: this means that after reading the relevant module
(and those preceding it), and looking at the helpful hints and model answer
associated with the exercise, students should be able to analyse these self-
study examples on their own, and then compare their answers with those
provided in the Students’ Answerbook. The remaining exercise examples
(i.e. those not marked with [M] or [S] after them) are intended for teachers
to use as the basis for hands-on problem-solving work in seminars, classes,
or assignments: answers to all of these non-self-study examples are pro-
vided in the Teachers’ Answerbook. Note that Exercise 3.1 tests you on
Module 3.1, Exercise 3.2 on Module 3.2, and so on; and that each exercise
in a given chapter presupposes familiarity with material covered in earlier
chapters and in earlier modules of the same chapter. Unparenthesised num-
bers like 4 refer to material in the exercises, whereas parenthesised numbers
like (22) refer to examples and conditions in the main text of the book.

EXERCISE 3.1

This exercise is designed to be attempted after reading Module 3.1 and the
chapters preceding it.

Discuss how Agreement, Case-marking and A‑Movement work in the

following sentences, focusing in each case on the syntax of the italicised
(pro)nominal constituents and the bold-printed (auxiliary/main) verbs:

1 He is telling the truth [M]

2 No traces of any DNA were found [S]
3 We workers are being exploited
4 Several new candidates have emerged [S]
5 The economy has grown considerably
6 The guards deny they tortured the prisoners

Helpful hints
Bear in mind the discussion of the VP-Internal Subject Hypothesis (and
the syntax of unaccusative/passive predicates) in the previous chapter,
and the role played by the Activity Conditions (5) in the main text/
 3.7 Workbook 211

summary Module 3.5, the Feature Valuation Conditions (6) in the main
text/summary, the EPP Condition (7) in the main text/Summary, and
the Case Conditions (11) in the main text/summary. In dealing with
agreement, concern yourself with the case, agreement, and movement
properties of goals, but not their internal structure: for example, take
several new candidates in 4 to be a QP, but don’t concern yourself with
its internal structure. In the case of complex QPs or DPs, assume that the
case/agreement properties of the overall phrase are those of its head, since
phrases are projections of their heads. For the purposes of this exercise,
make the simplifying assumption that finite auxiliaries originate in the
head T position of TP (setting aside the possibility that they may originate
in a lower auxiliary projection and from there move to T via Head
Movement). In the case of 5, assume that considerably is an ADVP which
adjoins to the VP headed by grown to form a larger VP.

EXERCISE 3.2

This exercise is designed to be attempted after reading Module 3.2 and the
module and chapters preceding it.

Account for the (un)grammaticality of the following expletive sentences,

paying particular attention to the syntax of Agreement and Case-marking:

1 It is said that he was taking bribes [M]

2 *There is said that he was taking bribes [M]
3 There has emerged a new strain of the virus [S]
4 *It has emerged a new strain of the virus [S]
5 It was raining heavily [S]
6 *There was raining heavily [S]
7 It does not matter that he lied
8 *There does not matter that he lied
9 There have been some changes to on-street parking
10 There has been some changes to on-street parking ([Link])
11 In her whole life, there has been only me (M. A. Graham, A Golden Forever,
Google Books, 2009: 94)
12 *In her whole life, there have been only I
13 It appears there was some unrest

Helpful hints
Bear in mind the VP-Internal Subject Hypothesis (and the syntax of
unaccusative predicates), and the role played by conditions like the
Activity Conditions, Valuation Conditions, EPP Condition, and Case
212 Agreement

Conditions outlined in (5), (6), (7) and (11) in the main text/Summary in
regulating how Agreement and Case-marking apply. For the purposes of
this exercise, make the following assumptions (but don’t concern yourself
with the internal structure of the relevant XPs): in 1, 2, 7, 8, the that-
clause is a CP; in 5, 6, heavily is an ADVP which adjoins to the VP headed
by raining to form an even larger VP; in 7, 8 not is an ADVP which serves
as the specifier of a NEGP with a null NEG head; in 9, 10, some changes to
on-street parking is a QP; and in 11, 12, in her whole life is a PP directly
merged in spec‑CP. Note that the use of has in 10 is generally taken to
be non-standard (with Standard English preferring to use have, as in 9):
consider the possibility that people who produce non-standard sentences
like 10 may treat expletive there as inherently third person singular. In
relation to 10–12, bear in mind that the verb b e can seemingly assign
accusative case to its complement in sentences like ‘If you were me, what
would you do?’ Bear in mind, too that (as noted in the book) a phrase like
only me can function as a third person singular expression in a sentence
like ‘Only me is working in the office at present’ (where only me is
paraphrasable as ‘nobody but/except me’).
In dealing with agreement, concern yourself with the overall case
and agreement properties of goals, but not with their internal structure:
for example, in the case of complex QPs or DPs, assume that the case/
agreement properties of the overall phrase are those of its head, since
phrases are projections of their heads. For the purposes of this exercise,
make the simplifying assumption that finite auxiliaries originate in the
head T position of TP (setting aside the possibility that they may originate
in a lower auxiliary projection and from there move to T via Head
Movement).
Bear in mind the following assumptions about expletive clauses made
in Module 3.2:

• expletive it enters the derivation with inherently valued third-person

and singular-number features and an unvalued case feature; its person/
number features value those of an affix/auxiliary in T
• expletive there enters the derivation with an inherently valued third
person feature and an unvalued case feature; the expletive and its as-
sociate jointly value the unvalued features on an affix/auxiliary in T
• expletive subjects (like other subjects) originate within VP, and from
there are attracted by the EPP feature on T to move to spec‑TP.
• case on an expletive subject (and any associate it has) is valued via
agreement with a complementiser
 3.7 Workbook 213

EXERCISE 3.3

This exercise is designed to be attempted after reading Module 3.3 and the
modules and chapters preceding it.

Discuss the syntax of the infinitive structures in the sentences below,

focusing on the syntax of the constituents involved in Agreement,
A‑Movement and Case-marking:
1 The suspect is expected to be charged [M]
2 They haven’t decided whether to charge the suspect [S]
3 *The suspect was decided to be charged (intended to be synonymous with ‘It
was decided to charge the suspect’) [S]
4 They don’t want it to rain [S]
5 It was said to be raining [S]
6 The enquiry found there to be no evidence of any malpractice [S]
7 There was found to be no evidence of any malpractice
8 He was thought to be the ringleader
9 *The chairman has been impossible for to attend the meeting (intended to
be synonymous with ‘It has been impossible for the chairman to attend the
meeting’)
10 It/*There is unlikely for there/*it to be any protests
11 There needs to be some changes (James Moore, [Link])
12 They were hoping for there to be some changes

In the case of 12 set aside the agreement-based account of movement

and case-marking outlined in the book, and instead follow the alternative
(agreement-free) account of movement and case-marking sketched in the
helpful hints below.

Helpful hints
In relation to the sentences in 1–11, say how the derivation of these
sentences would proceed if we adopt the agreement-based account of
Case-marking and A‑Movement outlined in this chapter. Accordingly
assume (as claimed in Module 3.3) that infinitival to enters the
derivation carrying an interpretable nonfinite tense feature [Nf-Tns], an
uninterpretable (and unvalued) person feature [u-Pers], and an EPP feature
which attracts the closest XP goal it agrees with in person to move to
spec‑TP, and is subsequently stripped from T‑to (and thereby removed from
the derivation entirely) because it is illegible at the PF and LF interfaces.
Bear in mind the VP-Internal Subject Hypothesis (and the syntax of
unaccusative and passive predicates), and the role played by conditions
like the Activity Conditions, Valuation Conditions, EPP Condition, and
214 Agreement

Case Conditions outlined in (5), (6), (7) and (11) in the main text/Summary
in regulating how Agreement and Case-marking apply. In the case of
2, concern yourself only with the syntax of the infinitival whether-
clause. In relation to 11, note that the use of the singular form needs is
generally taken to be non-standard (with Standard English preferring to
use the plural form need); see the suggestions about singular agreement
in sentences where there has a plural associate in the helpful hints for
example 10 in Exercise 3.2. In dealing with agreement, concern yourself
with the overall case and agreement properties of goals, but not with their
internal structure. For the purposes of this exercise, make the simplifying
assumption that finite auxiliaries originate in the head T position of TP
(setting aside the possibility that they may originate in a lower auxiliary
projection and from there move to T via Head Movement).
In relation to the sentence in 12 consider how its derivation would
work if we reject the agreement-based account of movement and case-
marking in the book (because it leads to a panoply of abstract agreement
features and operations which have no overt manifestation and thus little
empirical rationale), and instead explore an alternative agreement-free
account of movement and case-marking based on the revised assumptions
outlined in 13, 14 below:

13 A‑Movement is triggered not by person agreement nor by an EPP feature

on T, but rather by an [Attr-Pers] feature on a T‑head, meaning ‘Attract the
closest XP goal with a person feature’. This attraction feature is stripped from
T (and thereby removed from the derivation entirely) once it has done its
work, because it is illegible at PF and LF
14 Case-marking does not involve agreement, but rather a case-assignment
feature on a probe such as [Ass-Nom]/[Ass-Acc]/[Ass-Null] meaning ‘Assign
nominative/accusative/null case to any local/nearby XP goal(s) with an unvalued
case feature’; this case-assignment feature is stripped (i.e. removed from the
derivation) once it has done its work, because it is illegible at PF and LF

EXERCISE 3.4

This exercise is designed to be attempted after reading Module 3.4 and the
modules and chapters preceding it.

Discuss the syntax of the sentences in 1–8 below, saying why 2, 3, 4, 5, 8

are considered non-standard; in addition, discuss the case-marking of the
italicised items in the bracketed finite clauses in 9, 10:

1 It seems there were no arrests [M]

2 There seem were some problems … ([Link])
 3.7 Workbook 215

3 There seem there are two extremes ([Link]) [S]

4 There seems there are two parts ([Link])
5 You seem that are still missing the point ([Link])
6 The mayor looks like he has been re-elected [S]
7 Everybody looks like they have passed the exam
8 Certain groups are likely for to be under-represented ([Link]) [S]
9 He is someone [who/whom I think is making progress] [S]
10 I think [he/*him is making progress]

Helpful hints
Examples 1–8 are all complex sentences comprising a main clause and
a complement clause; in relation to these, consider the following three
issues:

(i) Is the complement clause a TP or a CP? If it is a CP, is the complementiser

heading the CP defective or non-defective?
(ii) Does the subject undergo Raising from spec‑TP in the complement clause
into spec‑TP in the main clause, and if so does the subject leave behind
an overt copy (and if so, how and why?). Or do the main and complement
clauses have separate subjects?
(iii) How do Agreement, A‑Movement and Case-marking work in these
structures?

Bear in mind that while for most speakers, expletive there is specified for
(third) person but unspecified for number, some speakers can treat it as
inherently third person singular. In relation to 6, 7, assume that perception
verbs like look are ambiguous between a thematic and non-thematic use.
In its thematic use, look theta-marks its subject and has an appearance
interpretation (so that 6 has a meaning paraphrasable as ‘The mayor has
an appearance suggesting he has been re-elected’ – for example, the mayor
has a broad smile on his face). By contrast, in its non-thematic use, look
does not theta-mark its subject (so allows an expletive subject as in ‘It looks
like there are problems’, or an idiomatic subject as in ‘Close tabs look like
they are being kept on the FBI’) and has a probability interpretation (so
that 6 is paraphrasable as ‘It looks like/It is probable that the mayor has
been re-elected’). For the purposes of this exercise, assume that like in 6, 7
is a complementiser, and adopt the binding analysis of look like structures
outlined in the book (adapted from Lappin 1984). In 8, take certain groups
to be a QP headed by the quantifier certain, and consider whether for is a
(defective or non-defective?) complementiser, or whether (in the relevant
variety) it is fused with the infinitive particle to, forming the complex infin-
itive particle (i.e. infinitival T‑head) forto.
216 Agreement

In relation to the bracketed relative clauses in 9, 10, set aside the

agreement-based account of movement and case-marking outlined in the
book, and instead make the following assumptions (outlined in the helpful
hints for Exercise 3.3):

11 A‑Movement is triggered not by person agreement nor by an EPP feature

on T, but rather by an [Attr-Pers] feature on a T‑head, meaning ‘Attract the
closest XP goal with a person feature’. This attraction feature is stripped from
T (and thereby removed from the derivation entirely) once it has done its
work, because it is illegible at PF and LF
12 Case-marking does not involve agreement, but rather a case-assignment
feature on a probe such as [Ass-Nom]/[Ass-Acc]/[Ass-Null] meaning ‘Assign
nominative/accusative/null case to any local/nearby XP goal(s) with an unvalued
case feature’; this case-assignment feature is stripped (i.e. removed from the
derivation) once it has done its work, because it is illegible at PF and LF

In addition, in 9 assume that who originates as the subject of making

progress, then becomes subject of is, then moves to the edge of the CP in
the making-clause (shown as CP1 below), and finally to the edge of the CP
in the think-clause (= CP2) – leaving behind copies in the positions itali-
cised below (where Af denotes a present-tense affix):

13 He is someone

[CP2 who [C2 ø] [TP I [T Af] think [CP1 who [C1 ø] [TP who [T is] who making progress]]]]

Assume that C1 and C2 carry an [Attr-Wh] feature enabling them to attract

an XP containing a wh‑word to become their specifier, and that this feature
(being illegible at the PF and LF interfaces) is stripped from C (and removed
from the derivation entirely) once it has triggered Wh‑Movement. Consider
the possibility that C1 may carry an optional [Ass-Nom] case feature which
(if present) assigns nominative case to the subject of the clause it intro-
duces, and that the verb think can carry an optional [Ass-Acc] feature
which (if present) enables think to assign accusative case to a local/nearby
XP with an unvalued case feature. Take who to be a relative wh‑pronoun (a
pronominal DP) which enters the derivation carrying the features [3‑Pers,
Sg‑Num, u‑Case] – plus an additional feature marking it as animate in
gender (in contrast to inanimate which), though you can ignore this gender
feature in your discussion, as it plays no role in agreement. Account for
why the subject of is making progress can be spelled out as either nom-
inative who or accusative whom in 9, but only as nominative he (not as
accusative him) in 10.
4 The clause periphery

4.0 Overview

As we saw in Chapter 1, the classic analysis of clauses in the 1980s took

them to comprise a VP headed by a lexical verb, and assumed that on top of
VP there were two different types of functional projection: a TP containing
a tense-marked auxiliary/affix/infinitive particle and its subject, and a CP
containing a complementiser and other peripheral constituents (e.g. fronted
constituents, clausal modifiers and so on). However, more recent research
has suggested CP and TP should each be ‘split’ into a number of distinct
functional projections. In this chapter and the next, we look at evidence
that the clause periphery (i.e. that part of the clause structure above TP
which is traditionally labelled CP) can be split up into a range of differ-
ent types of projection, including force phrase/FORCEP, topic phrase/TOPP,
focus phrase/FOCP, modifier phrase/MODP and finiteness phrase/FINP pro-
jections. In Chapter 6, we’ll look at evidence that the clause subperiphery
(i.e. the TP constituent housing a T auxiliary/affix/infinitive particle and
its subject) can likewise be split up into a number of different functional
projections. This approach has been termed ‘Cartographic’, since its aim is
to devise a ‘map of the left periphery’ (Rizzi 1997: 282).

4.1 The Cartographic approach

The key insight underlying the Cartographic approach to splitting up the

clause periphery into a number of distinct types of projection is that each
different type of peripheral constituent occupies a unique position on the
edge of a dedicated functional projection (i.e. a functional projection which
is dedicated to housing a constituent of the relevant type on the edge of
the projection). This implies (inter alia) that a peripheral topic will appear
on the edge of a topic projection, a peripheral focused constituent on the
edge of a focus projection, a peripheral relative constituent on the edge
of a relative projection … and so on. Rizzi (2014b) argues that the pos-
tulation of a range of different types of peripheral head hosting different
types of constituent gains cross-linguistic support from structures in other
218 The clause periphery

languages like those bracketed below which contain a bold-printed periph-

eral head with an italicised specifier (where Q denotes a question particle,
TOP denotes a topic particle, FOC denotes a focus particle, REL denotes a
relative particle, and EXCL denotes an exclamative particle):

(1) a. Ik weet niet [wie of [Jan gezien heeft]]

I know not who Q Jan seen has
‘I don’t know who Jan has seen’
(Dutch, Haegeman 1994)
b. Un sè do [dan lo yà [Kofi hu ì]]
I heard that snake the TOP Kofi killed it
‘I heard that the snake, Kofi killed it’
(Gungbe, Aboh 2004)
c. Un sè do [dan lo wè [Kofi hu]]
I heard that snake the FOC Kofi killed
‘I heard that the snake, Kofi killed’
(Gungbe, Aboh 2004)
d. Der Mantl [den wo [dea Hons gfundn hot]]
The coat which REL the Hans found has
‘The coat which Hans has found’
(Bavarian, Bayer 1984b)
e. [Che bel libro che [ho letto]]!
What nice book EXCL [Link] read
‘What a nice book I read!’
(Italian, Rizzi 2014b)

In each case the head of the relevant bracketed peripheral projection is the
bold-printed particle, the italicised constituent preceding the head is its spe-
cifier, and the bracketed constituent following the head is its complement.
The head of each peripheral projection assigns its specifier (if any) and
complement a specific interpretation at the semantics interface: for in-
stance, the head of the topic projection tells the semantic component that
‘my specifier is to be interpreted as the topic, and my complement as the
comment’ (Cinque & Rizzi 2010a: 51). In a similar way, a focus head ‘as-
signs the focus interpretation to its specifier, and the presupposition in-
terpretation to its complement’ (Cinque & Rizzi 2010a), where the focus
constitutes ‘new information, information assumed not to be shared by the
interlocutor’ (Rizzi 2014b: 8). A key assumption in this approach is that (in
conformity with the Linear Correspondence Axiom of Kayne 1994), no per-
ipheral projection can have more than one specifier. For obvious reasons,
this has been dubbed the ‘split CP’ analysis, because it involves splitting
what had previously been treated as a single type of peripheral constituent
 4.1 The Cartographic approach 219

(= CP) into a number of different types of projection (a topic phrase, focus

phrase etc.), each housing a different type of peripheral constituent.
Under this Cartographic approach, the head of each peripheral projection
is dedicated to housing a particular type of constituent on the edge of its
projection. This assumption follows from a featural uniqueness principle
posited by Cinque & Rizzi (2008), which can be outlined informally as fol-
lows (where an interpretable feature is one which contributes to semantic
interpretation/meaning):
(2) One Feature One Head Principle
Each functional head carries only one interpretable feature

It follows from (2) that no head can mark (for example) both force and
topic, or both focus and finiteness.
Constituents occupying the specifier position of a peripheral projection do
so in order to satisfy a requirement for them to occupy a criterial position in
a spec–head relation with a matching head (Rizzi 1997: 282): for example,
a peripheral topic must occupy the specifier position in a topic projection in
order to be in a spec–head relation with a topic head. Once a constituent is
in its criterial position, it is thereby frozen in place in consequence of a con-
straint termed the Criterial Freezing Condition in Rizzi (2005, 2006a, 2010,
2014a, 2014b) and Rizzi & Shlonsky (2006, 2007). For present purposes,
these requirements can be subsumed within the following condition:
(3) Criterial Freezing Condition/CFC
(i) A peripheral constituent must occupy a criterial position to be
interpretable at LF
(ii) A constituent which occupies its criterial position is frozen in place

Condition (3i) tells us that (for example) a peripheral constituent must end
up (at the end of the syntactic derivation) as the specifier of a topic projec-
tion in order to be interpretable as a topic in the semantic component; and
(3ii) specifies that any constituent occupying its criterial position is frozen
in place (and cannot move).
By way of an introduction to this approach, we’ll start by looking at the
rationale for positing Force and Topic projections.

4.1.1 Force and topic projections

As a starting point for our discussion, consider the structure of the brack-
eted clause produced by speaker b in the dialogue below:
(4) s p ea k e r a : The demonstrators have been looting shops and setting fire
to cars
s p ea k e r b : They must know [that this kind of behaviour, we will not

tolerate it]
220 The clause periphery

Here, the italicised phrase this kind of behaviour is the topic of the brack-
eted embedded clause in (4 b ), where a topic is a constituent which tells
us what the clause is about, and conveys old/familiar information known
to the speaker and hearer. In this case, the italicised topic refers back to
the activity of looting shops and setting fire to cars which is old/familiar
information by virtue of having been mentioned earlier by speaker a . But
whereabouts is the topic positioned within its containing clause?
Since topics can be phrases (e.g. the topic this kind of behaviour is a
DP in 4 b ) and since phrases can be specifiers, a plausible analysis is that
topics occupy the specifier position in a peripheral projection which is
positioned above the TP headed by will but below the CP headed by that.
But what is the nature of this peripheral projection? One possibility is
that it is a CP, and hence that the topic occupies spec‑CP. However, since
the italicised topic in (4 b ) is contained within a projection headed by the
complementiser that, a CP analysis requires us to assume that the brack-
eted embedded clause contains two separate CP projections – an outer
one containing the complementiser that, and an inner one containing the
topic this kind of behaviour. On this view, the bracketed embedded clause
will involve CP recursion (i.e. one CP stacked on top of another), and the
embedded clause bracketed in (4 b ) will have the structure shown in sim-
plified form below:

(5) CP

C CP
that
DP C′
this kind of
behaviour C TP
ø we will not
tolerate it
The advantage of a CP recursion analysis is that it accommodates topic
structures within the traditional CP+TP+VP model of clause structure,
without the need to posit additional types of projection. The disadvantage,
however, is that in treating the projection containing the complement-
iser that and the projection containing the topic this kind of behaviour as
exactly the same type of CP projection, the CP recursion analysis fails to
recognise important differences between the two types of CP. For example,
the higher CP headed by that has the semantic function of marking the
embedded clause as declarative in force/type, while the lower CP serves
to mark its specifier this kind of behaviour as the topic of the sentence.
 4.1 The Cartographic approach 221

Moreover, (contrary to what a CP recursion analysis might lead us to

expect), the two CPs are not interchangeable, in that the CP containing the
topic cannot be positioned above the CP containing that – as we see from
the ungrammaticality of:

(6) *They must know [CP this kind of behaviour [CP that [TP we will not
tolerate it]]]

These differences suggest that the two peripheral projections labelled CP in

(5) above are not simply two different instances of the same CP projection,
but rather constitute two distinct types of projection.
One way of capturing this (pioneered in influential Cartographic work by
Luigi Rizzi in the mid-1990s cited in the Bibliographical notes) is to split
CP up into a number of different types of functional projection. On this
‘split CP’ approach, the projection housing the complementiser that (which
serves to mark the clause as declarative in force and thus having the force
of a statement) is categorised as a force phrase (FORCEP), and its head com-
plementiser that is categorised as a FORCE head. By contrast, the projection
which serves to mark the DP this kind of behaviour as a topic is categorised
as a topic phrase (= TOPP), and is headed by a null topic particle (= TOP).
On this alternative approach, the embedded clause bracketed in (4 b ) does
not have the CP recursion structure in (5) above, but rather the split projec-
tion structure in (7) below (so called because CP has been split into distinct
FORCEP and TOPP projections):

(7) FORCEP

FORCE TOPP
that
DP TOP′
this kind of
behaviour TOP TP
ø we will not
tolerate it
The idea that the phrase housing the topic is a topic projection headed by
a topic particle (albeit one which is null in English) gains cross-linguistic
plausibility from the observation that there are a number of languages which
have an overt topic particle which is positioned after a topic – including
Gungbe, as we saw in (1b) above. This adds cross-linguistic plausibility to
the idea that topics are the specifiers of a topic phrase (= TOPP) projection
that is headed by a topic particle which is overt in languages like Gungbe,
but null in languages like English.
222 The clause periphery

4.1.2 In situ and ex situ topics

In a split projection structure like (7) above, the constituent which serves as
the specifier of the TOPP projection functions as a topic on which the TP
constituent makes a relevant comment. Since the topic is directly merged
in situ as the specifier of a TOPP projection, we can refer to it as an in situ
topic. In (7), the relevance of the comment to the topic is ensured by using
the resumptive pronoun it to ‘resume’ or ‘reprise’ (i.e. refer back to) the
topic this kind of behaviour, so ensuring that the comment in TP is rele-
vant to the topic in TOPP: accordingly, we can refer to this as a resumptive
topic structure. In topic structures containing a resumptive pronoun, the
topic is conventionally said to be dislocated, and a structure like (4 b , 7)
where the topic is to the left of the comment is said to be an instance of
Left Dislocation.
Although our example in (4 b ) involves a topic in spec‑TOPP being re-
prised by a resumptive personal pronoun (it), examples like those in (8)
below show that a variety of other types of (bold-printed) resumptive ex-
pression can be used to reprise an (italicised) in situ topic:

(8) a. High speed corners with a dry line, that’ll burn your tyres out (Martin
Brundle, BBC1 TV)
b. The Everton fans, there’s lots of different opinions here (Danny Kelly,
talkSPORT Radio)
c. Any Premier League team, take the three best players out of the team
and they’re gonna struggle (Dan Walker, BBC Radio 5)
d. Many of them, I have no idea where these teams are (Danny Kelly,
talkSPORT Radio)
e. The Championship, one of the greatest things about that league is
that it’s the most unpredictable league around (Listener, BBC Radio 5)

Thus, it would be more accurate to say that Topicalisation involves an in

situ topic reprised by a resumptive nominal or pronominal constituent.
Alongside resumptive topic structures like (4 b , 8), we also find topic
structures like (9b ) below which have a gap (marked by — ) in the position
occupied by resumptive it in (4 b ):

(9) speaker a : The demonstrators have been looting shops and setting fire
to cars
speaker b : They must know [that this kind of behaviour, we will not

tolerate — ]

There is a gap in (9 b ) in the sense that the transitive verb tolerate has no
complement/object immediately following it; in fact, it is the topic this kind
of behaviour which is interpreted as the object of tolerate. Since gaps can
arise via movement, a plausible analysis of (9b ) is to suppose that the DP
 4.1 The Cartographic approach 223

this kind of behaviour originates as the thematic complement of tolerate and

then (in order to mark it as a topic) moves to become the specifier of a TOPP
projection in the clause periphery. On this view, Topicalisation in the embed-
ded clause in (9 b ) involves the Topic Fronting operation arrowed below:

(10) FORCEP

FORCE TOPP
that
DP TOP′
this kind of
behaviour TOP TP
ø we will not tolerate ––

We can then suppose that topics are ex situ (i.e. displaced/fronted) constitu-
ents which move to spec‑TOPP in gap-containing topic structures like (9 b ,
10), but are in situ constituents directly merged in spec‑TOPP in gapless
topic structures like (4 b , 7). The overall complement clause is typed as
declarative in force by virtue of being a FORCEP headed by the indicative
complementiser that.
There is independent evidence that ex situ topic structures (= those con-
taining gaps) involve movement, whereas in situ topic structures do not.
One such piece of evidence comes from the observation that ex situ/gap
structures are subject to the Island Constraint of Ross (1967, 1986), where-
as in situ/gapless structures are not. Ross’s constraint can be characterised
informally as follows, where an island is a structure resistant to extraction
by movement operations which leave a gap behind, a gap is a null copy of
a moved constituent, and the antecedent of a gap is the moved constituent
which leaves the null copy behind:

(11) Island Constraint

No gap inside an island can have an antecedent outside the island

As we saw in §1.5.1, Ross identified a number of different types of constit-

uent as islands, including complex NPs (comprising an NP with a clause
embedded inside it). In the light of the assumption that complex NPs are
islands, consider the following contrast (where a bracketed relative clause
CP is contained inside a bracketed NP; a detail that has been set aside
below is that the laws-NP is modified by a null quantifier which denotes
‘an unspecified number of’):

(12) a. *Violent criminals, we need [NP laws [CP that send — to prison for life]]
b. Violent criminals, we need [NP laws [CP that send them to prison for
life]]
224 The clause periphery

On the assumption that the gap-containing topic structure in (12a)

involves movement (with the italicised topic originating in the gap position
marked — and subsequently moving from there to the italicised specifier
position in TOPP), we correctly predict that (12a) will be ungrammatical
because the moved topic has left behind an illicit gap inside the bracketed
complex NP island (i.e. inside a CP contained within an NP). Conversely, on
the assumption that no movement is involved in the resumptive structure
(12b) (because both the italicised topic and the bold-printed resumptive
pronoun are directly merged in situ), we correctly predict that resumptive
topic structures like (12b) will be grammatical, because there is no gap
inside the bracketed complex NP in (12b).
Thus far, we have seen two different types of topic structure: in situ
structures where the topic is associated with a resumptive, and ex situ
structures where the topic is associated with a gap. However, in colloquial
English, we also find a third type of topic structure involving an orphaned
topic which is not associated with any resumptive or gap – as the examples
in (13) below illustrate (where the orphaned topic is italicised):

(13) a. England, I agree with you completely on Capello (Listener, BBC

Radio 5)
b. Wales, it’s just gonna rain all day (Lisa O’Sullivan, talkSPORT Radio)
c. Your pension, you’ve been told you’ve got to pay more (Interviewee,
BBC Radio 5)
d. Defoe, even I could have scored that goal (Alan Green, BBC Radio 5,
boasting that even he could have scored a goal that Defoe missed)
e. Germany, I think Portugal were unlucky (Alan Brazil, talkSPORT
Radio, commenting on a match where Portugal were unlucky to be
beaten by Germany)

The examples in (13) contain an italicised topic in the periphery of the

clause, but the comment TP (after the comma) contains no resumptive or
gap linked to the topic. The topic typically has an as for interpretation: for
example, the topic England in (13a) has the paraphrase ‘As for England,
I agree with you completely on Capello’. The crucial requirement in topic
clauses is that the comment made in the TP following the topic should be
relevant to the topic: one way of ensuring relevance is by use of a resump-
tive or gap in the TP commenting on the topic; however, it is clear from
examples like those in (13) that relevance does not require the use of a
resumptive or gap (at least, in colloquial English).
On the assumptions made here, gap-containing topic structures like
(10) above involve movement, but gapless topic structures like (8, 13)
do not. Further evidence in support of this claim comes from case-
marking – as can be illustrated in relation to the replies given by speaker
 4.1 The Cartographic approach 225

b in the dialogue below (where capitals mark contrastive stress, and

strikethrough marks ellipsis):

(14) s p ea k e r a : What do you think about John and Mary?

s p ea k e r b : (i) he , I don’t think [CP [C ø] — is guilty]]

(ii) But hi , I suspect [CP [C that] he probably is guilty]

Since case is preserved under movement (in the sense that a moved con-
stituent retains any case that it was assigned before it moved), it follows
that the topic she in the gap-containing structure in (14 b i) carries the nom-
inative case it received in the gap position from the null complementiser
before moving to become the specifier of a topic phrase projection in the
clause periphery. By contrast, the fact that the topic him in (14 b ii) carries a
different (accusative) case from that carried by the nominative resumptive
pronoun he, provides empirical evidence that no movement is involved
in gapless topic structures. On one (widespread) assumption, an in situ
topic like him in (14 b ii) receives default case – that is, the case assigned
to a (pro)-nominal constituent which occupies a position where it does not
fall within the domain of any case assigner. The default case in English is
accusative (though in some other languages like German and Arabic, for
example, it is nominative). For present purposes, we can characterise this
phenomenon for English as follows:

(15) Default Case Assignment

A (pro)nominal which does not fall within the domain of (i.e. which
is not c‑commanded by) a local (i.e. nearby) case assigner receives
accusative case by default

Other types of constituent which have been argued to be assigned accusa-

tive case by default include those italicised below:

(16) a. Who wants an ice-cream? – Me!

b. She’s a bit taller than me
c. It was me that reported the incident
d. If you were me, what would you do?
e. Me cheat on you? No way!
f. Me being lazy, I ordered a takeaway

For more detailed discussion of default case, see Schütze (2001).

4.1.3 Multiple topic structures

All of the topic structures we have looked at so far have contained only a
single topic. However, sentences can in principle contain more than one
topic, as the following authentic examples of double topic structures illus-
trate (where the first topic is bold-printed and the second is italicised):
226 The clause periphery

(17) a. You just get the feeling that Arsenal, the way they keep the ball, it’s
particularly clever (Steve Claridge, BBC Radio 5)
b. Everton, Saha, he did look quite sharp (Danny Mills, BBC Radio 5)
c. A good friend of mine, his daughter, I promised to go and say hello
to her today (Alan Brazil, talkSPORT Radio)
d. The people on the street, the demands we are aware of, they have a
legitimacy (Alistair Burt, BBC Radio 5)

If each topic is housed in a separate topic projection, sentences like (17) will
involve TOPP recursion (i.e. a structure in which the periphery of a single clause
contains more than one TOPP projection). On this view, the periphery of the
that-clause in (17a) will have the following structure (where Arsenal is a DP
headed by a null D, and indeed the club is sometimes referred to as the Arsenal):

(18) FORCEP

FORCE TOPP
that
DP TOP′
ø Arsenal
TOP TOPP
ø
DP TOP′
the way they
keep the ball TOP TP
ø it’s
particularly
clever

Both topics highlighted in (18) are directly merged in situ (and assigned
accusative case by default), and both are reprised by a resumptive pro-
noun – the higher topic Arsenal being reprised by they, and the lower topic
the way they keep the ball being reprised by it.
The TOPP recursion analysis of multiple topic structures gains inde-
pendent motivation from the observation by Rizzi (2012) that languages
like Abidji which allow topic recursion use a separate topic particle (bold-
printed below and glossed as t o p ) to mark each (italicised) topic – as in
double topic sentences such as the following, where either highlighted topic
can be positioned in front of the other:

(19) a. kòfí έkέ òkókò έ έkέ è pìpjé nɩ̀

Kofi top banana the t o p he peel it
‘Kofi, the banana, he peeled it’
b. òkókò έ έkέ kòfí έkέ è pìpjé nɩ̀
banana the top Kofi t o p he peel it
‘The banana, Kofi, he peeled it’
 4.1 The Cartographic approach 227

This makes it plausible to suggest that English multiple topic structures like
(18) also involve TOPP recursion (i.e. they involve projecting more than one
topic phrase constituent within the periphery of a single clause).
However, while gapless multiple topic structures like (18) above are
grammatical, by contrast gap-containing/ex situ multiple topic structures
are ungrammatical – as illustrated by the contrasts below (attributed to
Rachel Nye and Ian Roberts in Rizzi 2013a: 214):

(20) a. John, I convinced — to buy your car

b. Your car, I convinced John to buy —
c. *John, your car, I convinced — to buy —
d. *Your car, John, I convinced — to buy —

As we see from (20a, 20b), either John or your car can freely be topicalised
by being moved into the specifier position in a topic projection. However,
we can’t topicalise both John and your car, as we see from the ungrammat-
icality of (20c, 20d). Why should this be? The answer lies in a constraint
(mentioned in §1.5.1) which was formulated by Abels (2012: 247) as fol-
lows (where likes denote ‘constituents which are alike’ – that is, constitu-
ents of the same type):

(21) Intervention Condition

Likes cannot cross likes

From this perspective, what goes wrong in (20c, 20d) is that one (bold-
printed) topic moves to the edge of an inner TOPP projection, and then
another (italicised) topic moves to the edge of an outer TOPP projection,
and in the process the (italicised) outer topic crosses the (bold-printed)
inner topic, in violation of the Intervention Condition (21). On this view,
one moved topic cannot cross another moved topic.
A further constraint which may cause structures like (20c, 20d) to be
degraded is the Distinctness Condition of Richards (2010: 5), which we
can characterise informally as follows (where non-distinct constituents are
constituents which are alike, and ‘at PF’ means ‘at the end of the syntactic
derivation, at the point where syntactic structures are inputted into the PF
component for processing’):

(22) Distinctness Condition

Any structure at PF which contains a sequence of two non-distinct
constituents <α, α> is degraded

It follows from the constraint in (22) that the succession of two highlighted
topicalised DPs in (20c, 20d) leads to degraded grammaticality.
Emonds (2004: 107) claims (on the basis of contrasts like that shown in
slightly adapted form below) that the clause periphery can contain two
228 The clause periphery

different types of topic (an in situ and an ex situ topic), but that when the
two co-occur, an (italicised) in situ topic must precede a bold-printed ex
situ/fronted topic:

(23) a. My mother, a man like that, I don’t think she would hire —
b. *A man like that, my mother, I don’t think she would hire —

This restriction can again be argued to follow from the Intervention

Condition. In (23a) the italicised outer topic is directly generated in situ,
and so does not move across the (bold-printed) inner topic. By contrast, in
(23b) the italicised ex situ topic moves across the bold-printed in situ inner
topic, and the resulting movement of one topic across another leads to an
intervention violation (though for me at least, the violation feels stronger
where one fronted topic crosses another fronted topic, as in 20c and 20d;
this suggests that the topics are more alike if both are fronted, rather than
if one is fronted and the other is not).
You should now be able to tackle Exercise 4.1.

4.2 Focus projections

In Module 4.1, we saw that the clause periphery can comprise two distinct
projections, one housing a complementiser (FORCEP), and another housing
a topic (TOPP). In this module, we will see that the clause periphery can
also contain a third type of functional projection – namely a focus phrase/
FOCP projection containing a focus (i.e. focused/focalised constituent) as
its specifier.

4.2.1 Peripheral focused constituents

From a discourse perspective, a focused constituent represents new infor-
mation (i.e. information not previously mentioned in the discourse and
assumed to be unfamiliar to the hearer). Because focusing introduces new
information and wh‑questions are typically used to ask for new informa-
tion (e.g. Who did you see? asks for the identity of the person you saw),
focusing can be used in replies to wh‑questions. By way of illustration,
consider the following dialogue:

(24) speaker a : How many goals did Cristiano Ronaldo score for Real
Madrid?
speaker b : Would you believe [that 451 goals he scored — in 438

games]?!
 4.2 Focus projections 229

Here, the italicised phrase produced by speaker b originates in the gap pos-
ition as the complement of the verb scored, and is then moved to the per-
iphery of the bracketed clause, in order to focus it (i.e. mark it as conveying
new information). Some authentic examples of focusing (which I recorded
from a variety of radio and TV programmes) are given below, where the
focused constituent is italicised:

(25) a. We told you there were going to be goals, and goals there were (Jake
Humphreys, BT Sport TV) (21)
b. Straight up the other end Crusaders went (Ian Abrahams, talkSPORT
Radio)
c. Another good cross he delivers (Rob Palmer, Sky Sports TV)
d. Excellent football that was (Ray Wilkins, Sky Sports TV)
e. Very relaxed they were in the interview, as far as I could see (Sir
Ming Campbell, BBC Radio 5)

But whereabouts in the clause periphery are focused constituents posi-

tioned, and how do they get there?
A relevant observation to make in this regard is that there are languages
in which a focused peripheral constituent is followed by an overt focus
particle – as we see from the examples in (26) below (where the focused
constituent is italicised, and the focus particle is bold-printed and glossed
as foc ; note that p r t denotes a particle, ag r an agreement marker, and p e rf
a perfect aspect marker):

(26) a. Ùn lέn ɖɔ wémà lɔ wε Sέna xìá.

I think-perf that book the f o c Sena read-p e rf
‘I think that it is the book that Sena has read’
(Gungbe, Aboh 2004: 238)
b. Gà mālàm nē na maid dà littāfin
to teacher foc I return prt book
‘It was to the teacher that I returned the book’
(Hausa, Green 2007: 62)
c. Ekitabu kyo Kambale a-asoma
book foc Kambale agr -read
‘It was the book that Kambale read’
(Kinande, Schneider-Zioga 2007: 412)

This makes it plausible to suggest that peripheral focused constituents are

the specifiers of a focus phrase (= FOCP) projection which is headed by an
overt (bold-printed) focus particle in languages like Gungbe, Hausa and
230 The clause periphery

Kinande, but by a null focus particle in languages like English. On this

view, the that-clause in a sentence like (24 b ) above will be derived as
shown below (where 451 goals is treated as a NUMP/numeral phrase, and
the internal structure of TP is not shown):

(27) FORCEP

FORCE FOCP
that
NUMP FOC′
451 goals
FOC TP
ø he scored — in 438 games

Since this is a gap structure (i.e. a structure in which there is a gap —

internally within TP), we can suppose that the italicised focused constitu-
ent originates within TP in the position marked by the gap and from there
moves to the edge of the focus phrase projection in the clause periphery, as
shown by the arrow in (27). It would seem that Focusing always involves
movement, whereas (as we saw earlier) topics can either be directly merged
in situ in spec‑TOPP or moved there.
The assumption that clauses can contain peripheral topic and focus pro-
jections raises the question of whether the periphery of a given clause can
contain both topicalised and focused constituents. The answer is ‘yes’, as
we see from authentic examples such as the following:

(28) a. Fernando Alonso, 189 points he’ll have — (David Croft, BBC Radio 5)
b. Kevin Pietersen, five he was out for — (George Reilly, BBC Radio 5)
c. The goal, none more important he has scored — than that one (Gary
Neville, Sky Sports TV)
d. Harry Kane, 100 he scores (Banner in Liverpool’s Anfield Stadium,
celebrating Harry Kane scoring his 100th goal)
e. Stuart Broad, a fantastic piece of bowling that was — in the last test
match (Elliot Cook, talkSPORT Radio)

The bold-printed constituent on the periphery of each of the clauses in

(28) is an in situ topic (directly merged in spec‑TOPP), reprised by the
underlined resumptive constituent within the main body of the clause in
(28a–28d), but unreprised in (28e). The italicised constituent is focused and
appears to have been fronted (i.e. positioned at the front of the clause by
movement), since it leaves a gap behind lower down in the clause (marked
by — ). Given these assumptions (and the further assumption that clauses
canonically contain a FORCEP projection marking clause type), the clause
periphery in (28a) will have the following structure:
 4.2 Focus projections 231

(29) FORCEP

FORCE TOPP
ø
DP TOP′
Fernando
Alonso TOP FOCP
ø
NUMP FOC′
189 points
FOC TP
ø he’ll have —

The null indicative FORCE head of FORCEP in (29) serves to type the rele-
vant clause as declarative in force. The TOPP projection below FORCEP
contains an (in situ) topic Fernando Alonso which is directly merged in
spec‑TOPP, and reprised by the resumptive pronoun he in spec‑TP. The
focused NUMP/numeral projection 189 points originates as the comple-
ment of the verb have and moves from its initial position to become the
specifier of the focus phrase/FOCP projection below TOPP, as shown by the
arrow in (29).
Although in sentences like (28) above, the bold-printed topic precedes
the italicised focused constituent, the converse order (in which an italicised
focused constituent precedes a bold-printed in situ topic) is also possible –
as we see from sentences such as the following:

(30) a. 189 points, Fernando Alonso, he now has —

b. Five, Kevin Pietersen, he was out for —, not fifty-five
c. Anything Messi can do, Ronaldo, he can do — even better

Sentences like (28, 30) are in keeping with the claim made by Rizzi (1997:
291) that ‘A focus and one or more topics can be combined in the same
structure. In that case, the focal constituent can be both preceded and fol-
lowed by topics.’

4.2.2 Comparing topic and focus

An important question arising from the postulation of separate topic and
focus projections in the clause periphery is what differences there are
between topicalised and focused constituents. One important difference
lies in their semantic properties: topics represent old/familiar informa-
tion and tell us what a sentence is about, and are often paraphrasable by
‘As for … ’. By contrast, focused constituents represent new/unfamiliar
information, and are often paraphrasable by a cleft or pseudo-cleft sen-
tence. Consequently, (28a) can be paraphrased as follows: ‘As for Fernando
232 The clause periphery

Alonso, what he’ll have is 189 points’. Topicalised and focused constitu-
ents also differ in that there is typically a slight pause immediately after
a topic (marked by the comma after the bold-printed topics in 28), but
not necessarily after a focused constituent (like those italicised in 25).
Furthermore, a topic can be positioned on the edge of a clause either by
Move or by Merge (and a topic merged in situ in the clause periphery may
or may not be reprised by a resumptive expression internally within the
clause), whereas a focused constituent can only get into the clause per-
iphery by movement (and so is always associated with a gap internally
within the main body of the clause, and cannot be reprised by a resump-
tive constituent).
A further difference between topicalised and focused constituents is that
topics (by virtue of referring to a specific entity assumed to be familiar to
the hearer) are referential expressions, and consequently non-referential/
non-specific pronouns like something/someone/somebody, anything/anyone/
anybody, or everything/everyone/everybody can only be focused, not topic-
alised – as we see from the ungrammaticality which results from using a
(bold-printed) referential resumptive pronoun (i.e. the kind of pronoun used
to reprise a topic) to reprise a fronted non-specific pronoun in sentences
such as:

(31) a. Everything, the tanks destroyed (*it)

b. Something, I think (*it) must have upset him
c. Anything, people are willing to confess to (*it) under torture

If the italicised constituents in (31) were topics, it ought to be possible for

them to be reprised by a (bold-printed) resumptive pronoun; the fact that
this is not possible suggests that non-specific expressions are not possible
topics, and hence that the italicised constituents in (31) are focused.
An additional difference between focused and topicalised constituents is
that although clauses can contain more than one topic, they cannot contain
more than one focused constituent – as we see from the impossibility of
focusing both something and everyone in (32b) below:

(32) a. You must give something to everyone at Christmas

b. *Everyone, something you must give — to — at Xmas

The ungrammaticality of multiple-focus structures like (32b) results from

violation of the Intervention Condition (21), in that (32b) involves illicit
movement of one fronted focused QP (everyone) across another fronted
focused QP (something), thereby inducing an intervention violation.
The numerous differences outlined above between peripheral topicalised
and focused constituents make it plausible to posit that the two serve as the
specifiers of different types of functional head (a peripheral topic serving
 4.3 Modifier and finiteness projections 233

as the specifier of a TOP head, and a peripheral focused constituent serving

as the specifier of a FOC head).
You should now be able to tackle Exercise 4.2.

4.3 Modifier and finiteness projections

In this module, we will see that in addition to FORCEP, TOPP and FOCP
constituents, the clause periphery can also contain one or more modifier
projections/MODP, and a finiteness projection/FINP.

4.3.1 Modifier projections

In addition to complementisers, topics and focused constituents, the per-
iphery can also contain one or more clausal adjuncts. These are constitu-
ents (like those italicised below) which serve to modify the particular clause
(bracketed below) in whose periphery they are situated:

(33) a. I suspect [that in spite of the rain the game will go ahead]
b. I wonder [if during the holidays you could find time to read my
dissertation]
c. I have heard [that tomorrow he will fly to Paris]
d. I know [that occasionally he plays poker]
e. I would think [that if it rains we should abandon the game]
f. I’d suggest [that when he comes home you should talk to him]

Such adjuncts typically provide information about the circumstances (e.g.

time, place, manner, conditions etc.) under which the event described in the
modified clause takes place, and so they are termed circumstantial. Typical
circumstantial peripheral constituents include prepositional phrases (as in
33a, 33b) above, or adverbials (as in 33c, 33d), or clauses introduced by
subordinating conjunctions (as in 33e, 33f). Circumstantial modifiers like
these don’t give rise to intervention effects, as we see from the fact that in
sentences like (34) below, a (bold-printed) topicalised constituent can move
out of the gap position in which it originates into the clause periphery
across an intervening (italicised) modifier:

(34) a. I must warn you that this kind of behaviour, in this kind of town,
people will not tolerate —
b. I must say that the government’s response, in the present crisis, I
find — wholly unacceptable
c. Cheese, often people have strong feelings about — (Davison 1984:
807)
d. Wine, in some restaurants, you have to pay a small fortune for —
234 The clause periphery

This suggests that the italicised modifiers are directly merged in situ in the
position they occupy in the clause periphery. Since they serve to modify
the clause containing them, this raises the possibility that they serve as
specifiers of a modifier projection/MODP in the clause periphery. If so, the
that-clause in a sentence like (34a) above will have the structure shown in
simplified form below, with the fronted topic this kind of behaviour cross-
ing the intervening in situ PP modifier in this kind of town without indu-
cing any intervention effect:

(35) FORCEP

FORCE TOPP
that
DP TOP′
this kind of
behaviour TOP MODP
ø
PP MOD′
in this kind
of town MOD TP
ø people
will not
tolerate —

MODP projections can occupy a wide range of different positions in the

clause periphery, as illustrated by the authentic examples below (which
I recorded from a variety of live, unscripted radio and TV programmes),
where the italicised constituents are modifiers, the bold-printed constitu-
ents are dislocated topics, and the underlined constituents are resumptive
pronouns:

(36) a. That tells you that at the highest level, the big teams, they don’t
fancy it (Darren Lewis, talkSPORT Radio)
b. At the end of the day, the players, they’re in there to win things
(Listener, talkSPORT Radio)
c. Fabio Capello, whether we like it or not, he’s still part of the England
team (Adrian Durham, talkSPORT Radio)
d. The director tonight, if we had extra time, he would hang, draw and
quarter me (Dotun Adebayo, BBC Radio 5)
e. You have to say that Higuain, normally he would have hit the target
(Jon Driscoll, Sky Sports TV)

Moreover, as sentences like those in (37) below illustrate, the periphery of a

single clause can contain multiple (italicised) modifiers (following a bold-
printed dislocated topic in the case of 37b):
 4.3 Modifier and finiteness projections 235

(37) a. Last week, in Paris, after a hard day’s work, he met his agent again
(Haegeman 2012: 73)
b. I just felt that Roy Hodgson, a few weeks ago, when Liverpool lost to
Everton, he was in a minority of one (John Motson, BBC Radio 5)

This suggests that modifiers can (in general) be freely positioned between
other constituents within the periphery. If each MODP projection contains only
one clausal modifier, this in turn means that the periphery of a given clause
can optionally contain one or more MODP projections (giving rise to MODP
recursion – i.e. clauses whose periphery contains more than one MODP).
Our discussion so far leads to the more general conclusion that each XP
in the clause periphery is housed in (and serves as the specifier of) a dedi-
cated functional projection (i.e. a functional projection which only allows
a specific type of specifier). Thus, a topic serves as the specifier of a topic
projection/TOPP, a focused constituent as the specifier of a focus projec-
tion/FOCP, and a clausal modifier as the specifier of a modifier projection/
MODP. Note in particular that this approach to clausal modifiers treats
them as specifiers contained in a separate MODP projection of their own,
rather than (as in earlier work discussed in Chapter 1) as adjuncts which
adjoin to a TP or CP to form an even larger TP or CP.

4.3.2 Finiteness projections

A further type of peripheral projection identified by Rizzi (1997) is a finite-
ness projection/FINP that is positioned below all other peripheral projections
and immediately above TP, and serves to mark whether the clause contain-
ing it is finite or nonfinite. One type of (nonfinite) FIN head found in English
is the infinitival complementiser for, for example, in sentences such as:

(38) She never intended for him to get hurt

As would be expected if for is the head FIN constituent of a FINP projection

which is positioned below all other peripheral projections and immediately
above TP, for is positioned below other (italicised) constituents in the per-
iphery of the same clause and immediately above the (underlined) clause
subject in spec‑TP:

(39) a. What’s critical is [if people saw the helicopter, for them to contact us]
(Police spokeswoman, BBC Radio 5)
b. We would not want [in a game like the one tomorrow, for any
spectators to get hurt]
c. The police advice was [on no account for any member of the public
to approach the escaped convict]

In these examples, infinitival for is preceded by an (italicised) clausal modi-

fier contained within a MODP projection – the modifier being a conditional
236 The clause periphery

if-clause in (39a), and a PP in (39b, 39c). Assuming that for is in FIN, such
examples are consistent with FINP being the lowest projection in the clause
periphery, hence positioned immediately above the TP housing the (under-
lined) subject of the clause. If case-marking by a transitive head like for
requires the transitive head to c‑command and be immediately adjacent to
the constituent it case-marks, it follows that the FIN head for in (39) will
have to immediately precede the underlined infinitive subject (and not be
able to precede the italicised modifier). This will be the case if for is in FIN,
and FINP is the lowest projection in the periphery, so that FINP is posi-
tioned immediately above the subject in spec‑TP.
Given the assumption that the head FIN constituent of FINP marks
whether a clause is finite or nonfinite, we might expect English to have
a finiteness marker which is the finite counterpart of the infinitival FIN
constituent for. In this connection, it is interesting to observe double-that
clauses such as those bracketed below:

(40) a. And now we’re told [that, for a further two years, that we’re going to
have that pay restraint] (Dave Prentice, BBC Radio 4)
b. He knows [that, at the end of the summer, whether we do anything
at the Euros or not, that he’s going to be going] (Adrian Durham,
talkSPORT Radio)
c. The party opposite said [that if we cut 6 billion from the budget
that it would end in catastrophe] (David Cameron, Prime Minister’s
Questions, BBC Radio 5)
d. My hope is [that, by the time we meet, that we’ll have made some
progress] (Barack Obama, Press conference, BBC Radio 5)

The bracketed clauses in these examples contain an initial occurrence of

that, then one or two (italicised) modifying adjunct phrases/clauses, then
a second occurrence of that, and then the (underlined) clause subject. The
first occurrence of that in a sentence like (40d) can plausibly be taken to
occupy the head FORCE position of FORCEP (marking the clause it intro-
duces as indicative in mood and declarative in force), but where is the sec-
ond occurrence positioned? It might at first sight seem as if the second that
could occupy the head MOD position of the MODP containing the clausal
modifier by the time we meet. However, any such structure would fall foul
of a filter posited by Chomsky & Lasnik (1977), which was given the fol-
lowing informal characterisation in §1.5.2:

(41) Doubly Filled COMP Filter

At the end of a syntactic derivation, any structure in which the edge
of a projection headed by an overt complementiser is doubly filled (i.e.
contains some other overt constituent) is filtered out as ungrammatical
 4.3 Modifier and finiteness projections 237

One way of avoiding any violation of the filter in (41) is to take the second
occurrence of that to occupy the head FIN position of FINP. If so, the per-
iphery of the clause bracketed in (40d) will have the structure shown below:

(42)
FORCEP

FORCE MODP
that
PP MOD′
by the time
we meet MOD FINP
ø
FIN TP
that we’ll have
made some
progress
This would mean that, in spoken English, the particle that in an indicative
clause can either be used to spell out a FORCE head (as with the bold-printed
occurrence of that in 42), or can be used to spell out a FIN head (as with the
italicised occurrence of that in 42). We can suppose that in the same way as
a transitive infinitival FIN head like for assigns accusative case to its subject,
so too the finite FIN head that in (42) serves to assign nominative case to its
subject we – whereas other peripheral heads (including the FORCE head that)
are not case assigners. The assumption that FIN is a case assigner would mean
that the Case Conditions presented in §3.1.2 need to be revised along the lines
specified below, by replacing mention of C as a case assigner by FIN in (43i):

(43) Case Conditions

A case-assigning probe values an unvalued case feature on an XP goal
and assigns the goal a case value that depends on the nature of the
probe – namely:
(i) nominative if the probe is a finite FIN with a TP complement (where
a finite FIN is one that is indicative, subjunctive or imperative in
mood)
(ii) null if the probe is an intransitive infinitival FIN with a TP
complement
(iii) accusative if the probe is transitive

The claim embodied in (43) that FIN (not FORCE) is the case assigner in
the clause periphery has interesting consequences for a clause such as that
bracketed in (44a) below, which we can take to have the peripheral struc-
ture shown in simplified form in (44b):

(44) a. I must admit [that me too, I hate garlic]

238 The clause periphery

b. FORCEP

FORCE TOPP
that
DP MOD′
me too
MOD FINP
ø
FIN TP
ø I hate garlic

On the assumptions made in (43), the null FIN head ø in (44b) assigns nom-
inative case to the subject I (because FIN is finite, FIN c‑commands I, and
FIN is the closest case-assigner c‑commanding I), but the FORCE head that
is not a case assigner – hence it does not assign nominative case to the
immediately subjacent topic me too. Instead, the topic receives accusative
case by default by virtue of not falling within the domain of a local case
assigner: see the Default Case Assignment Condition in (15) above, and the
discussion of (14, 16).
You should now be able to tackle Exercise 4.3.

4.4 Complete and truncated clauses

In this module, we’ll see that some clauses have a more complex peripheral
structure than others. More specifically, we’ll see that some are complete
clauses which include both FINP and FORCEP projections (and can option-
ally include other projections like TOPP, FOCP and MODP positioned
between FORCEP and FINP), whereas others are truncated clauses which
project only as far as FINP, and thus lack the FORCEP projection found in
complete clauses (as well as lacking TOPP and FOCP projections).

4.4.1 Complete clauses

An interesting theoretical issue arising from the split projection view of
the clause periphery concerns how long/cross-clausal movement works in
the case of complete clauses (i.e. movement out of one complete clause
into another). In this connection, consider how Topicalisation works in the
structure below:

(45) speaker a : The demonstrators have been looting shops and setting fire
to cars
speaker b : This kind of behaviour, they must know [that, with the full

force of the law, we will crack down on — ]

 4.4 Complete and truncated clauses 239

Let’s suppose that all root clauses are complete clauses (since they require
a FORCEP projection to mark their illocutionary force), and that clauses
introduced by the indicative complementiser that are likewise complete
clauses (because indicative that clauses are declarative in force). Given
these assumptions, (45 b ) will involve movement out of one complete clause
into another. More specifically, the italicised topic this kind of behaviour
originates in the gap position (as the complement of the preposition on)
and from there moves across the edge of the that-clause to the specifier
position in a topic projection at the front of the main clause. But how does
this movement take place?
Locality constraints require long/cross-clausal movement to take place
one clause at a time, so that a constituent moving out of one clause into an-
other transits through the edge of the highest projection in the lower clause
before moving into the higher clause. One of these locality constraints is
Chomsky’s (1998) Impenetrability Condition (discussed in §1.5.1), which
we can adapt from the CP framework to the split projection framework by
reformulating it as:

(46) Impenetrability Condition

Anything below the head H of the highest projection HP in a complete
clause is impenetrable to anything above HP

Above/below here mean ‘c-commanding/c-commanded by’, and a complete

clause is one containing a FORCEP projection. Let’s therefore suppose that
the topic transits through the highest peripheral projection in the comple-
ment clause, and thus moves into spec‑FORCEP in the that-clause, before
subsequently going on to move into its criterial position in spec‑TOPP
in the main clause. On this view, the topic first undergoes the movement
arrowed below in the complement clause:

(47) FORCEP

DP FORCE′
this kind of
behaviour FORCE MODP
that
PP MOD′
with the
full force MOD FINP
of the law ø
FIN TP
ø we will crack
down on—
240 The clause periphery

The arrowed movement will not violate the Impenetrability Condition/IC

outlined in (46) above, since IC bars direct movement from a position below
a FORCE head to one above/outside FORCEP (and the fronted topic in 47
moves from a position below the FORCE head that into a specifier position
inside the FORCEP headed by that).
Once it has moved to the edge of the FORCEP projection in the comple-
ment clause, the fronted topic can then move from there into the specifier
position of a TOPP projection in the main clause, as shown by the arrow
below:

(48) FORCEP2

FORCE TOPP
ø
DP TOP′
this kind of
behaviour TOP FINP
ø
FIN TP
ø
DP T′
they
T VP
must
DP V′
they
V FORCEP1
know
DP FORCE′
—
FORCE MODP
that with the
full force
of the law,
we will
crack
down on
Since the topic moves from a position on the edge of highest projection
(FORCEP1) in the periphery of the lower clause to a position within the per-
iphery of the higher clause, the arrowed movement in (48) does not violate
the Impenetrability Condition (46). Since spec‑FORCEP is not the criterial
position for a topic (but rather is simply a transit position through which
any XP constituent moving out of the complement clause can transit),
the movement arrowed in (48) will not incur any violation of the Criterial
Freezing Condition (3) either.
 4.4 Complete and truncated clauses 241

4.4.2 Truncated clauses

In addition to complete clauses which have the status of FORCEP projec-
tions able to house a full range of peripheral constituents (e.g. topics and/
or focused constituents), there is also a class of truncated clauses whose
periphery projects only as far as the FINP projection. Rizzi & Shlonsky
(2007) and Rizzi (2014a) posit that bare declarative complement clauses
(i.e. those not introduced by the overt complementiser that) are truncated at
the FINP level (in the sense that they project only as far as FINP): accord-
ingly, Rizzi (2014a: 30) claims that ‘finite clauses with null complementis-
ers in English involve truncation of the higher structure of the C system,
of FORCE (hence the absence of the declarative force-marker that) and of
the topic-focus system’ and thus contain only ‘a vestigial presence of the
C‑system’ comprising a projection of ‘the lowest head, FIN’.
The difference between complete and truncated (i.e. incomplete/defective)
clauses can be illustrated by the contrast below:

(49) a. John believes [that Mary wrote this book]

b. John believes [that this book, Mary wrote]

(50) a. John believes [Mary wrote this book]

b. *John believes [this book, Mary wrote]

Given Rizzi’s assumptions, the bracketed that-clauses in (49) will be com-

plete clauses projecting up to FORCEP, and consequently the FORCE head
that allows a TOPP complement containing the fronted topic this book in
(49b). By contrast, the bare complement clauses bracketed in (50) are trun-
cated clauses that project only as far as FINP, and since a FIN head does
not allow a TOPP complement, topicalisation of the italicised DP this book
is not possible in (50b).
In this connection, it is interesting to note that for-infinitives don’t allow
a peripheral topic like that italicised in (51a) below or a peripheral focused
constituent like that italicised in (51b), or a peripheral FORCE head like
interrogative whether in (51c):

(51) a. *He is anxious [the inheritance dispute for us to settle it amicably]

b. *He is anxious [everything for them to thrash out in detail]
c. *He is wondering [whether for us to have a staycation this summer]

One way of ruling out structures like (51) would be to suppose that for-
infinitives are truncated clauses which project only as far as FINP. Such an
analysis could also be extended to Control infinitives with a PRO subject
like that bracketed in (52a) below, since such clauses do not allow a periph-
eral focus constituent (like that italicised in 52b), or a peripheral dislocated
topic (like that italicised in (52c):
242 The clause periphery

(52) a. He decided [PRO to leave everything to his children]

b. *He decided [everything PRO to leave to his children]
c. *He decided [his children PRO to leave everything to them]

By contrast, interrogative Control infinitives like those bracketed below

have the status of complete clauses, since they require a FORCEP projec-
tion to house the italicised interrogative operator and type the clause as a
question:

(53) a. I didn’t know [whether PRO to laugh or cry]

b. I didn’t know [what PRO to say]
c. I didn’t know [who PRO to contact]

In sentences like (52, 53), we can suppose that PRO is assigned null case
by a null, intransitive infinitival FIN head; by contrast, in for-infinitives
the infinitive subject is assigned accusative case by the transitive FIN
head for.
Thus, some verbs/predicates select a FORCEP complement (and allow
other peripheral constituents like a topic or focus), whereas others select
a (smaller) FINP complement (and do not allow other peripheral constitu-
ents). Still, this picture is seemingly challenged by clauses like those brack-
eted below:

(54) a. And I’m hoping [on Friday night that we can turn up and get the
points] (Dean Saunders, BBC Radio 5)
b. The problem we’ve got is [in an ideal world, that all patients would
be on single-sex wards] (Hospital spokesperson, BBC Radio 5)
c. I’m sure [behind the scenes that he’s got the backing] (Steve
McClaren, BBC1 TV)
d. Arsène Wenger said [if a defensive midfield player becomes available,
that they will be in for him] (Mark Chapman, BBC Radio 5)
e. I’m not sure [going into the Premier League, that their squad’s that
strong] (Jason Burt, talkSPORT Radio)

Here, the bracketed complement clauses are bare (in the sense that they do
not begin with the overt complementiser that) and thus, on the assump-
tion that a declarative FORCE head is always lexicalised as that) must have
the status of FINP rather than FORCEP. But if so, how come they allow an
italicised PP or subordinate clause in their periphery? A plausible answer
comes from the suggestion made by Rizzi (2014a) that natural language
grammars allow FINP recursion structures of the form FINP+MODP+FINP
(where MODP is a projection which houses circumstantial clause modifiers).
On this view, the V‑bar headed by the verb hoping in (54a) above would
have the FINP recursion structure below:
 4.5 Summary 243

(55) V′

V FINP2
hoping
FIN2 MODP
ø
PP MOD′
on Friday
night MOD FINP1
ø
FIN1 TP
that we can
turn up and
get the points

If an overt complementiser like that serves to mark the upper or lower

boundary of the clause periphery (and hence can either spell out the high-
est FORCE head or the lowest FIN head in the periphery), then it follows
that only the lower FIN head (FIN1) will be lexicalised as that. (It should be
noted, however, that some speakers do not like sentences in which FIN is
spelled out as that.)
The key assumptions made about clause structure in this module can be
summarised as follows:

(56) Assumptions about the clause periphery

(i) Clauses which can contain peripheral topic or focus constituents
are complete clauses which have FORCEP as the highest and FINP
as the lowest projection in the periphery (with any other peripheral
projections being positioned between FORCEP and FINP)
(ii) Clauses which cannot contain peripheral force-marking, topicalised
or focused constituents are truncated clauses which have a
periphery that projects only as far as FINP
(iii) FINP recursion allows for FINP+MODP+FINP structures

You should now be able to tackle Exercise 4.4.

4.5 Summary

In this chapter, we have seen that the clause periphery (traditionally ana-
lysed as containing one or more CP projections) has been split up in recent
Cartographic work into a number of distinct projections (including FORCEP,
TOPP, FOCP, MODP and FINP). In Module 4.1, I argued that all (non-
defective) clauses contain a FORCEP projection which serves to specify the
illocutionary force of the clause (e.g. whether it is declarative, interrogative
244 The clause periphery

or imperative in force); I noted that a declarative FORCE head in a com-

plement clause can be spelled out as the complementiser that. I went on to
argue that the criterial position for a peripheral topic constituent is as the
specifier of a TOPP projection below FORCEP, and that clauses can contain
one or more TOPP projections. I distinguished between two different types
of topic: in situ topics which are directly merged in spec‑TOP and may be
reprised by a resumptive constituent lower down in the structure; and ex situ
topics which originate in some position within TP and from there move into
spec‑TOPP, leaving a gap behind. I noted that the two types of topic differ
in that ex situ topics are case-marked in the position which they originate
in (and obey constraints on movement), whereas in situ topics are assigned
accusative case by default, and are not subject to movement constraints.
In Module 4.2, I argued that peripheral focused constituents move to their
criterial position in spec‑FOCP from a position within TP, leaving a gap be-
hind in the position which they move from. I noted that a FOCP constituent
can either precede or follow any TOPP constituent in the same clause, but
that the two types of constituent differ in three ways: (i) focused constitu-
ents always result from movement into their criterial position and always
leave a gap behind, whereas topicalised constitituents can either be gener-
ated in situ in their criterial position, or be moved there; (ii) the periphery
of a given clause can only contain a single FOCP, but can contain multi-
ple TOPP constituents; (iii) non-specific pronouns like something/someone/
somebody, anything/anyone/anybody, or everything/everyone/everybody can
only be focused, not topicalised.
In Module 4.3, I noted that in addition to a force-marking complementis-
er, a focus, and one or more topics, the clause periphery can also con-
tain one or more circumstantial modifiers: these are typically PP, ADVP,
or subordinate clause constituents which provide information about the
circumstances (e.g. time, place, manner, conditions etc.) under which the
event described in the modified clause takes place. I argued that these cir-
cumstantial modifiers are directly merged in situ as the specifiers of MODP
projections, and that the MODP containing them can occupy a range of
periphery-medial positions between the highest and lowest peripheral pro-
jections in a given clause. I went on to argue that the periphery of a clause
also contains a FINP projection (in the lowest position in the periphery,
immediately above TP) whose head serves to mark its containing clause as
finite or nonfinite, and to case-mark the subject (e.g. the infinitival comple-
mentiser for is a FIN head which serves to mark its clause as infinitival, and
serves to case-mark the subject of the for-clause as accusative).
In Module 4.4, I outlined the claim made by Rizzi that although complete
clauses have the status of FORCEP projections which can house a full range
of peripheral constituents (e.g. topics and/or focused constituents), there
 4.5 Summary 245

is also a class of truncated clauses whose periphery projects only as far as

the FINP projection (and so do not allow peripheral topic or focus constitu-
ents). I took a brief look at how cross-clausal movement works when an XP
moves out of one complete clause into another, and argued that (in order
to satisfy locality constraints on movement such as the Impenetrability
Condition), a constituent which moves out of one clause into another has
to transit through the edge of the highest peripheral projection in the lower
clause (FORCEP) before moving into the periphery of the higher clause. I
then went on to look at the structure of truncated clauses, arguing that
these include declarative complement clauses not introduced by the overt
complementiser that as their first word, and for-infinitives. I noted that
Rizzi allows for the possibility that the periphery of truncated clauses may
contain a MODP constituent, and handles this in terms of FINP recursion
structures of the form FINP+MODP+FINP.
Thus, the overall picture which our discussion in this module leads us
to is that the periphery of a complete clause comprises FORCEP and FINP
projections between which there can optionally be TOPP, FOCP and MODP
projections; but by contrast, the periphery of truncated clauses projects
only as far as FINP, though the mechanism of FINP recursion allows for
FINP+MODP+FINP structures.
Constructs and assumptions which have played a key role in our discus-
sion in this chapter include the following:

(2) One Feature One Head Principle

Each functional head carries only one interpretable feature

(3) Criterial Freezing Condition/CFC

(i) A peripheral constituent must occupy a criterial position to be
interpretable at LF
(ii) A constituent which occupies its criterial position is frozen in place

(11) Island Constraint

No gap inside an island can have an antecedent outside the island

(15) Default Case Assignment

A (pro)nominal which does not fall within the domain of (i.e. which
is not c‑commanded by) a local (i.e. nearby) case assigner receives
accusative case by default

(21) Intervention Condition

Likes cannot cross likes

(22) Distinctness Condition

Any structure at PF which contains a sequence of two non-distinct
constituents <α, α> is degraded
246 The clause periphery

(41) Doubly Filled COMP Filter

At the end of a syntactic derivation, any structure in which the edge
of a projection headed by an overt complementiser is doubly filled (i.e.
contains some other overt constituent) is filtered out as ungrammatical

(43) Case Conditions

A case-assigning probe values an unvalued case feature on an XP goal
and assigns the goal a case value that depends on the nature of the
probe – namely:
(i) nominative if the probe is a finite FIN with a TP complement (where
a finite FIN is one that is indicative, subjunctive or imperative in
mood)
(ii) null if the probe is an intransitive infinitival FIN with a TP
complement
(iii) accusative if the probe is transitive

(46) Impenetrability Condition

Anything below the head H of the highest projection HP in a complete
clause is impenetrable to anything above HP

(56) Assumptions about the clause periphery

(i) Clauses which can contain peripheral topic or focus constituents
are complete clauses which have FORCEP as the highest and FINP
as the lowest projection in the periphery (with any other peripheral
projections being positioned between FORCEP and FINP)
(ii) Clauses which cannot contain peripheral force-marking, topicalised
or focused constituents are truncated clauses which have a
periphery that projects only as far as FINP
(iii) FINP recursion allows for FINP+MODP+FINP structures

4.6 Bibliographical notes

The Cartographic approach to the periphery has been developed in work by

Luigi Rizzi over several decades (see Rizzi 1996, 1997, 2000a, 2000b, 2001,
2004a, 2004b, 2005, 2006a, 2006b, 2006c, 2007, 2010, 2011, 2012, 2013a,
2013b, 2013c, 2014a, 2014b, 2015a, 2015b, 2016a, 2016b, 2017, 2018,
2020; Rizzi & Shlonsky 2006, 2007; Cinque & Rizzi 2008, 2010a, 2010b;
Rizzi & Cinque 2016; Rizzi & Bocci 2017), and built on by numerous other
researchers, including Aboh (2004, 2005, 2006, 2007, 2010, 2016), Aboh
& Pfau (2010), Badan (2007), Badan & Del Gobbo (2010), Baltin (2010),
Belletti (2004a, 2004b, 2009), Benincà (2001, 2006, 2010, 2012a, 2012b),
Benincà & Cinque (2010), Benincà & Munaro (2010), Benincà & Poletto
 4.6 Bibliographical notes 247

(2004), Bianchi & Cruschina (2016), Bianchi & Frascarelli (2010), Bianchi
et al. (2016), Biloa (2013), Bocci (2004, 2007, 2009, 2013), Brugé et al.
(2012), Cardinaletti (2004, 2009), Cardinaletti et al. (2014), Cinque (2002,
2023), Cruschina (2006, 2008, 2010a, 2010b, 2011a, 2011b, 2012), Cruschina
& Remberger (2008, 2017), Danckaert (2011, 2012), Demonte & Fernández
Soriano (2009, 2013, 2014), Durrleman (2008), Endo (2007a, 2007b, 2014,
2015a, 2015b, 2017, 2018), Franco (2009), Frascarelli (2000), Frascarelli &
Hinterhölzl (2007), Frascarelli & Puglielli (2007), Garzonio (2005), Giusti
(1996, 2012), Grewendorf (2002), Grewendorf & Poletto (2009), Haegeman
(2000a, 2003, 2004, 2006a, 2006b, 2007a, 2007b, 2009, 2010, 2012, 2013,
2017, 2019, 2021), Haegeman & Hill (2014), Jayaseelan (2003, 2008),
Jiménez-Fernández (2011, 2015), Krapova (2002), Krapova & Cinque (2008),
Laenzlinger (1999), Legate (2002), Munaro (2003), Nye (2013), Paoli (2003,
2007), Paul (2005, 2014), Pearce (1999), Poletto (2000), Puskás (2000),
Roberts (2004), Roussou (2000), Salvi (2005), Shlonsky (1997, 2010, 2014),
Shlonsky & Soare (2011), Shlonsky & Bocci (2019), Speas & Tenny (2003),
Torrence (2013), Tsai (2008), Villa-García (2010, 2011a, 2011b, 2012a,
2012b, 2012c, 2015), and Villalba (2000). For overviews of the Cartographic
approach, see Cinque & Rizzi (2008, 2010a), Shlonsky (2010), Rizzi (2013a),
Rizzi & Cinque (2016), Rizzi & Bocci (2017), and Shlonsky & Bocci (2019).
It should be noted that much Cartographic research posits a far wider range
of types of peripheral projection than that assumed here; however, to sim-
plify exposition I have consciously limited the number of different types of
peripheral head utilised here.
On topics, see Gruber (1967), Ross (1967, 1973), Lakoff (1969), Neubau-
er (1970), Emonds (1970, 1976, 2004), Postal (1971, 1972), Higgins (1973),
Hirschbühler (1973, 1974, 1975), Rodman (1974, 1997), van Riemsdijk &
Zwarts (1974, 1997), Gundel (1975, 1985, 1988), Contreras (1976), Keenan
& Schieffelin (1976a, 1976b), Chomsky (1977), Cinque (1977, 1983), Rive-
ro (1978, 1980), Prince (1981a, 1981b, 1984, 1997), Reinhart (1981), Vat
(1981), Kitagawa (1982), van Haaften et al. (1983), Greenberg (1984), Peset-
sky (1989), Demirdache (1991), Ward & Prince (1991), Authier (1992), Ge-
luykens (1992), Lasnik & Saito (1992), Watanabe (1993), Lambrecht (1994),
Ziv (1994), Büring (1997, 1999, 2003), Grohmann (1997, 2000, 2003, 2006),
Zaenen (1997), Birner & Ward (1998a, 1998b), Portner & Yabushita (1998),
Maki et al. (1999), Cormack & Smith (2000), Frascarelli (2000), Gregory &
Michaelis (2001), Platzack (2001, 2004), Grewendorf (2002), Boeckx (2003,
2012), Belletti (2004a), Benincà & Poletto (2004), Casielles-Suárez (2004), Frey
(2004, 2005), Shaer & Frey (2004), Boeckx & Grohmann (2005), Grohmann
(2006), Samek-Lodovici (2006, 2009), Féry (2007), Frascarelli & Hinterhölzl
(2007), Manetta (2007), Bianchi & Frascarelli (2010), Ott (2012a, 2012b, 2014,
2015), Constant (2014), Villa-García (2015), Bayer & Dasgupta (2016), Büring
248 The clause periphery

(2016), Miyagawa (2017a), and Radford (2018). Some researchers argue for
there being more than one type of topic projection: for discussion, see Belletti
(2004a), Benincà & Poletto (2004), Bianchi & Frascarelli (2010), Frascarelli &
Hinterhölzl (2007), and Samek-Lodovici (2006, 2009).
On Left Dislocation, see van Riemsdisk & Zwarts (1974, 1977), Rodman
(1974, 1977), Gundel (1975), Cinque (1977), Vat (1981), van Haaften et al.
(1983), Greenberg (1984), Prince (1984), Anagnastopoulou et al. (1997),
Grohmann (1997), Gregory & Michaelis (2001), Grewendorf (2002), Frey
(2004, 2005), Boeckx & Grohmann (2005), and Manetta (2007).
On focus, see Schachter (1973), Akmajian (1979), Prince (1981a), Culi-
cover & Rochemont (1983), Taglicht (1984), Erteschik-Shir (1986), Brody
(1990), Kiss (1995, 1998), Rooth (1985, 1992, 1996), Horvath (1986, 1995,
2000, 2007), Rochemont (1986), König (1991, 2017), Rizzi (1996, 1997,
2013c), Büring (1997, 2009), Kiss (1998), Lahiri (1998), Romero (1998),
Zubizaretta (1998, 2009), Ambar (1999), Kidwai (1999), Pearce (1999), Cor-
mack & Smith (2000), Frascarelli (2000), Benincà (2001), Szendröi (2001),
Kim (2002, 2006), Drubig (2003), Green & Jaggar (2003), Benincà & Poletto
(2004), Brunetti (2004), Futagi (2004), Lee (2004), Park (2005), Cruschina
(2006, 2012), Féry & Samek-Lodovici (2006), Samek-Lodovici (2006, 2009),
Aboh (2007), Eckardt (2007), Frascarelli & Puglielli (2007), Green (2007),
Jayaseelan (1996, 2003, 2008), Hartman & Ai (2009), Badan & Del Gobbo
(2010), Fanselow & Lenertová (2010), Miyagawa (2010), Dehé et al. (2011),
Cheng & Downing (2012), M. Wagner (2012), Ortega-Santos (2013), Rizzi
(2013c, 2016a), Dominguez (2014), Irani (2014), Wierzba (2014, 2019), De
Cesare (2015), Jiménez-Fernández (2015), Arregi (2016), Authier & Haege-
man (2016), Bianchi & Cruschina (2016), Bianchi et al. (2016), van der Wal
(2016), Cruschina & Remberger (2017), Erlewine (2017b), Badan & Crocco
(2019), and Borise (2019).
On finiteness, see Lasser (1997), Klein (1998, 2006), Bianchi (2003),
Landau (2004), Adger (2007), Anderson (2007), Nikolaeva (2007), Rad-
ford (2008), Amritavalli (2013), Kissock (2013), Ritter & Wiltschko (2014),
O’Neill (2015), Cowper (2016), Eide (2016), Goodall (2017), Manzini & Sa-
voia (2018), Lowe (2019), and Manzini & Roussou (2019).
On intervention effects, see Starke (2001), Kim (2002, 2006), Boeckx &
Jeong (2004), Rizzi (2006b), Beck (2006), Beck & Kim (2006), Boeckx &
Lasnik (2006), Endo (2007a), Friedmann et al. (2009), Abels (2012), Haege-
man (2012), and Haegeman et al. (2014).
On the Doubly Filled COMP Filter, see Chomsky & Lasnik (1977),
Seppänen & Trotta (2000), Koopman (2000), Koopman & Szabolsci (2000),
Zwicky (2002), Collins (2007), Bayer & Brandner (2008), Baltin (2010),
Schönenberger (2010), Bayer (2015, 2016), Collins & Radford (2015), and
Radford (2013, 2018).
 4.7 Workbook 249

4.7 Workbook

In the exercises below, an [M] after an example sentence indicates that a

model answer is provided in the free-to-download Students’ Answerbook
(and in the Teachers’ Answerbook as well). You will find it helpful to look at
this model answer before attempting to tackle other examples in the exer-
cise. Some other exercise examples have an [S] after them, indicating that
they are for self-study: this means that after reading the relevant module
(and those preceding it), and looking at the helpful hints and model answer
associated with the exercise, students should be able to analyse these self-
study examples on their own, and then compare their answers with those
provided in the Students’ Answerbook. The remaining exercise examples
(i.e. those not marked with [M] or [S] after them) are intended for teachers
to use as the basis for hands-on problem-solving work in seminars, classes,
or assignments: answers to all of these non-self-study examples are pro-
vided in the Teachers’ Answerbook. Note that Exercise 4.1 tests you on
Module 4.1, Exercise 4.2 on Module 4.2, and so on; and that each exercise
in a given chapter presupposes familiarity with material covered in earlier
chapters and in earlier modules of the same chapter. Unparenthesised num-
bers like 4 refer to material in the exercises, whereas parenthesised numbers
like (22) refer to examples and conditions in the main text of the book.

EXERCISE 4.1

This exercise is designed to be attempted after reading Module 4.1 and the
chapters preceding it.
Discuss the structure of the periphery of the clauses in the sentences
below (in the case of sentences 1–3, concerning yourself only with the
bracketed clauses):

1 I must admit [that us republicans, we don’t trust you democrats] [M]

2 I think [that Red Bull, one of their spotters actually got arrested] (David
Coulthard, BBC1 TV)
3 I don’t know [whether this high tempo, we can sustain (it) for a full
season] [S]
4 This particular case, the FBI, they solved (it) very quickly [S]
5 Diesel cars, the pollution they cause, people underestimate (it)
6 The flat tyre, John explained that there had been nails on the ground (after
Rodman 1997: 38)
7 Hillary Clinton, her husband Bill Clinton, his advisor was a Ku Klux Klan
member (Political commentator, BBC Radio 5)
250 The clause periphery

8 The assignments, most of *(them) are fine [S]

9 This statue, the top of *(it), someone has vandalised

In 8 and 9, the notation *(them)/*(it) means that the sentence is grammatical

if them/it is used, but ungrammatical if it is not.

Helpful hints
In each of the sentences, concern yourself with the structure of the
clause periphery; do not concern yourself with the internal structure
of TPs below the periphery. In the case of sentences 1–3, contrast CP
recursion and Cartographic analyses; but for the remaining sentences,
present only Cartographic analyses. Take nominal constituents in
the periphery to be topics, and bear in mind that a topic can be (i) a
fronted/ex situ topic which leaves behind a gap when it moves, or (ii)
an in situ topic reprised by a resumptive pronoun like it/them, or (iii)
an in situ topic which is unreprised. Take the phrases us republicans,
you democrats, Red Bull, this high tempo, this particular case, the FBI,
the pollution they cause, the flat tyre, Hillary Clinton, her husband Bill
Clinton, his adviser, the assignments, this statue, and the top of it to be
DPs headed by an overt or null D, but don’t concern yourself with their
internal structure. In addition, take one of their spotters to be a NUMP,
and diesel cars to be a QP headed by a null quantifier (meaning ‘an
unspecified number of’), but don’t concern yourself with the internal
structure of these phrases. In 3, take whether to be an interrogative
complementiser which has a null yes–no operator OpYNQ as its specifier,
and take this operator to be an ADVP.
In arriving at your analysis, consider the possible role played by
conditions like the following in relevant examples:

10 Constraint on Extraction Domains/CED (from §1.5.1)

Only complements allow material to be extracted out of them, not
specifiers or adjuncts

11 Freezing Principle (from §1.5.1)

The constituents of a moved phrase are frozen internally within (and so
cannot be extracted out of) the moved phrase

12 Impenetrability Condition (= 46)

Anything below the head H of the highest projection HP in a complete
clause is impenetrable to anything above HP

13 Intervention Condition (= 21)

Likes cannot cross likes
 4.7 Workbook 251

14 Clause Typing Conditions (from §1.4.3):

(i) A clause is interpreted as being of a given type (interrogative or
exclamative etc.) if it contains a peripheral clause-typing specifier of
the relevant type
(ii) An indicative clause is interpreted as declarative in type by default if
it does not contain a peripheral clause-typing specifier

EXERCISE 4.2

This exercise is designed to be attempted after reading Module 4.2 and the
module and chapters preceding it.

Discuss the structure of the periphery of the clauses in the sentences

below (concentrating on the bracketed clauses alone in the case of 1a–1c):

1a I doubt [that anyone at all, you would dare complain to about my

cooking] [M]
1b I doubt [that my cooking, anyone at all, you would dare complain to about
it] [M]
1c *I doubt [that anyone at all, my cooking, you would dare complain to
about] [M]
2a Everything else, I hereby bequeath to my wife [S]
2b My wife, everything else, I hereby bequeath to her [S]
2c *Everything else, my wife, I hereby bequeath to [S]
3a Something special, he gave to everyone
3b Everyone, he gave something special to
3c *Something special, everyone, he gave to
4a Everything possible he has done for his sick mother
4b His sick mother, everything possible he has done for her
4c *Everything possible, his sick mother, he has done for
5a The rumours about everything else, he flatly denied [S]
5b ??Everything else, he flatly denied the rumours about [S]
5c *Everything else, the rumours about he flatly denied [S]
6a Everything about the president, I thoroughly dislike
6b The president, everything about him, I thoroughly dislike
6c *The president, everything about, I thoroughly dislike

Helpful hints
In each of the above sentences, concern yourself with the structure
of the clause periphery, but do not concern yourself with the internal
252 The clause periphery

structure of TPs below the periphery. Treat my cooking, my wife, his sick
mother, the president, and the rumours about everything else as DPs,
but do not attempt to analyse their internal structure; and treat anyone
at all, something special, everyone, everything else, everything possible,
and everything about the president/him as QPs, but do not attempt to
analyse their internal structure. In 5 and 6, treat the about-phrase as an
adjunct to the expression it modifies (so that in the rumours about him,
the PP about him is an adjunct to the NP rumours: this idea gains support
from one-pronominalisation in ‘Which rumours? The ones about Prince
Androgen?’).
In relevant sentences, consider the role played by one or more of the
following:

7 Distinctness Condition (adapted from Richards 2010: 5)

At PF, any sequence <α, α> of two adjacent non-distinct constituents is
degraded

8 Intervention Condition (Abels 2012: 247)

Likes cannot cross likes

9 Path Containment Condition (Pesetsky 1982a: 309)

If two paths overlap, one must contain the other (NB: paths = movement
paths)

10 Constraint on Extraction Domains/CED (adapted from Huang 1982)

Only complements allow material to be extracted out of them, not
specifiers or adjuncts

11 Specificity Condition (adapted from Fiengo & Higginbotham 1981)

Specific nominals are opaque domains for extraction (i.e. nominals
introduced by specific expressions such as the/this/that/my/your/his etc.
are resistant to having anything extracted out of them)

12 Freezing Principle (adapted from Wexler & Culicover 1980)

The constituents of a moved phrase are frozen internally within (and so
cannot be extracted out of) the moved phrase

Note that many of these conditions/principles/constraints are not dis-

cussed in this chapter (although some were introduced in §1.5.1), and
two – 9 and 11 – are introduced for the first time here, to test your ability
to determine how constraints which are unfamiliar to you block illicit
derivations.
 4.7 Workbook 253

EXERCISE 4.3

This exercise is designed to be attempted after reading Module 4.3 and the
modules and chapters preceding it.

Discuss the structure of the periphery of the clauses in the sentences

below (concentrating on the bracketed clauses alone in sentences
containing these):

1 I doubt [if Lord Scumme-Bagge, in his will, anything at all, he left to charity]
[M]
2 I know [that, generally, the FBI, in cases like this, they find the killer] [S]
3 Everton, if it wasn’t for their goalkeeper, Arsenal, they could have scored
eight (Darren Gough, talkSPORT Radio)
4 I feel [that the president, his supporters, everything possible they have done]
5 Do you think [that Paris Saint-Germain, with the money they have behind
them, that they’re the new powerhouse in French football]? (Alan Brazil,
talkSPORT Radio) [S]
6 I wonder [if, one Thursday, that you wouldn’t mind coming into the studio]
(Andy Goldstein, talkSPORT Radio)
7 It means that they can be enticed, [because sometimes, then, that they have
less of a bond with their home country] (Tim Vickery, BBC Radio 5) [S]
8 All of a sudden, in the big interviews with football journalists, that Sir Alex
has changed his tune (Neil Ashton, talkSPORT Radio)

Note that the use of that in sentences like 6–8 is non-standard, and some
speakers also don’t like the second occurrence of that in sentences like 5 –
though I have recorded numerous examples of structures like 5 and several
of structures like 6–8 from live, unscripted radio and TV broadcasts (see
Radford 2018).

Helpful hints
In the clauses you analyse, concern yourself with the structure of the
clause periphery, but not with the internal structure of TPs below the
periphery. Treat Lord Scumme-Bagge, the FBI, Everton, Arsenal, Paris
Saint-Germain, the president, and his supporters as DPs, but do not
attempt to analyse their internal structure; treat everything possible
as a QP, but do not attempt to analyse its internal structure; and treat
one Thursday as a PP (headed by a null preposition – perhaps a null
counterpart of on) and likewise treat all of a sudden as a PP (perhaps
with all being in spec‑PP), but do not concern yourself with the internal
structure of these PPs. Treat the if-clause in 3 and the because-clause in
254 The clause periphery

7 as a projection of the CONJ/subordinating conjunction if/because and

hence a CONJP, but do not concern yourself with the internal structure
of the if-clause in 3. However, treat if in 1 and 6 as an interrogative
complementiser/FORCE head, and assume that it has a null yes–no
question operator OpYNQ as its specifier (and that this is an ADVP). For the
purposes of this exercise, treat sometimes/then in 7 as ADVP constituents
(though they could more abstractly be taken to be PP constituents headed
by a null preposition).

EXERCISE 4.4

This exercise is designed to be attempted after reading Module 4.4 and the
modules and chapters preceding it.

Discuss the structure of the periphery of the clauses in the sentences

below (concentrating on the bracketed clauses alone in the case of
sentences containing these):

1 Some cases, I suspect that the FBI, they never solve [M]
2 The president, everyone says that *(he) is stubborn
3 The key file, he found that several parts of *(it), they had redacted [S]
4 I don’t think [that for the sake of your own well-being, that if you are in a
bilingual classroom, that once you have completed the homework in one
language, that you should have to do it all over again in the second one]
(example which Eiichi Iwasaki tells me that Jim McCloskey informed him that
Chris Potts devised)
5 He is now planning, in the next few months, to run marathons for charity [S]
6 An expensive diamond ring like this, a man like that, I don’t think would ever
buy for his wife (Radford 2018: 65) [S]
7 I think [previously, that in trying to protect his off stump, he flicked the ball
to the on side] (Ian Chappell, BBC Radio 5 Live Sports Extra)
8 I just think [Arsenal, honestly, that they are gonna fall behind if they don’t
sign a striker] (John Cross, talkSPORT Radio)
9 It is not impossible [for in future for Google’s G-mail to crash] ([Link].
com) [S]
10 It’s time [for now Rooney to go] (Darren Gough, talkSPORT Radio)

Note that the notation *(he)/*(it) in 2, 3 means that the sentence is gram-
matical if he/it is used, but ungrammatical if he/it is omitted. Note too that
the use of the complementisers that/for in sentences 7–10 is non-standard:
say what is odd about their use.
 4.7 Workbook 255

Helpful hints
In each of the relevant clauses, concern yourself with the structure of
the clause periphery, and with how nominals in the periphery are case-
marked; do not concern yourself with the internal structure of TPs below
the periphery. Make the following assumptions about the status of various
constituents, but don’t attempt to analyse their internal structure: the
FBI, the president, the key file, Arsenal are DPs headed by an overt or
null D; some cases, everyone, several parts of (it) are QPs headed by a
quantifier; for the sake of your own well-being, now, in the next few
months, in trying to protect his off stump, and in future are PPs headed
by an overt or null preposition (with now paraphrasable as ‘at the
present time’); the subordinate clauses if there is a problem, if you are
in a bilingual classroom, and once you have completed the homework in
one language are CONJP constituents (i.e. projections of a subordinating
conjunction/CONJ like if/once); an expensive diamond ring like this and
a man like that are ARTP constituents; honestly and previously are ADVP
constituents.
Bear in mind the following assumptions made about the clause
periphery in Module 4.4:

11 Assumptions about the clause periphery (= 56)

(i) Clauses which can contain peripheral topic or focus constituents
are complete clauses which have FORCEP as the highest and FINP
as the lowest projection in the periphery (with any other peripheral
projections being positioned between FORCEP and FINP)
(ii) Clauses which cannot contain peripheral force-marking, topicalised
or focused constituents are truncated clauses which have a
periphery that projects only as far as FINP
(iii) FINP recursion allows for FINP+MODP+FINP structures

In addition, in relevant clauses, consider the role played by the following:

12 Default Case Assignment (= 15)

A (pro)nominal which does not fall within the domain of (i.e. which
is not c‑commanded by) a local (i.e. nearby) case assigner receives
accusative case by default

13 Criterial Freezing Condition/CFC (= 3)

(i) A peripheral constituent must occupy a criterial position to be
interpretable at LF
(ii) A constituent which occupies its criterial position is frozen in place

14 Intervention Condition (= 21)

Likes cannot cross likes
256 The clause periphery

15 Impenetrability Condition (= 46)

Anything below the head H of the highest projection HP in the periphery
of a complete clause (i.e. one containing FORCEP) is impenetrable to
anything above HP

16 Constraint on Extraction Domains/CED (from §1.5.1)

Only complements allow material to be extracted out of them, not
specifiers or adjuncts (adapted from Huang 1982)

17 Freezing Principle (from §1.5.1)

The constituents of a moved phrase are frozen internally within (and so
cannot be extracted out of) the moved phrase (adapted from Wexler &
Culicover 1980)

18 Doubly Filled COMP Filter (= 41)

At the end of a syntactic derivation, any structure in which the edge
of a projection headed by an overt complementiser is doubly filled (i.e.
contains some other overt constituent) is filtered out as ungrammatical

19 COMP‑Trace Filter/CTF (from §1.5.2)

Any structure at the end of a syntactic derivation in which an overt
COMP/complementiser is immediately adjacent to and c‑commands a
trace (i.e. a gap/null copy left behind by a moved constituent) is filtered
out as ill-formed

20 Distinctness Condition (= 22)

Any structure at PF which contains a sequence of two non-distinct
constituents <α, α> is degraded

In relation to 19, assume that (for the purposes of CTF) two constituents are
immediately adjacent if one precedes the other, and there is no overt con-
stituent intervening between them.
5 More peripheral constituents

5.0 Overview

In this chapter we turn to look at what positions in the clause periphery are
occupied by a range of different types of peripheral constituent which were
not discussed in the previous chapter, including fronted negative constitu-
ents, wh‑constituents of various kinds, and inverted auxiliaries. We’ll see
that these can be accommodated within a Cartographic model in which the
clause periphery canonically comprises FORCEP and FINP projections, with
optional TOPP, FOCP and MODP constituents positioned between them.
However, two notes of caution should be sounded at the outset. The first
of these is that inter-speaker variation means that some sentences which
are fine for me may not be fine for some other native English speakers. By
way of illustration, I note that Emonds (1970) argued that some syntac-
tic operations are restricted to applying in root/main clauses, and conse-
quently do not apply in complement clauses. One such operation, Emonds
claimed, is (Subject–Auxiliary) Inversion, which (by virtue of being a root
operation) can apply in root/main clauses like those in (1a, 1b) below, but
not in italicised complement clauses like (1c, 1d) (the examples and judge-
ments being from Emonds 1970: 6–7):
(1) a. Never in my life have I spoken to him
b. When is he coming, and where is he from?
c. *John thought that Bill hadn’t come and that neither had Mary
d. *Mary doesn’t know why Susan in leaving, and we don’t know why
is she either

In these examples, Inversion is triggered by a peripheral negative constitu-

ent (never in my life/neither) in (1a, 1c), and so I’ll refer to the phenomenon
as Negative Inversion. By contrast in (1b, 1d) Inversion is triggered by
a peripheral interrogative constituent (when/where/why), and so I’ll refer
to the phenomenon as Interrogative Inversion. In Emonds’s, (American)
variety of English, both Negative Inversion and Interrogative Inversion are
restricted to occurring in root clauses. By contrast, in my own (British)
variety, Interrogative Inversion is restricted to applying in root clauses, but
Negative Inversion is not. Thus, for me, the italicised auxiliary-inverted
negative complement clause in (1c) is perfectly grammatical, but the
258 More peripheral constituents

italicised auxiliary-inverted interrogative complement clause in (1d) is not.

In the discussion in this chapter, some of the illustrative examples I use
(which are – naturally enough – based on my own variety of English)
will involve Negative Inversion in a complement clause: bear in mind that
speakers who treat Negative Inversion as a root phenomenon will find such
sentences (like those in 2c, 2d, 5, 7e, 15, 26, 27b, 29, 62, 66, and 69 below)
degraded. Note that this does not undermine the argumentation put for-
ward here, since theories of grammar (and grammars of English) have to
be able to account for the full range of structures found in all varieties of
English (including mine!).
A second note of caution is the following. In this module, I shall be seek-
ing to establish what combinations of different types of peripheral projec-
tion (e.g. FORCEP, TOPP, FOCP, MODP, FINP) are possible, and what kinds
of conditions/constraints rule out combinations that are not possible. This
means that some of the example sentences below have an extremely com-
plex periphery, and may feel somewhat contrived. This inevitably means
that they cause parsing difficulties, especially where they involve constitu-
ents moving to the periphery across other constituents moving to the per-
iphery, and where anyone parsing the sentence has to establish where each
moved peripheral constituent originated prior to movement. In the case of
structures with a complex periphery that is difficult to parse, many speakers
reject the sentences as unacceptable (perhaps because parsing such sen-
tences imposes a heavy memory load). However, the (unasterisked) complex
structures illustrated in this chapter are grammatical for me, especially if
read with a clear pause between each peripheral constituent.

5.1 Negative and Interrogative Inversion

In this module, we look at the syntax of Negative and Interrogative Inversion,

and attempt to determine the position occupied by the fronted negative or
interrogative constituent on the one hand, and the inverted auxiliary on
the other. Please bear in mind the cautionary note in Module 5.0 that my
variety of English allows Negative Inversion to apply in root and non-root
clauses alike, whereas some speakers restrict Negative Inversion to applying
in root clauses only.

5.1.1 Negative Inversion

We’ll begin by examining the position in the clause periphery occupied by
fronted negative constituents and inverted auxiliaries in a type of structure
(illustrated below) traditionally said to involve Negative Inversion:
 5.1 Negative and Interrogative Inversion 259

(2) a. Nothing would I ever do to upset you

b. On no account should you leave the country
c. I can assure you [that at no time did he misbehave in any way]
d. He was adamant [that not a single drop of alcohol had he touched]

Such structures contain an (italicised) fronted negative XP followed by a

(bold-printed) inverted auxiliary. But what position in the clause periphery
do the negative constituent and the inverted auxiliary occupy, and how do
they get there?
Rizzi (1997) and Haegeman (2012) argue that Negative Inversion involves
focusing of the fronted negative constituent. One reason for thinking this
is that (as we saw in §4.2.2), non-referential/non-specific pronouns like
something/anything/everything can be focused (but not topicalised) in Eng-
lish, and this suggests that fronting of the non-referential pronoun nothing
in (2a) is the result of focusing rather than topicalisation. This assumption
would also account for why nothing cannot be reprised by a (bold-printed)
resumptive pronoun in a sentence like (3) below:

(3) Nothing would I ever do (*it) to upset you

Since only topicalised (and not focused) constituents can be reprised by a

resumptive pronoun, the ungrammaticality of sentences like (3) when they
contain a resumptive is consistent with the peripheral negative constituent
nothing being focused.
A further reason for thinking that peripheral negative constituents are
focused is that (like typical focused constituents) they can be used to supply
new information in reply to a question, as in the dialogue below:

(4) s p ea k e r a : How much alcohol have you had?

s p ea k e r b : Not a single drop have I touched in the past 6 months

A third reason for treating peripheral negative constituents as focused is

that they can contain a focus marker like only or even, as in:

(5) a. It is unfortunate [that only one of the hostages have they released]
b. It is unfortunate [that not even one of the hostages have they released]

For reasons such as these, Rizzi and Haegeman suppose that the peripheral
negative constituent in such structures occupies spec‑FOCP. Accordingly,
in a sentence like (2a) ‘Nothing would I ever do to upset you’, the pronom-
inal QP nothing moves to spec‑FOCP from its initial position as the the-
matic complement of the verb do by Focus Movement (a subtype of A‑bar
Movement).
It seems clear that Auxiliary Inversion in Negative Inversion structures is
triggered by the focusing of a negative constituent, since we find Inversion
260 More peripheral constituents

in structures like (2, 4 b , 5) with a focused negative constituent, but not in

structures where other kinds of constituent are focused (as we saw in Mod-
ule 4.2). Moreover, there is no Auxiliary Inversion in structures like those
below which contain an unfocused peripheral negative constituent occupy-
ing the specifier position in a peripheral MODP projection:

(6) a. In no time at all, the fire reached the outskirts of the town
b. For no apparent reason, the dam burst
c. With no money, it is difficult to make ends meet
d. Not long ago, they discovered a new Inca temple using Lidar

Rizzi and Haegeman account for this triggering effect by supposing that
the fronted auxiliary moves to the edge of the same FOCP projection as the
focused negative constituent, so that FOC attracts an auxiliary to adjoin
to FOC when FOC attracts a focused negative constituent to become its
specifier.
However, sentences such as the following (illustrating a phenomenon
termed non-adjacent inversion by Haegeman 2000b) cast doubt on the as-
sumption that the inverted auxiliary moves to FOC in clauses containing a
preposed negative constituent:

(7) a. Nothing at all, when questioned, did he say

b. On no account, during the vacation, would I go into the office
(Haegeman 2000b: 133)
c. Under no circumstances, however desperate I was, would I ask her for
money
d. Not one of the wounded soldiers, hard though the surgeons tried,
could they save
e. He prayed that never again atrocities like these would he witness
(Haegeman 2012: 48, fn. 46)
f. On very few occasions, people like him, do they get the credit they
deserve

The reason is that the italicised fronted negative constituent (in spec‑FOCP)
is separated from the bold-printed inverted auxiliary by a variety of other
(underlined) peripheral constituents: these include a prepositional phrase
(in spec‑MODP) in (7b), a subordinate clause (again in spec‑MODP) in (7a,
7c, 7d), a fronted topic (in spec‑TOPP) in (7e), and a dislocated topic (again
in spec‑TOPP) in (7f). Sentences like those in (7) provide evidence that the
inverted auxiliary must be the head of a lower peripheral projection than
that containing the focused negative constituent. But what precisely is the
projection containing the inverted auxiliary?
Since the inverted auxiliary immediately precedes the subject I/he/they in
(7), and since FINP is the projection immediately above the TP containing
 5.1 Negative and Interrogative Inversion 261

the subject, a plausible answer is that the inverted auxiliary moves from T
to FIN via Head Movement. Evidence in support of this comes from the ob-
servation that the inverted auxiliary and its subject cannot be separated by
intervening constituents, as we see from the ungrammaticality which results
if we re-position the underlined peripheral constituents in (7) between the
(bold-printed) inverted auxiliary and its (italicised) subject, as in (8) below:

(8) a. *Nothing at all did, when questioned, he say

b. *On no account would, during the vacation, I go into the office
c. *Under no circumstances would, however desperate I was, I ask her
for money
d. *Not one of the wounded soldiers could, hard though the surgeons
tried, they save
e. *He prayed that never again would, atrocities like these, he witness
f. *On very few occasions, do, people like him, they get the credit they
deserve

Sentences such as (7, 8) thus suggest that a fronted negative constituent

moves to the specifier position in FOCP, whereas a fronted auxiliary moves
to the head FIN position of FINP.
However, this conclusion raises the question of why a fronted negative
constituent moving to spec‑FOCP should trigger movement of an auxiliary
from T to FIN, when FOCP and FINP are two separate peripheral projec-
tions. A plausible answer is that a focused negative constituent transits
through spec‑FINP on its way to spec‑FOCP, and that (at the point where
it is in spec‑FINP) it attracts an auxiliary in T to move to adjoin to FIN.
On this view, (7a) above would have the derivation shown below, if we
take when questioned to be a CONJP projection headed by the CONJ/sub-
ordinating conjunction when (perhaps an elliptical variant of when he was
questioned):

(9) FORCEP

FORCE FOCP
ø
QP FOC′
nothing
at all FOC MODP
ø
CONJP MOD′
when ques-
tioned MOD FINP
ø
QP FIN′
nothing
at all FIN TP
did+ø he did say nothing at all
262 More peripheral constituents

The focused negative QP nothing at all originates as the thematic comple-

ment of the verb say, and from there initially moves to become the specifier
of a null FIN head ø which concomitantly attracts the auxiliary did to move
from T to adjoin to FIN-ø. Subsequently, the focused negative QP moves
into its criterial position in spec‑FOCP, and lower copies of the focused QP
and the inverted auxiliary receive a null spellout (marked by strikethrough
above).

5.1.2 Interrogative Inversion

A further type of context in which we find Auxiliary Inversion is in
wh‑questions like those below, illustrating a phenomenon traditionally
termed Interrogative Inversion:

(10) a. Which dress was she wearing?

b. Who have they arrested?
c. How many cars does he own?
d. At what time are they arriving?

Rizzi (1997, 2001) and Haegeman & Guéron (1999) argue that the fronted
(italicised) wh‑XPs in Interrogative Inversion structures are focused and
move to spec‑FOCP. From a semantic perspective, such an analysis is far
from implausible: interrogative wh‑constituents ask for new information,
and in this respect they fulfil a similar role to other focused constituents
(which provide new information). Moreover, a focus analysis is also plaus-
ible from a typological perspective, since in some languages an (italicised)
fronted interrogative constituent can be followed by a (bold-printed) focus
particle, as the examples below illustrate (where foc denotes a focus par-
ticle, and agr an agreement affix):

(11) a. Manwa ad izra hassan?

who foc saw hassan
‘Who was it that Hassan saw?’
(Berber, Alami 2011)
b. Inā nḕ sukà jē?
Where foc have go?
‘Where is it that they have gone?’
(Hausa, Green 2007: 84)
c. Iyondi yo Kambale a-alangira?
who foc Kambale ag r -saw
‘Who is it that Kambale saw?’
(Kinande, Schneider-Zioga 2007: 412)
 5.1 Negative and Interrogative Inversion 263

It is plausible to take the italicised fronted wh‑constituent to be the spe-

cifier of the bold-printed focus particle in such sentences, and to treat the
focus particle as the head FOC constituent of a FOCP projection.
Further evidence in support of claiming that interrogative wh‑constituents
move to spec‑FOCP in Interrogative Inversion structures comes from the
ungrammaticality of sentences like the following:

(12) a. *What to whom did he say?

b. *What to nobody did he say?

If, as claimed here, fronted negative and interrogative XPs are focused in
clauses where Auxiliary Inversion takes place, then the ungrammaticality
of sentences like (12) can be attributed to them violating the constraint that
there can be no more than one FOCP projection in the periphery of a given
clause.
Let’s now turn to the question of where the inverted auxiliary is posi-
tioned in Interrogative Inversion structures. There is evidence that (as in
the case of Negative Inversion) the inverted auxiliary does not move into
the head FOC position of FOCP, but rather moves into the head position in
a lower FINP projection. Such an assumption would predict that other per-
ipheral constituents can be positioned between the focused interrogative
constituent and the inverted auxiliary – and this prediction is borne out by
sentences such as those in (13) below, where an italicised interrogative con-
stituent is separated from a bold-printed inverted auxiliary by an under-
lined peripheral constituent which is a PP modifier in (13a–13d), a reduced
subordinate clause in (13e), and a dislocated topic in (13f):

(13) a. What at present do you need? ([Link])

b. How long, in the light of this promotion, will you stay here? (Ross
1967: 31)
c. What else, with no money to speak of, could he possibly have done?
(Kayne 1998: 155, fn. 66)
d. Under what circumstances, during the holidays, would you go into
the office? (Sobin 2003: 193)
e. Why, when asked, did all the students I interviewed say they would
take ‘study drugs’ again? ([Link])
f. How often, that book, has he praised it in public? (Kayne 1998: 155,
fn. 66)

Such sentences suggest that Interrogative Inversion can be treated in much

the same way as Negative Inversion. If so, the fronted wh‑constituent will
first move into spec‑FINP (at that point, triggering the head FIN constituent
264 More peripheral constituents

of FINP to attract a finite auxiliary to move from T to adjoin to FIN), and

thereafter move on its own to spec‑FOCP. On this view, (13a) will have the
derivation shown below:

(14)

FORCEP

FORCE FOCP
ø
QP FOC′
what
FOC MODP
ø
PP MOD′
at present
MOD FINP
ø
QP FIN′
what
FIN TP
do+ø you do need what

We can handle Auxiliary Inversion here in the same way as for Negative
Inversion if we suppose that a null FIN head which attracts a focused nega-
tive or interrogative XP to become its specifier also attracts a T auxiliary to
adjoin to FIN (as shown by the dotted arrow in 14). Spec‑FINP being simply
a transit position, the focused QP what in (14) subsequently moves from
spec‑FINP into its criterial position in spec‑FOCP.
An interesting property of focus structures is that (as we see from the
clause whose structure is shown in highly simplified form in 15 below) a
TOPP containing an (italicised) peripheral topic can be positioned between
a FORCEP containing a (bold-printed) force-marking complementiser and a
FOCP containing an (underlined) focused constituent:

(15) He prayed [FORCEP that [TOPP atrocities like those [FOCP never again
[FINP would [TP he witness]]]]] (Radford 2004a: 329)

Accordingly, a prediction made by the analysis of Interrogative Inversion

sketched in (14) above is that in front of the focused interrogative wh‑
constituent we should be able to have an (in situ or ex situ) topic – and this
prediction is borne out by sentences like those below (where an underlined
topic is positioned in front of an italicised focused interrogative constituent):

(16) a. A book like this, to whom would you give? (Delahunty 1983)
b. That kind of pen, what can you use it for? (Radford 1988: 532)
 5.1 Negative and Interrogative Inversion 265

c. Tom, why would anyone want to meet? (Bianchi 1999: 179)

d. Tottenham, where do they go from here? (Ronnie Irani, talkSPORT
Radio)
e. This kind of performance, how do you think the fans will react to it?
(Mark Chapman, BBC Radio 5)
f. Agbonlahore, how sharp was he? (Jason Cundy, talkSPORT Radio)

On the assumptions made here, a sentence like (16a) will have the structure
shown in highly simplified form below:

(17) [FORCEP ø [TOPP a book like this [FOCP to whom [FINP would [TP you give]]]]]

On this view, the TOPP projection containing the topic a book like this
is positioned immediately below FORCEP, and the FOCP containing the
focused interrogative phrase to whom is positioned between TOPP and FINP.
Although a structure like (17) in which two different phrases (the ARTP
a book like this and the PP to whom) are both moved to the periphery by
A‑bar Movement aren’t too bad for me (perhaps because I am an experi-
enced linguist used to parsing structures with a complex periphery), many
speakers find don’t find them acceptable. One reason may be that they vio-
late the following constraint (from Abels 2012: 247) introduced in §1.5.1:

(18) Intervention Condition/IC

Likes cannot cross likes

The putative IC violation arises because one A‑bar moved argument crosses
another (in that the fronted ARTP a book like this crosses the intervening
fronted ARTP to whom, and both phrases are arguments of the verb give).
Another reason may be that they violate the following condition devised
by Pesetsky (1982a: 309), where the term ‘paths’ denotes ‘movement paths’:

(19) Path Containment Condition/PCC

If two paths overlap, one must contain the other

The constraint in (19) bars movement paths from cutting across each other –
a condition clearly violated in (16a), as shown schematically below:

(20)
[FORCEP ø [TOPP a book like this [FOCP to whom [FINP would [TP you give — —]]]]]

However, if the topic is generated in situ and reprised by a resumptive

pronoun like it (as in 16b ‘That kind of pen, what can you use it for?’), no
intervention or path containment violation arises.
The examples discussed above all involve short (clause-internal) move-
ment of an interrogative wh‑constituent to spec‑FOCP. But what about
266 More peripheral constituents

cases such as the following, which involve long (cross-clausal) movement

of an interrogative wh‑constituent:

(21) Where is it thought [that he will go]?

Let’s suppose that where is a wh‑PP with a meaning paraphrasable as ‘to

what place’. In (21), where originates as the complement of the verb go in
the bracketed complement clause, and ultimately ends up in spec‑FOCP
in the main clause. But what positions does the wh‑PP transit through?
Locality considerations restrict long-moved constituents to moving one
clause at a time – that is, first to the periphery of the closest clause con-
taining the wh‑QP, then to the periphery of the next closest clause, and so
on. In part, this is the consequence of the Impenetrability Condition, which
was given the following formulation in §4.4.1 (where above/below mean
‘c‑commanding’/‘c‑commanded by’, and a complete clause is one whose
periphery contains FORCEP):

(22) Impenetrability Condition/IC

Anything below the head H of the highest projection HP in a complete
clause is impenetrable to anything above HP

The condition in (22) rules out movement across the edge of FORCEP,
because this is the highest projection in the periphery and so IC would
prevent where from moving into the periphery of the main clause without
transiting through the edge of the FORCEP projection in the complement
clause. Thus, (22) would rule out the movement represented by the lower
(dotted) arrow below:

(23)
[FORCEP ø [FOCP where ø [FINP where is [TP it is thought
[FORCEP that [FINP he will go where]]]]]]

This is because where illicitly moves across the edge of the complement
clause FORCEP projection headed by that, in violation of the Impenetrability
Condition (22). However, this locality violation can be avoided if where
transits through the edge of the complement clause FORCEP, as shown
below:

(24)
[FORCEP ø [FOCP where ø [FINP where is [TP it is thought
[FORCEP where that [FINP he will go where]]]]]]

On this view, where moves first into spec‑FORCEP in the complement clause
(as shown by the dotted arrow), then into spec‑FINP in the main clause
 5.2 Wh‑questions without Interrogative Inversion 267

(where it triggers Auxiliary Inversion), and finally into its criterial position
in spec‑FOCP in the main clause. There is no locality violation in (24), since
where moves through the edge of FORCEP in the complement clause, and
not (illicitly) across the edge of FORCEP.
You should now be able to tackle Exercise 5.1.

5.2 Wh‑questions without Interrogative Inversion

In this module, we examine the syntax of wh‑questions used without

Auxiliary Inversion. We look at two types of these: on the one hand,
embedded wh‑questions (e.g. wh‑question clauses embedded as the com-
plement of some predicate), and on the other hand how come questions
(which have the idiosyncratic property that for most speakers, they don’t
allow Auxiliary Inversion even in main/root clauses).

5.2.1 Embedded wh‑questions

Given the arguments in Module 5.1 that root wh‑questions involve moving
an interrogative wh‑XP into spec‑FOCP, it might at first sight seem plaus-
ible to suppose that embedded wh‑questions like those bracketed below
have a parallel syntax, with the italicised interrogative wh‑XP occupying
spec‑FOCP:

(25) a. I wonder [where she has gone]

b. He enquired [which apartment she was living in]
c. He asked [how much time the test would take]
d. He was dying to find out [what hotel she was staying in]

However, there are a number of reasons for thinking that the italicised
wh‑XPs are not focused in such structures. For one reason, if they were,
we would expect them to trigger Auxiliary Inversion in the same way as
they do in root/main clause questions like (10) above. After all, negative
constituents which are focused in embedded clauses trigger Auxiliary
Inversion (in varieties of English like mine), as we see from structures like
those bracketed in (26) below:

(26) a. I reiterated [that under no circumstances would I accede to his

demands]
b. He swore [that never again would he get involved in family disputes]
c. Boris maintained [that in no way had he acted improperly]

The fact that there is no Auxiliary Inversion in embedded questions like

those bracketed in (25) makes it implausible that the fronted wh‑constituent
is focused.
268 More peripheral constituents

Further evidence in support of the conclusion that embedded wh‑

questions don’t involve wh‑focusing comes from the observation that they
can themselves contain an (italicised) focused constituent, as we see from
structures such as:

(27) a. He asked [why absolutely everyone the police had detained — ]

b. He asked [why nobody at all had the police detained — ]

The italicised QPs in (27) exhibit characteristics typical of focused constitu-

ents: for instance, they are associated with a gap in the comment TP, and
cannot be reprised by a resumptive pronoun – as we see from the ungram-
maticality of:

(28) a. *He asked [why absolutely everyone the police had detained them]
b. *He asked [why nobody at all had the police detained them]

Furthermore, the fronted QP nobody at all in (27b) has the further property
(typical of focused negative constituents) of triggering Auxiliary Inversion
(in varieties of English like mine). Since the italicised constituents in (27)
are focused and since the clause periphery can only contain one FOCP con-
stituent, it follows that the peripheral interrogative constituent why cannot
be focused in embedded questions like (27) if the italicised constituent fol-
lowing it is focused. (It should be noted that speakers who treat Auxiliary
Inversion as a root operation won’t accept sentences like 27b; and that
parsing problems caused by sentences which involve A‑bar Movement of
multiple XPs to the periphery will further reduce the acceptability of sen-
tences like 27 because they involve both Focus Movement of absolutely
everyone/nobody at all, and Wh‑Movement of why – though both sentences
are OK for me).
But if why is not in spec‑FOCP in (27), where is it positioned? A rea-
sonable supposition is that why is positioned on the edge of the leftmost
projection in the bracketed clause (i.e. on the edge of FORCEP) – and in-
deed this is what is claimed by Rizzi (1997) and Haegeman (2000b). This
assumption would predict that a wide range of other constituents can be
positioned below/after the FORCEP constituent containing the interrogative
wh‑XP – and this prediction is borne out (for speakers like me who don’t
restrict Auxiliary Inversion to applying in root clauses), as we see from the
(heavily contrived) embedded wh‑question structure shown in highly sim-
plified form below (which needs to be read with pauses between the differ-
ent peripheral constituents, to facilitate parsing):

(29) We must question [FORCEP why [MODP in a civilised society [TOPP drug addicts
[MODP when they ask for help [FOCP only rarely [FINP do [TP they get get
it]]]]]]]
 5.2 Wh‑questions without Interrogative Inversion 269

The assumption that interrogative wh‑constituents in embedded questions

are in spec‑FORCEP is plausible from a semantic perspective, since the
main function of FORCEP is to mark a clause as declarative/interrogative/
exclamative/imperative etc. in force/type. And in §1.4.3, we saw that the
following conditions determine how clauses are typed:

(30) Clause Typing Conditions

(i) A clause is interpreted as being of a given type (interrogative or
exclamative etc.) if it contains a peripheral clause-typing specifier of
the relevant type
(ii) An indicative clause is interpreted as declarative in type by default if
it does not contain a peripheral clause-typing specifier

It follows from condition (30i) that positioning why in spec‑FORCEP in (29)

will type the bracketed complement clause as a wh‑question.
A further argument in support of positing that interrogative wh‑
constituents occupy spec‑FORCEP in embedded questions can be formu-
lated in relation to the selectional properties of predicates. Question-asking
predicates (like ask, enquire, wonder etc.) select an interrogative clause as
their complement. Given the assumptions made here, this means that the
complement of an interrogative predicate must be a projection with an
interrogative specifier. This selectional requirement will be met in a struc-
ture like that bracketed in (29) above if why is positioned on the edge of the
highest (FORCEP) projection in the embedded clause, and if why thereby
immediately follows the predicate question.
The assumption that a wh‑XP in an embedded wh‑question occupies
spec‑FORCEP and that FORCEP has to be the highest projection in an em-
bedded clause in order to satisfy selectional requirements predicts that no
other constituents will be able to precede an interrogative wh‑XP in an
embedded wh‑question. So, although (as discussed earlier in relation to 16
above) a topic can precede an interrogative wh‑XP in a main clause ques-
tion, the same (topic+wh) order is ungrammatical in embedded questions
like those bracketed in (31b, 32b) below:

(31) a. I’m curious [FORCEP why [TOPP politicians [FINP people mistrust them]]
b. *I’m curious [TOPP politicians [FORCEP why [FINP people mistrust them]]

(32) a. I’m unsure [FORCEP why [FOCP absolutely everyone [FINP he mistrusts]]
b. *I’m unsure [FOCP absolutely everyone [FORCEP why [FINP he mistrusts]]

According to Rizzi (1997: 301), FORCEP has to be the highest projection

in embedded clauses (not TOPP or FOCP) because ‘verbs select for declara-
tives or questions, not for clauses with or without topic (or focus)’. In more
270 More peripheral constituents

general terms, this selectional requirement can be seen as the consequence

of a condition such as the following:

(33) Selection Condition

A selector and its selectee (e.g. a verb and its complement) must be as
close as possible (ideally, immediately adjacent); the further apart they
are, the more degraded the structure becomes

The conclusion to be drawn from our discussion of wh‑questions in

this section and the last is that in root/main-clause/direct speech wh‑
questions, the fronted interrogative wh‑XP is focused, and moves through
spec‑FINP (where it triggers Auxiliary Inversion) into its criterial position
in spec‑FOCP. By contrast, in embedded wh‑questions, the interrogative
wh‑XP is unfocused and moves directly into a different criterial position in
spec‑FORCEP. This suggests the following generalisation about the syntax
of interrogative wh‑XPs:

(34) Interrogative Generalisation

(i) A peripheral interrogative XP that triggers Auxiliary Inversion moves
through spec‑FINP (with FIN attracting a finite auxiliary to adjoin to
it) into its criterial position in spec‑FOCP
(ii) The criterial position for a peripheral interrogative XP that does not
trigger Auxiliary Inversion is spec‑FORCEP

It should be noted, however, that (34) is simply a descriptive stipulation,

and not a theoretical principle.

5.2.2 How come questions

Our discussion has so far neglected a somewhat idiosyncratic type of
wh‑question illustrated below:

(35) a. How come we can’t get things through Congress? (Barack Obama,
cited in [Link])
b. How come you’re not at work today? ([Link])
c. How come you told him? ([Link])
d. How come you’ve ended up here? ([Link])

Such questions are introduced by how come, and have a number of idiosyn-
cratic properties. One of these is that (in most varieties) how come doesn’t
trigger Auxiliary Inversion (unlike why in e.g. Why can’t we get things
through Congress?). Another is that how and come seem to have become
fused together to form a single word – and indeed, they are sometimes written
as such, as in the following conversation (from the online urbandictionary):

(36) Dave: yo, u hungry? Todd: no, u? Dave: howcome?

 5.2 Wh‑questions without Interrogative Inversion 271

There are good reasons for taking how come to function as a single word
in present-day English. Firstly, how come is invariable, in the sense that
how cannot be substituted by other wh‑words (as we see in 37a below),
and come cannot be substituted by other forms of the same verb (as we
see in 37b):

(37) a. How/*However/*When/*Where come he resigned?

b. How come/*comes/*came/*coming he resigned?

Secondly, the string how come is indivisible, in the sense that no other
constituents (like those italicised below) can intrude between the two
items:

(38) How (*in your view,) (*Charles,) come he resigned?

Thirdly, come cannot be deleted via an ellipsis operation like Sluicing (leav-
ing how on its own) in a sentence like (39) below (from Kim & Kim 2011: 4):

(39) You’re always grinning about something. How come/*How?

The fact that the string how come is invariable and indivisible is consistent
with it having become a single word in present-day English. (An important
caveat to note, however, is that the observations made about how come
above and below hold for most but not for all speakers: there is a great
deal of variation between speakers in how they use how come, discussed at
length in Radford 2018: 216–92.)
Having established that how come functions as a single word in collo-
quial English, let’s ask what category it belongs to. Collins (1991) argues
that it functions as a complementiser in contemporary English. However,
categorising how come as a complementiser proves problematic in certain
respects. For one thing, how come differs from typical interrogative com-
plementisers like if/whether in being able to occur in main/root/principal/
independent clauses like:

(40) How come/*If/*Whether Boris has resigned?

Moreover, unlike typical complementisers (but like typical wh‑words), how

come allows Sluicing/ellipsis of its complement in sentences such as:

(41) I suspected he had left, but I wasn’t sure why/how come/*whether/*if

he had left

Furthermore, like the adverb why (but unlike the complementisers whether/
if), how come can be modified by the adverb exactly, for example, in:

(42) I have no idea how come exactly/why exactly/*whether

exactly/*if exactly the server is working
272 More peripheral constituents

In addition, how come can be coordinated with other wh‑adverbs, as we see

from the internet-sourced examples below:

(43) a. I don’t know why, or where, or how come, but about 6 years
ago my dad decided we didn’t need ‘The Weasley’s’ in our life
([Link])
b. I spend many hours on the computer assessing why, when and how
come – Jose did not care (M. Sampedro-Iglesia, The Heroes Among
Us, WestBow Press, 2011: 32)

Thus, there are numerous respects in which how come behaves more like
the adverb why than like the complementisers whether/if. Let’s therefore
take it to be an ADVP.
But whereabouts in the clause periphery is how come positioned? An
interesting property of how come questions is that (as illustrated below)
how come can be followed by a range of other peripheral constituents:

(44) a. How come in the daylight we can only see the sky but when it’s
night time you can see space? ([Link])
b. How come, if this is such a huge problem, that there isn’t more
useful and legit information on how to overcome it? (nextscientist
.com)
c. How come the things you did last night, you can’t remember them?
(Radford 2018: 247)
d. How come at no point, even when asked to, did he apologise?
(Radford 2018: 243)

Thus, in (44a) how come precedes a MODP projection containing the per-
ipheral PP in the daylight; in (44b) how come precedes a MODP containing
an if-clause and a FINP containing that; in (44c), how come precedes the
topic DP the things you did last night; and in (44d), how come precedes a
FOCP containing the focused negative PP at no point, a MODP containing
the reduced subordinate clause even when asked to, and a FINP containing
the inverted auxiliary did. Why does how come precede other peripheral
constituents in sentences like (44)?
Since how come marks the clause it introduces as interrogative in force,
and since FORCEP is the highest projection in the clause periphery, a plaus-
ible answer is to suppose that how come occupies spec‑FORCEP – that is,
the specifier position in FORCEP. On this assumption, (44d) would have the
peripheral structure shown in simplified form below (if we take how come
to be an ADVP, and treat even when asked to as a CONJP headed by the
CONJ/subordinating conjunction when):
 5.2 Wh‑questions without Interrogative Inversion 273

(45) FORCEP

ADVP FORCE′
how
come FORCE FOCP
ø
PP FOC′
at no
point FOC MODP
ø
CONJP MOD′
even when
asked to MOD FINP
ø
PP FIN′
—
FIN TP
did+ø he did
at no point
apologise
Here, the null FIN head ø attracts the focused negative PP at no point to
become its specifier, and at the same time also attracts the T auxiliary did
to adjoin to it. Subsequently, the FOC head attracts the focused PP at no
point to move to its criterial position in spec‑FOCP. The overall clause is
typed as a wh‑question by virtue of its periphery containing the interroga-
tive clause-typing ADVP how come in spec‑FORCEP.
The FORCEP analysis of how come in (45) accounts for how come gener-
ally being the highest constituent in the periphery of the clause containing
it. For example, in the complement clause bracketed below, how come pre-
cedes the dislocated topic my passport and the PP modifier at the airport
and cannot follow them:

(46) a. It is unclear [how come, at the airport, my passport, they didn’t

stamp it]
b. *It is unclear [at the airport, my passport, how come they didn’t
stamp it]

This follows from the assumption that how come is in spec‑FORCEP, since
FORCEP is canonically the highest projection in the clause periphery (in
embedded clauses, at least: but see the discussion of 48 below).
A second property accounted for under the FORCEP analysis is that
root how come questions (unlike other root wh‑questions) don’t trigger
Auxiliary Inversion – as we see from sentences like (35) above. This can
be accounted for if we suppose that how come is directly merged in situ
in spec‑FORCEP, and that only an interrogative XP which is focused and
274 More peripheral constituents

transits through spec‑FINP triggers Auxiliary Inversion. (Accordingly, In-

version of the auxiliary did in 45 above is not triggered by how come but
rather by the focused negative PP at no point transiting through spec‑FINP
on its way to spec‑FOCP.)
A third property which can be accounted for if how come is directly
merged in situ in spec‑FORCEP is that (as observed by Collins 1991) how
come always has scope over the clause it introduces, never over an embed-
ded clause. In this respect, it differs from why – as illustrated below:

(47) a. Why did John say Mary left?

b. How come John said Mary left?

In (47a), why can either question the reason for John saying what he said,
or the reason for Mary leaving. This ambiguity can be accounted for if
why can originate either as an adjunct to the VP in the left-clause, or as
an adjunct to the VP in the say-clause: in either eventuality, why moves
through spec‑FINP in the main clause (triggering Auxiliary Inversion) into
spec‑FOCP. By contrast, in (47b) how come questions the reason for John
saying what he said, not the reason for Mary leaving. This follows if how
come is directly merged in spec‑FORCEP in the main clause and can only
have scope over the clause in which it originates (not over a lower clause).
Furthermore, because how come is directly generated in spec‑FORCEP (and
does not move there from spec‑FINP), it will not trigger Auxiliary Inversion.
However, a potential complication arising from the claim that how come
is directly merged in spec‑FORCEP is posed by the observation that in root
clauses, how come can be preceded by peripheral constituents like those
italicised below:

(48) a. Before you go to bed, after you’ve had a shower, how come you check
your email? (Radford 2018: 244)
b. On Thursday, how come there were no lectures? (Radford 2018: 247)
c. The mistakes we made when we were young, how come we repeat
them when we are old? (Radford 2018: 244)
d. The mayor of Trumpton, after campaigning so hard, how come he
didn’t get re-elected?

Thus, how come is preceded by two MODPs containing subordinate clauses

introduced by before/after in (48a), by a MODP containing the PP on
Thursday in (48b), by a TOPP containing the dislocated topic the mistakes
we made when we were young in (48c), and by a TOPP containing the dis-
located topic the mayor of Trumpton and a MODP containing after cam-
paigning so hard in (48d). Moreover, as we see from the (highly contrived)
example below, how come can be both preceded and followed by other
peripheral constituents within the same clause:
 5.2 Wh‑questions without Interrogative Inversion 275

(49) Humpty Dumpty, if he was so popular, how come his friends, at no

point, even after he fell off the wall, did they rally round him?

Here, how come is preceded by a TOPP containing the dislocated topic

Humpty Dumpty, and a MODP containing a subordinate clause introduced
by if; at the same time, how come is also followed by a TOPP containing
the dislocated topic his friends, a FOCP containing the focused negative PP
at no point, a MODP containing the subordinate clause even after he fell
off the wall, and a FINP containing the inverted auxiliary did. Of course,
the periphery of (49) is so complex that, to facilitate parsing, the sentence
needs to be read with a pause between successive peripheral constituents –
and even then, parsing such a complex structure remains difficult because
of the memory load required.
The question posed by our discussion above is how we can account for
this seemingly idiosyncratic behavior of how come. A plausible answer is to
maintain the FORCEP analysis of how come, but to posit that FORCEP is only
required to be the highest peripheral projection in complement clauses like
those bracketed in (46) because of requirement imposed by the Selection Con-
dition (33) for selector and selectee to be as close as possible – more specifically,
because the predicate unclear selects a complement headed by an interrogative
FORCEP. We can then suppose that in root/main clauses (since they are not se-
lected by any higher predicate), certain peripheral projections (including TOPP
and MODP projections) can be projected above FORCEP, leading to sentences
like (48) and (49) above. On this view, a sentence like (50a) below would have
the structure in (50b) (simplified by not showing the internal structure of FINP,
or the position within FINP in which at no point originates):

(50) a. Humpty Dumpty, how come, at no point did anyone help him?
b. TOPP

DP TOP′
Humpty
Dumpty TOP FORCEP
ø
ADVP FORCE′
how come
FORCE FOCP
ø
PP FOC′
at no
point FOC FINP
ø did anyone help him

The overall sentence would be interpreted as interrogative in force in accord-

ance with the Clause Typing Condition in (30i) by virtue of containing the
interrogative clause-typing specifier how come in its periphery.
276 More peripheral constituents

However, an alternative approach to accounting for data like (44–49)

above is offered by Shlonsky & Soare (2011). They argue that how come
is directly merged in situ as the specifier of an INTP/interrogative phrase
which is a medial peripheral projection (in the sense that it is positioned
between FORCEP and FINP): INTP allows one or more TOPP and MODP
constituents to be positioned above or below it, and also allows a FOCP and
a FINP projection below it. On this view, the order of the peripheral projec-
tions in (49) above is as shown in highly simplified form below (where ø is
a null FORCE head):

(51) [FORCEP ø [TOPP Humpty Dumpty [MODP if he was so popular [INTP how come
[TOPP his friends [FOCP at no point [MODP even after he fell off the wall
[FINP did [TP they rally round him]]]]]]]]]

However, there are two potential problems posed by the INTP analysis. One
is that it fails to account for why (even for speakers like me who allow
Auxiliary Inversion in non-root clauses) a structure like (51) can’t be embed-
ded as the complement of a predicate like unclear which selects an inter-
rogative FORCEP complement – as we see from the ungrammaticality of:

(52) *It is unclear [Humpty Dumpty, if he was so popular, how come his friends,
at no point, even after he fell off the wall, did they rally round him]

After all, under the INTP analysis in (51), the predicate unclear would
indeed select an interrogative FORCEP as its complement. By contrast,
under the FORCEP analysis of how come in (50), what’s wrong with (52)
is that unclear has a TOPP complement, and not the interrogative FORCEP
complement that it requires.
The second potential problem with the INTP analysis of how come is that
it is not in accordance with the Interrogative Generalisation in (34) above,
repeated as (53) below:

(53) Interrogative Generalisation

(i) A peripheral interrogative XP that triggers Auxiliary Inversion moves
through spec‑FINP (with FIN attracting a finite auxiliary to adjoin to
it) into its criterial position in spec‑FOCP
(ii) The criterial position for a peripheral interrogative XP that does not
trigger Auxiliary Inversion is spec‑FORCEP

The INTP analysis in (51) is clearly inconsistent with (53ii), which spe-
cifies that the criterial position for interrogative XPs that do not trigger
Interrogative Inversion is spec‑FORCEP. So, from this perspective at least,
it can be argued that the FORCEP analysis of how come in (45, 50b) is
preferable.
You should now be able to tackle Exercise 5.2.
 5.3 Yes–no questions 277

5.3 Yes–no questions

The analyses of root and embedded wh‑questions outlined in §5.1.2 and

§5.2.1 respectively raise the question of what happens in root and embed-
ded yes–no questions – an issue taken up in this module.

5.3.1 Root yes–no questions

There is evidence (outlined more fully in Radford 2016: 259–65, and
Radford 2020: 220–25) that root yes–no questions contain an abstract/null/
silent yes–no question operator, and that this operator is responsible for
triggering Auxiliary Inversion. The null question operator analysis gains
plausibility from the observation that Shakespearean English had main
clause yes–no questions containing an inverted auxiliary preceded by the
adverbial yes–no question operator whether, as illustrated below:

(54) a. Whether had you rather lead mine eyes or eye your master’s heels?
(Mrs Page, The Merry Wives of Windsor, [Link])
b. Whether dost thou profess thyself a knave or a fool? (Lafeu, All’s
Well that Ends Well, IV.v)

The postulation of an adverbial operator which received an overt spellout

as whether in earlier varieties of English but which has a null spellout in
present-day English raises the possibility that yes–no questions have a par-
allel syntax to that of wh‑question operators (like who, what, where, when
etc.). If so, this would mean that in root yes–no questions, the yes–no ques-
tion operator (which I will abbreviate as OpYNQ, and treat as an ADVP) is
focused, and moves from spec‑FINP (where it triggers Auxiliary Inversion)
into its criterial position in spec‑FOCP. On this view, the periphery of a
sentence like:

(55) Last night, did he go out?

would have the derivation below (where last night is treated as a PP

headed by a null temporal preposition with a meaning similar to on/dur-
ing, and the sentence is paraphrasable as ‘Is it the case that last night he
went out?’):
278 More peripheral constituents

(56) FORCEP

FORCE FOCP
ø
ADVP FOC′
OpYNQ
FOC MODP
ø
PP MOD′
last
night MOD FINP
ø
ADVP FIN′
OpYNQ
FIN TP
did+ø he did go out

Let’s make the following assumption about Auxiliary Inversion:

(57) Auxiliary Inversion
A null FIN head which has a focused interrogative/negative XP as its
specifier attracts a finite auxiliary in T to adjoin to FIN

On this view, FIN (at the point where it has the interrogative operator ADVP
OpYNQ as its specifier) attracts the T auxiliary did to adjoin to the null FIN
head. Subsequently, the yes–no question operator (being focused) moves to
its criterial position in spec‑FOCP. The resulting structure is then spelled out
at PF as (55), Last night, did he go out? (A possibility that I will set aside
here is that the MODP containing last night may alternatively occupy a
position above FOCP – or perhaps even above FORCEP.)
At first sight, it might seem implausible to suppose that a focused con-
stituent can be null. However, American Sign Language allows null in-
terrogative wh‑operators (Petronio & Lillo-Martin 1997). Likewise (as
noted by Yamakoshi 2002), wh‑drop in questions occurs in Child English
(illustrated by the first utterance in 58a below produced by a two-year-old
girl), and in colloquial adult Bavarian (illustrated by the omission of was
‘what’ in 58b, where 2p denotes second person plural, and p r t a discourse
particle – and angle brackets indicate constituents present in the syntax but
unpronounced in the phonology):
(58) a. <What> the porcupine doing? What the porcupine doing?
(Claire, 2;1; Hill 1983)
b. <Was> dea-ts-n es do?
what do-2p -prt you (t)here
‘What are you guys doing (t)here?’
(Bayer 2010: 32)
 5.3 Yes–no questions 279

The fact that we find null operators in contexts which require a focused
interrogative wh‑operator adds plausibility to the idea that a null question
operator can be focused.
A key claim made in the analysis in (56) above is that the yes–no ques-
tion operator originates as the specifier of FINP (in which position, it trig-
gers Auxiliary Inversion) and from there moves to its criterial position in
spec‑FOCP. Such an analysis predicts that the yes–no question operator
can only have scope over the clause containing it, not over a lower clause.
This claim is borne out by the observation that a question like Did John say
Mary left? can only question whether John said something, not whether
Mary left. This means that the null yes–no question operator is like how
come in (47b) in respect of only being able to have scope over the clause
containing it – but unlike why in (47a), in that why can originate in a clause
below that in which it ends up.
A piece of evidence in support of the focused interrogative operator analy
sis of yes–no questions is distributional in nature. As we saw in our earlier
discussion of wh‑questions, an interrogative wh‑operator can only be fo-
cused (and trigger concomitant Auxiliary Inversion) in a root clause ques-
tion, not in an embedded question like that bracketed in (59a) below; and
as expected under the focused yes–no question operator analysis, yes–no
questions show the same restriction against Auxiliary Inversion in embed-
ded questions, as we see in (59b):

(59) a. *The FBI finally discovered [where had the hostages been held]
b. *The FBI finally discovered [OpYNQ had the hostages been held in a
cave]

This can be accounted for if we suppose that Interrogative Inversion

involves focusing of an interrogative operator and is restricted to root
clauses. Embedded questions, by contrast, contain an unfocused interroga-
tive operator in spec‑FORCEP, and this operator (since it is not focused)
does not trigger Auxiliary Inversion – as we see from the (bold-printed)
uninverted auxiliary in:

(60) The FBI finally discovered [where the hostages had been held]

But what of the syntax of embedded yes–no questions?

5.3.2 Embedded yes–no questions

Embedded yes–no questions are introduced by the complementisers
whether/if, which have been argued to have an abstract yes–no ques-
tion operator as their specifier (see e.g. Radford 2016: 263–65, or Radford
2020: 223–35). If we treat embedded yes–no questions in the same way
as embedded wh‑questions, they will contain a null interrogative operator
280 More peripheral constituents

in spec‑FORCEP; and (as illustrated below) this operator will not trigger
Inversion of the bold-printed auxiliary because the (italicised) operator is
not focused:

(61) The FBI were hoping to discover [OpYNQ if/whether the hostages had been
held in a cave]

Since FORCEP can be followed by a wide range of other peripheral con-

stituents (including TOPP, FOCP and MODP), we would expect to find that
embedded yes–no questions can contain a wide range of peripheral con-
stituents following if/whether. And as the (heavily contrived) embedded
question structure bracketed below illustrates, this is indeed the case:

(62) We all need to probe deep into our consciences and constantly keep
asking [FORCEP OpYNQ whether [MODP in a civilised society [TOPP people like
that [MODP when they are in trouble [FOCP never ever [FINP should [TP we turn
our back on them]]]]]]]

On this view, embedded yes–no questions contain a null yes–no question

operator in spec‑FORCEP which serves to type the clause as a yes–no ques-
tion. (Two caveats should be noted here, however. Firstly, speakers who
treat Auxiliary Inversion as a root phenomenon will not accept structures
like 62; and secondly, the extreme complexity of the periphery of 62 makes
it difficult to parse, and for this reason many speakers will find it unaccept-
able: to aid parsing, it needs to be read with an audible pause between suc-
cessive peripheral constituents.)
The analysis outlined here accounts for why other peripheral constitu-
ents cannot be positioned between a predicate and a whether/if-clause that
it selects as its complement – as we see from the contrasts below:

(63) a. The weather forecasters are now unsure [if it will snow at Xmas]
b. *The weather forecasters are now unsure [at Xmas if it will snow]

(64) a. I’m curious [whether the second attack, anyone else witnessed it]
b. *I’m curious [the second attack, whether anyone else witnessed it]

In (63a, 64a) the adjectives unsure/curious immediately precede the if/

whether projection that serves as their complement, in accordance with the
locality requirement (imposed by the Selection Condition 33) for a selector
to be as close as possible to its selectee. By contrast, in (63b) an underlined
PP modifier (and in 64b an underlined topic) are positioned between the
selector (unsure/curious) and its selectee (the if/whether projection), in vio-
lation of the locality requirement on selection.
Considerations such as those outlined above suggest that the com-
plementisers if/whether occupy the head FORCE position of FORCEP in
 5.4 Exclamative and relative clauses 281

embedded yes–no questions, and have an abstract yes–no question opera-

tor as their specifier. This operator can serve to block extraction out of an
embedded FORCEP, for example, in a structure such as the following (where
how is construed as modifying react):

(65) *How did you wonder [FORCEP OpYNQ [FORCE whether] he would react — ]?

The presence of the yes–no question operator in the bracketed spec‑

FORCEP prevents how from moving out of the gap position and trans-
iting through spec‑FORCEP in the complement clause and spec‑FIN in
the main clause, on its way to the italicised spec‑FOCP position in the
main clause. So, the only option is for how to ‘jump across’ FORCEP and
move directly through spec‑FINP into spec‑FOCP in the main clause,
violating both locality requirements (since how illicitly moves across the
edge of FORCEP in violation of the Impenetrability Condition 22), and
the Intervention Condition (18) (since one interrogative operator crosses
another).
You should now be able to tackle Exercise 5.3.

5.4 Exclamative and relative clauses

Having looked at the syntax of wh‑questions, we now turn to look at the

syntax of two other types of clause that involve Wh‑Movement – namely
exclamatives on the one hand, and relatives on the other.

5.4.1 Exclamative clauses

Let’s begin by looking at the syntax of exclamative wh‑clauses. Rizzi (1997:
283) takes a FORCEP projection to serve the function of typing a clause as
declarative/interrogative/exclamative, etc. Accordingly, Radford (2004a:
332) suggests that ‘exclamative wh‑expressions … move into the specifier
position within FORCEP’, and thereby type the clause as exclamative. What
this would lead us to expect is that when a predicate selects an exclamative
clause as its complement, the locality requirement on selection will mean
that the exclamative FORCE head will be the highest head in the periphery
of the clause, and hence the exclamative wh‑constituent will precede any
topic, focus or modifier in the periphery of the clause. And indeed, this
claim seems to be borne out by an (admittedly contrived) sentence structure
such as the following (which needs to be read with a pause between adja-
cent peripheral projections, as an aid to parsing):

(66) It’s amazing [FORCEP in how many countries [TOPP that kind of behaviour
[FOCP under no circumstances at all [FINP would [TP it be tolerated]]]]]
282 More peripheral constituents

Here, amazing is a predicate which selects an exclamative clause as its

complement. Given the adjacency requirement on selection imposed by
the Selection Condition (33), this means that amazing must be immediately
followed by an exclamative FORCEP – as indeed it is in (66). The exclama-
tive FORCEP is in turn followed by a TOPP projection containing the dis-
located topic that kind of behaviour; this TOPP is itself followed by a FOCP
containing the focused negative PP under no circumstances at all; FOCP
is followed by a FINP containing the inverted auxiliary would; and FINP
is followed by a TP containing the subject it. (An important point to note is
that 66 will be unacceptable to speakers for whom Auxiliary Inversion is a
root phenomenon, and to speakers for whom its complex periphery makes
the sentence difficult to parse.)
Note that positioning some other constituent between the selecting
predicate amazing and the selected exclamative FORCEP containing in how
many countries results in degradation, as illustrated below:

(67) a. *It’s amazing [TOPP that kind of behaviour [FORCEP in how many
countries [FOCP under no circumstances at all [FINP would [TP it be
tolerated]]]]]
b. *It’s amazing [FOCP under no circumstances at all [TOPP that kind
of behaviour [FORCEP in how many countries [FINP would [TP it be
tolerated]]]]]

This is because sentences like (67) violate the Selection Condition (33), in
that the selector (amazing) is separated from the selectee (the exclamative
FORCEP containing the exclamative PP in how many countries) by one or
more intervening peripheral projections (by TOPP in 67a, and by FOCP
and TOPP in 67b). By contrast, in a root exclamative clause (which, by
its nature, is not selected by a higher predicate), there is no selectional
requirement for FORCEP to be the highest projection in the periphery –
and hence a structure such as the following is fine as a root/main clause
(if read with a pause between adjacent peripheral projections, as an aid to
parsing):

(68) [TOPP That kind of behaviour [FORCEP in how many countries [FOCP under no
circumstances at all [FINP would [TP it be tolerated]]]]]

In (68), the dislocated topic that kind of behaviour can be positioned

above the exclamative FORCEP containing in how many countries pre-
cisely because there is no higher predicate like amazing which requires the
exclamative FORCEP to be the highest projection in the periphery of the
clause containing it. (Of course, a familiar caveat needs to be made about
68, namely that it will be unacceptable to speakers for whom its complex
periphery makes the sentence difficult to parse.)
 5.4 Exclamative and relative clauses 283

5.4.2 Relative clauses

Having examined the structure of the clause periphery in interrogative and
exclamative clauses so far in this chapter, let’s conclude by taking a look
at the periphery of (restrictive) relative clauses. Rizzi (1997: 289) maintains
that ‘relative operators occupy the highest specifier position, the spec of
FORCE’. If FORCEP is indeed the highest/leftmost projection in the periph-
ery of a relative clause, it follows that other peripheral constituents (e.g.
topic, focus, or modifier constituents) will follow the FORCEP housing the
relativiser, as illustrated by the bracketed relative clause in the (heavily
contrived) sentence below:

(69) A university is the kind of place [where, that kind of behaviour, under
no circumstances, despite pressure from students, will the authorities
tolerate it]

As would be expected if FORCEP is the highest projection in a relative

clause, the FORCEP containing the italicised relative wh‑XP where pre-
cedes a (bold-printed) topic that kind of behaviour, an (underlined) focused
negative phrase under no circumstances, an (italicised) PP modifier despite
pressure from students, and a bold-printed inverted auxiliary will. On this
view, the bracketed relative clause in (69) will have the peripheral structure
shown in skeletal form below:

(70) A university is the kind of place [FORCEP where [TOPP that kind of behaviour
[FOCP under no circumstances [MODP despite pressure from students [FINP will
[TP the authorities tolerate it]]]]]]

We can then suppose that the FORCEP projection housing relative where
serves to mark the overall clause as relative in type, as well as marking the
clause as declarative in force. (Of course, the caveat should be made about
69/70 that it will be unacceptable to speakers for whom Auxiliary Inversion
is a root phenomenon, and to speakers for whom its complex periphery
imposes a heavy processing load which makes the sentence extremely dif-
ficult to parse.)
However, the FORCEP analysis of relative clauses proves problematic in
certain respects. For one thing, it is at variance with the following principle
posited by Cinque & Rizzi (2008), introduced in Module 4.1:

(71) One Feature One Head Principle

Each functional head carries only one interpretable feature

The principle in (71) rules out the possibility of an analysis in which a

FORCE head marks both relative clause type and declarative force.
An additional problem with positing that relative clauses are projections
of a FORCE head which marks both the relative type and the declarative
284 More peripheral constituents

force of the clause is that not all relative clauses are declarative in force.
For example, the bracketed relative clauses in (72) below are interrogative
in force:

(72) a. Mike’s leadership was second to none, [without which how else
could six novices laugh our way through ten days at minus 40?]
([Link])
b. It’s one of them situations now [where Harry, what does he do?] (Ray
Parlour, talkSPORT Radio)
c. Then you launched these podcasts, [which, how many times have
these been downloaded]? (Richard Bacon, BBC Radio 5)
d. This is a vital piece of evidence [which how come the police
overlooked?]

Likewise, the relative clauses bracketed in (73) below are imperative in

force:

(73) a. The top speed [which please don’t try to reach] is 220 miles an hour
(Ferrari test driver, BBC Radio 5)
b. You will please correct the false statement made last week in your
next issue, [after which please don’t handle my name at all] (Robin
Sterling 2013, Tales of Old Blount County, Alabama, Google Books,
p. 278)
c. You will start with completing their application form, [which please
be careful to triple check …] ([Link])
d. Just a couple of comments, Peter. One regards the spelling of Kanté,
[on which see the attached screenshot] (email from me to Peter
Trudgill)

And those bracketed in (74) below are exclamative in force:

(74) a. We were put up in a hotel [where, when we arrived, what a lovely

surprise there was waiting for us!]
b. It was a war [in which, what unspeakable acts of brutality people
committed in the name of freedom!]
c. It was a time [when, how happy people were to lead a simple life!]
d. It was something [which, how often he had dreamt about!]

Such examples suggest that relative clauses can be not only declarative in
force, but alternatively interrogative, imperative or exclamative; this leads
us to the conclusion that force and type are different features which (under
the One Feature One Head Principle) must be realised on different heads.
One way of implementing this idea is to suppose that on top of the
FORCEP constituent which marks a relative clause as declarative/inter-
rogative/imperative etc. in force, there is a separate projection marking
 5.4 Exclamative and relative clauses 285

the clause as relative in type, which I will denote as RELP; and indeed
this is the position argued for in more recent work by Rizzi (2005, 2013a,
2015a) and Radford (2019). This RELP constituent will be the highest
projection in the periphery of a relative clause (in order to be adjacent
to its antecedent in non-extraposed restrictive relatives), and will house
a wh‑XP which serves to type the clause as relative. If so, relative oper-
ators like who/which/where/when will not be positioned on the edge of
FORCEP, but rather on the edge of RELP – as shown in skeletal form for
(74a) above in (75) below:

(75) We were put up in a hotel [RELP where [MODP when we arrived [FORCEP what a
lovely surprise [FINP there was waiting for us!]]]]

Although the (italicised) relative operator is often overt (like which in 76a
below), the edge of RELP can sometimes contain a null relative operator
(below shown as OpREL) in place of an overt wh‑operator like which, as in
(76b). Likewise, the (bold-printed) head REL constituent of RELP can either
be spelled out overtly (as with that in 76b below) or receive a null spellout
as ø (as in 76a):

(76) a. It was something [RELP which [REL ø] how often he had dreamed about!]
b. It was something [RELP OpREL [REL that] how often he had dreamt
about!]

On the assumptions made here, the relative clauses bracketed in (76) above
will have the peripheral structure shown in simplified form in (77) below
(where — represents the gap left behind by movement of the relative oper-
ator which/OpREL to spec‑FORCEP, and the relative operator is treated as a
pronominal DP (like the relative pronoun lequel ‘[Link]’ in French):

(77) RELP

DP REL′
which/
OpREL REL FORCEP
that/ø
ADVP FORCE′
how often
FORCE FINP
ø
FIN TP
ø he had dreamt about—

The relative operator in spec‑RELP can be either overt or null, and serves
to type the overall structure as a relative clause. The REL head can likewise
either be spelled out as the overt complementiser that, or as a null head ø.
286 More peripheral constituents

However, RELP cannot have both an overt specifier and an overt head, as
we see from the ungrammaticality of:

(78) *It was something [RELP which that [FORCEP how often he had dreamed
about!]]

A structure like (78) is ruled out because it violates a filter introduced in

§1.5.2, where it was given the following formulation:

(79) Doubly Filled COMP Filter/DFCF

At the end of a syntactic derivation, any structure in which the edge
of a projection headed by an overt complementiser is doubly filled (i.e.
contains some other overt constituent) is filtered out as ungrammatical

The DFCF violation arises in (78) because the overt relative complementiser
that has the overt specifier which, leading to the ungrammatical collocation
*which that.
An issue which arises from the structure in (77) relates to the movement
of the relative operator (either which or its null counterpart) from being the
complement of about to becoming the specifier of the REL head. One ques-
tion which arises is whether the arrowed Relative Movement operation in
(77) satisfies locality requirements on movement, such as the Impenetrabil-
ity Condition in (22) above, repeated as (80) below:

(80) Impenetrability Condition

Anything below the head H of the highest projection HP in a complete
clause is impenetrable to anything above HP

The assumption embodied in (80) allows which (or its null operator coun-
terpart) to move directly to spec‑RELP across the edge of FINP and across
the edge of FORCEP, without violating any locality requirement: the edge
of FORCEP and FINP are not barriers to movement in (77) because neither
is the highest projection in the periphery (the highest peripheral projection
being RELP).
A second question which arises in relation to the derivation in (77) is
whether the relative operator (in crossing the exclamative phrase how of-
ten) violates the following constraint (repeated from 22 above):

(81) Intervention Condition

Likes cannot cross likes

We might expect some kind of intervention effect, since if how often moves
to the periphery from some position internally within TP, this would mean
that one fronted wh‑constituent (which) moves across another (how often).
Still, for me at least, the intervention effect is not particularly strong
here: this may be because the two types of wh‑constituent are different
 5.4 Exclamative and relative clauses 287

(e.g. which is relative, and how often is exclamative; and which is an argu-
ment but how often is an adjunct).
The relative clause structures we have looked at so far can be termed
gap relatives, since they all involve an ex situ relative operator moving to
spec‑RELP from some position within TP, leaving a gap behind in the pos-
ition out of which it moves. However, alongside gap relatives we also find
resumptive relatives in which an (overt or null) relative operator is reprised
by an (underlined) resumptive pronoun, so resulting in relative clauses like
those bracketed below:

(82) a. King Kong is a movie [which you’ll laugh yourself sick if you see it]
(Ross 1967: 433)
b. There really may be a lot of interaction [which the kid doesn’t hear
any of it] (Prince 1995: 4)
c. We were first introduced by a friend [who, we will try to protect her
privacy] (Prince Harry, BBC Radio 5)
d. This is a girl [who, in real life, you’d never let her have the keys to
your house, right?] (Jeremy Clarkson, BBC2 TV)
e. I really liked flying in an airplane [that I understand how it works]
(Bever et al. 1976: 150)
f. We’re afraid of things [that we don’t know what they are] (Ferreira &
Swets 2005: 263)
g. He’s a guy [he’s got an incredible knowledge of the game] (Steve
Smith, talkSPORT Radio)
h. Dimitar Berbatov, he’s the kind of player [he scores those fancy lobs]
(Connor MacNamara, BBC Radio 5)

If we treat such structures in a parallel fashion to the in situ resumptive

topic structures discussed in §4.1.2, they will involve an overt or null rela-
tive wh‑operator directly merged in situ in spec‑RELP being reprised by an
underlined resumptive pronoun within TP. In (82a–82d) the edge of RELP
comprises an overt relative operator (which/who) which is the specifier of
a null REL head; in (82e–82h) the edge of RELP comprises a null relative
operator which is the specifier of a REL head which is spelled out as the
overt relative complementiser that in (82e, 82f), and as a null relative com-
plementiser in (82g, 82h). Thus, the RELP projections in (82a, 82c, 82e, 82g)
will have the structures shown in highly simplified form below:

(83) a. … a movie [RELP which [REL ø] you’ll laugh yourself sick if you see it]
b. … a friend [RELP who [REL ø] we will try to protect her privacy]
c. … an airplane [RELP OpREL [REL that] I understand how it works]
d. … a guy [RELP OpREL [REL ø] he’s got an incredible knowledge of the
game]
288 More peripheral constituents

Where an overt relative operator is used, it is spelled out in its default form
(e.g. the relative operator is spelled out as who in 82c and 82d because w h o
is spelled out as whose if genitive, whom if accusative in formal styles, and
who otherwise). For a monograph-length treatment of the syntax of relative
clauses in everyday English, see Radford (2019).
You should now be able to tackle Exercise 5.4.

5.5 Summary

This chapter began in Module 5.1 by looking at the phenomenon of Negative

Inversion (in structures like No evidence did they offer) and Interrogative
Inversion (in sentences like What evidence did they offer?). I argued that
the fronted negative/interrogative XP is focused in both types of structure:
it originates internally within TP (here, as the complement of offer), and is
attracted by a FIN head to transit through spec‑FINP (with the FIN head
concomitantly triggering Auxiliary Inversion and attracting a finite T aux-
iliary to adjoin to FIN), before subsequently being attracted by a FOC head
to move to its criterial position in spec‑FOCP.
In Module 5.2, I looked at wh‑questions that don’t trigger Auxiliary
Inversion. I started by examining complement clause questions (like that
italicised in ‘I wonder what evidence they offered’), and I argued that in
such cases, the wh‑QP what evidence originates within TP (here, as the
complement of offered) and from there moves directly into spec‑FORCEP,
where it types the relevant clause as a wh‑question. I went on to look at
(main and embedded clause) how come questions like How come you were
late? I suggested that these involve how come functioning as an ADVP
which is directly merged in situ in spec‑FORCEP, and does not move there
from spec‑FINP (so does not trigger Auxiliary Inversion).
In Module 5.3, I turned to look at the syntax of yes–no questions. I
started by looking at root yes–no questions with Auxiliary Inversion (like
Did they offer any evidence?), and argued that these involve a focused (but
silent) yes–no operator/OpYNQ which originates in spec‑FINP (where it trig-
gers Auxiliary Inversion) and from there moves into its criterial position
in spec‑FOCP. I went on to discuss embedded yes–no questions like that
italicised in ‘I wonder if/whether they offered any evidence.’ I argued that
these are FORCEP constituents headed by the interrogative complementis-
er if/whether, and that they contain a null yes–no question operator in
spec‑FORCEP.
In Module 5.4, I examined exclamative and relative clauses. I began by
looking at exclamatives (like How little evidence they offered!), and I argued
that these involve a FORCE head attracting an exclamative wh‑XP (like how
 5.5 Summary 289

little evidence) to move from its initial position inside TP (here, as the com-
plement of offer) into its criterial position in spec‑FORCEP, where it serves
to type the clause as exclamative in force. I then turned to look at relative
clauses, and I argued that these are RELP constituents which contain within
them a separate FORCEP constituent which is usually declarative in force,
but can alternatively be interrogative, exclamative or imperative. I argued
that the head REL constituent of RELP contains an overt or null comple-
mentiser, and that REL has an XP specifier which contains an overt or null
relative operator that serves to type the clause as relative, and to link the
relative clause to its antecedent. This relative operator moves into spec‑RELP
from its initial position within TP in gap relatives like that italicised in ‘He’s
someone who very few people like — ’; but the relative operator is generated
in situ in and reprised by a (bold-printed) resumptive constituent in resump-
tive relatives like ‘He’s someone who very few people like him’.
Constructs and operations which have played an important role in our
discussion in this chapter include the following:

(18) Intervention Condition/IC

Likes cannot cross likes

(19) Path Containment Condition/PCC

If two [movement] paths overlap, one must contain the other

(22) Impenetrability Condition

Anything below the head H of the highest projection HP in a complete
clause is impenetrable to anything above HP

(30) Clause Typing Conditions

(i) A clause is interpreted as being of a given type (interrogative or
exclamative etc.) if it contains a peripheral clause-typing specifier of
the relevant type
(ii) An indicative clause is interpreted as declarative in type by default if
it does not contain a peripheral clause-typing specifier

(33) Selection Condition

A selector and its selectee (e.g. a verb and its complement) must be as
close as possible (ideally, immediately adjacent); the further apart they
are, the more degraded the structure becomes

(34) Interrogative Generalisation

(i) A peripheral interrogative XP that triggers Auxiliary Inversion moves
through spec‑FINP (with FIN attracting a finite auxiliary to adjoin to
it) into its criterial position in spec‑FOCP
(ii) The criterial position for a peripheral interrogative XP that does not
trigger Auxiliary Inversion is spec‑FORCEP
290 More peripheral constituents

(57) Auxiliary Inversion

A null FIN head which has a focused interrogative/negative XP as its
specifier attracts a finite auxiliary in T to adjoin to FIN

(71) One Feature One Head Principle

Each functional head carries only one interpretable feature

(79) Doubly Filled COMP Filter/DFCF

At the end of a syntactic derivation, any structure in which the edge
of a projection headed by an overt complementiser is doubly filled (i.e.
contains some other overt constituent) is filtered out as ungrammatical

5.6 Bibliographical notes

On Negative Inversion in standard varieties of English see Hooper &

Thompson (1973), Radford (1988, 2009a, 2009b), Authier (1992), Haegeman
(1995, 2000c), Sobin (2003), Büring (2004), Jacobsson (2007), Maekawa
(2007), de Clerk (2010), Haegeman (2012), Collins & Postal (2014), Radford &
Iwasaki (2015), Francis (2017), Jiménez-Fernández (2018), and Martín (2019).
On wh‑questions see Chomsky (1957, 1968, 1972, 1973, 1975, 1977, 1980,
1981, 1986a), Katz & Postal (1964), Jacobs & Rosenbaum (1968), Baker (1970),
Emonds (1970), Bresnan (1970, 1972, 1976, 1979), Chomsky & Lasnik (1977),
Baltin (1982, 2010), Erteschik-Shir (1986), Pesetsky (1989), Lasnik & Saito
(1992), Henry (1995), Jayaseelan (1996, 2003), Hallman (1997), Rizzi (1996,
1997, 2000a, 2000b, 2001, 2004a, 2004b, 2005, 2006a, 2006b, 2006c, 2010,
2011, 2012, 2013a, 2013b, 2013c, 2014a, 2014b, 2015a, 2015b), Haegeman &
Guéron (1999), Koopman (2000), Haegeman (2000b, 2012), Sobin (2003), Riz-
zi & Shlonsky (2006, 2007), Cable (2007, 2008, 2010a, 2010b), Cinque & Rizzi
(2008, 2010a), Goedegebuure (2009), Shlonsky & Soare (2011), Bayer (2014,
2015, 2016), Rizzi & Cinque (2016), Endo (2014, 2015a, 2015b, 2017, 2018),
Rizzi & Bocci (2017), Radford (2018), and Badan & Crocco (2019). On how
come questions, see Zwicky & Zwicky (1971), Collins (1991), Ochi (2004),
Fitzpatrick (2005a), Conroy (2006), Kim & Kim (2011), Shlonsky & Soare
(2011), Claridge (2012), Endo (2014, 2015a, 2015b, 2017, 2018), and Radford
(2018). On yes–no questions, see Katz & Postal (1964), Bresnan (1970), Lar-
son (1985), Grimshaw (1993), Roberts (1993a), den Dikken (2006), Haegeman
(2012), and Bianchi & Cruschina (2016).
On exclamatives see Elliot (1974), Milner (1978), Grimshaw (1979),
Oomen (1979), Gérard (1980), Radford (1980, 1982, 1989, 1997c, 2016,
2020), Huddleston (1993), Obenauer (1994), Benincà (1995, 1996), Michaelis
& Lambrecht (1996), Portner & Zanuttini (2000), Zanuttini & Portner (2000,
2003), Gutiérrez-Rexach (2001, 2008), Michaelis (2001), Villalba (2001,
 5.6 Bibliographical notes 291

2003, 2008), Ambar (2002), d’Avis (2002, 2016), Beijer (2002), Munaro
(2003), Collins (2004), Castroviejo (2006, 2007, 2008, 2010, 2019), Lipták
(2006), Miró (2006, 2008), Ono (2006), Ono & Fujii (2006), Sæbø (2006,
2010), Abels (2007, 2010), Marandin (2008), Beyssade (2009), Brandner
(2010), Delsing (2010), Jónsson (2010), Yamato (2010), Honda (2011),
Kalinina (2011), Chernilovskaya et al. (2012), Rett (2012), Zevakhina (2013),
Chernilovskaya (2014), Delfitto & Fiorin (2014), Badan & Cheng (2015),
Nouwen & Chernilovskaya (2015), Siemund (2015), Bosque (2017), Zubi-
aguirre Sebastián (2017), Palacios Roman (2018), Trotzke & Giannakidou
(2019), Vishenkova & Zevakhina (2019), Al-Bataineh (2020), and de Moraes
Sieiro (2020).
On relative clauses see Quirk (1957), Kuroda (1968), Morgan (1972),
Vergnaud (1974), Hirschbühler (1976), Carlson (1977), Bresnan & Grimshaw
(1978), Emonds (1979), Erdmann (1980), Ihalainen (1980), Weisler (1980),
Harris & Vincent (1980), Cornilescu (1981), Groos & van Riemsdijk (1981),
McCawley (1981), Shnukal (1981), Shorrocks (1982), Hirschbühler & Rivero
(1983), Borer (1984), Kekalainen (1985), Safir (1986b, 1999), Heim (1987),
Kjellmer (1988, 2010), Miller (1988), Kameshima (1989), Fabb (1990), Le
beaux (1991), Borsley (1984, 1992, 1997, 2001), de Vries (1993), Åfarli
(1994), Doherty (1994), Jacobsson (1994), Rooryck (1994), Guy & Bayley
(1995), Keenan (1985), Prince (1995), Seppänen & Kjellmer (1995), Grosu
(1996, 2003), Dubinsky (1997), Reid (1997), Sag (1997), Seppänen (1997),
Sigley (1997), Tottie (1997a, 1997b), Geisler (1998), Grosu & Landman
(1998), Newbrook (1998), Sauerland (1998, 1999, 2003), Wiltschko (1998),
Bianchi (1999, 2000), Comrie (1999, 2002), Khalifa (1999), Alexiadou et al.
(2000), Citko (2000, 2001, 2002, 2004, 2008), Zwart (2000), Izvorski (2001),
Munaro (2001), Akiyama (2002), Bhatt (2002, 2015), Caponigro (2002, 2004),
de Vries (2002, 2006), Tagliamonte (2002), Aoun & Li (2003), Benincà (2003,
2012a, 2012b), del Gobbo (2003), Herrmann (2003, 2005), Loock (2003,
2005, 2007a, 2007b, 2010), Arnold (2004, 2007), Authier & Reed (2005),
Adger & Ramchand (2005), Cable (2005), Cecchetto (2005), Heycock (2005),
Donati (2006), Hulsey & Sauerland (2006), van Riemsdijk (2006), Fiorentino
(2007), Inada (2007), Kayne (2007, 2015), Parry (2007), Cinque (2008, 2011,
2013, 2017b, 2020), Herdan (2008), Sportiche (2008, 2015, 2017), Caponigro
& Pearl (2008, 2009), Castillo (2009), Yang (2009), Berizzi (2010), Berizzi &
Rossi (2010), Cecchetto & Donati (2010, 2015), Donati & Cecchetto (2011),
Benincà & Cinque (2010, 2014), Bertollo & Cavall (2012), Hackl & Nissen-
baum (2012), Simonin (2012), Sistrunk (2012), Boef (2013), Caponigro et
al. (2013), Suárez-Gómez (2014), Collins (2015), Collins & Radford (2015),
Haegeman et al. (2015), Hinrichs et al. (2015), Kalivoda & Zyman (2015),
Sportiche (2015, 2017), Deal (2016), Douglas (2016), Patterson & Caponigro
(2016), Radford (2016, 2019), Burke (2017), and Wen (2020).
292 More peripheral constituents

On resumptive relatives (and the use of resumptives more generally),

see Kroch (1981), Chao & Sells (1983), Sells (1984, 1987), Engdahl (1985),
Prince (1990, 1995), Demirdache (1991), Contreras (1991), Erteschik-Shir
(1992), Shlonsky (1992), Bondaruk (1995), Dickey (1996), Suñer (1998),
McDaniel & Cowart (1999), Sharvit (1999), Aoun (2000), Aoun et al. (2001),
McKee & McDaniel (2001), Alexopoulou & Keller (2002, 2007), Cresswell
(2002), McCloskey (2006a, 2017a, 2017b), Rouveret (2002, 2011), Boeckx
(2003), Asudeh (2004, 2011a, 2012), Merchant (2004b), Cann et al. (2005),
Ferreira & Swets (2005), Grolla (2005), Szczegielniak (2005), Šímová (2005),
Alexopoulou (2006, 2010), Salzmann (2006, 2011), Bianchi (2008), Bošković
(2009), Friedmann et al. (2009), Nakamura (2009), Omaki & Nakao (2010),
Heestand et al. (2011), Keffala (2011), Keffala & Goodall (2011), Clemens et
al. (2012), Han et al. (2012), Beltrama (2013), Hofmeister & Norcliffe (2013),
Polinsky et al. (2013), Ackerman et al. (2014), Müller (2014), Sichel (2014),
Cerron-Palomino (2015), Francis et al. (2015), Hladnik (2015), Beltrama &
Xiang (2016), Blythe (2016), Guz (2017), Keshev & Meltzer-Asscher (2017),
Morgan & Wagers (2018), Checa-García (2019), Radford (2019), and Loss &
Wicklund (2020).

5.7 Workbook

In the exercises below, an [M] after an example sentence indicates that a

model answer is provided in the free-to-download Students’ Answerbook
(and in the Teachers’ Answerbook as well). You will find it helpful to look
at this model answer before attempting to tackle other examples in the
exercise. Some other exercise examples have an [S] after them, indicat-
ing that they are for self-study: this means that after reading the rele-
vant module (and those preceding it), and looking at the helpful hints
and model answer associated with the exercise, students should be able to
analyse these self-study examples on their own, and then compare their
answers with those provided in the Students’ Answerbook. The remaining
exercise examples (i.e. those not marked with [M] or [S] after them) are
intended for teachers to use as the basis for hands-on problem-solving
work in seminars, classes, or assignments: answers to all of these non-
self-study examples are provided in the Teachers’ Answerbook. Note that
Exercise 5.1 tests you on Module 5.1, Exercise 5.2 on Module 5.2, and so
on; and that each exercise in a given chapter presupposes familiarity with
material covered in earlier chapters and in earlier modules of the same
chapter. Unparenthesised numbers like 4 refer to material in the exercises,
whereas parenthesised numbers like (22) refer to examples and conditions
in the main text of the book.
 5.7 Workbook 293

EXERCISE 5.1

This exercise is designed to be attempted after reading Module 5.1 and the
chapters preceding it.

Discuss the structure of the periphery of the clauses in the sentences

below (in the case of sentences containing a bracketed clause, concerning
yourself only with the periphery of the bracketed clause):
1 I must admit [that a scene as grotesque as that, never before, in all my life,
have I witnessed] [M]
2 Nothing at all, even under duress, did he say [S]
3 I’m sure [that a player like that, no way, under any circumstances, would
they sell to a rival team]
4 That painting, how much, in the end, did you sell *(it) for? [S]
5 How much, in an average year, of your salary do you lose in tax?
6 At no point, for any reason, would, under any circumstances, the men give up
7 What, when you met him, did he say that on no account, at the meeting,
should you do? [S]
8 What did Mary claim that they stole? [S]
9 What did Mary claim did they steal? (Belfast English, Henry 1995: 108)
10 What now do you think you are? ([Link])

Note that the notation *(it) in 4 means ‘the sentence is grammatical if it is

used, but ungrammatical if it is omitted’.

Helpful hints
Assume that all the structures in 1–8 are grammatical – as they are in my
own variety of English (though speakers who treat Auxiliary Inversion as a
root phenomenon, or who find sentences with a complex periphery hard to
parse may find some of the sentences degraded). In each example, concern
yourself only with the structure of the peripheral projections in the
relevant clauses, not with the internal structure of the TP beneath these.
In 2 take nothing at all to be a QP and even under duress to be a PP; in 3
take a player like that to be an ARTP, and no way to be a PP headed by
a null variant of in; in 4 take how much to be a QP (perhaps an elliptical
form of how much money); in 5 take how much of your salary to originate
as a QP which is the complement of lose, and bear in mind the possibility
of split/discontinuous spellout (and the spellout rules in 14 and 15 below);
in 7 take when you met him to be a CONJP headed by the subordinating
conjunction/CONJ when. In all these cases, don’t concern yourself with the
internal structure of the relevant QP/ARTP/PP/CONJP constituents, or with
the internal structure of TP constituents. In relation to 6, note that most
294 More peripheral constituents

speakers find it unacceptable (why do you think this is?), though it is OK

for me: bear in mind the possibility that FINP recursion can give rise to
FINP+MODP+FINP structures. Note that 9 is an example of interrogative
Inversion in Belfast English, and that the complement of claim in this
sentence cannot be introduced by the complementiser that, so might be a
truncated clause which projects only as far as FINP.
Where relevant, consider the possible role played by one or more of the
following:

11 Intervention Condition (= 18)

Likes cannot cross likes

12 Impenetrability Condition (= 22)

Anything below the head H of the highest projection HP in the periphery
of a complete clause is impenetrable to anything above HP

13 Path Containment Condition (= 19)

If two movement paths overlap, one must contain the other

14 Low Spellout Rule (from §2.3.3)

A KP/PP/CP which is the lowest/rightmost constituent of a larger phrase
that undergoes movement can be spelled out on a lower copy of the
moved phrase

15 Default Spellout Rule/DSR (from §1.4.1)

For a constituent whose spellout is not determined in some other way,
the highest copy of the constituent is overtly pronounced, and any lower
copies are silent

EXERCISE 5.2

This exercise is designed to be attempted after reading Module 5.2 and the
module and chapters preceding it.

Discuss the structure of the periphery of the interrogative clauses in the

sentences below (analysing only the bracketed complement clauses in
examples so marked):

1 I wonder [where, the homeless, in times of crisis, they go] [M]

2 He was unaware [what kind of food, on no account, if you are diabetic,
should you eat] [S]
3 Capello has to know [who, when the chips are down, that he can trust]
(Graham Taylor, BBC Radio 5) [S]
4 I wonder [how many other clubs, if their entire first team is on that kind of
money] (Bob Beech, talkSPORT Radio)
 5.7 Workbook 295

5 Why, at no point, despite being ill, did he call a doctor?

6 You never know [life, where it takes you] (Ray Parlour, talkSPORT Radio) [S]
7 He protested [that where else, at the time, could he have gone?]
8 How come, in spite of subsidies, that the bank collapsed? [S]
9 How come Humpty Dumpty, after the fall, they couldn’t mend?
10 How come at no point, even after the scandal, did they investigate him? [S]
11 The alibi he gave, how come at no point, in spite of requests, did anyone
check *(it)?
12 (*I wonder) during the trial, the main defence witness, how come the DA
didn’t cross-examine him?

Say which structures are non-standard, and why: note that speakers who
treat Auxiliary Inversion as a root phenomenon, or who find sentences
with an extremely complex periphery hard to parse may find some of the
sentences degraded. Note too that the use of the complementiser that in
3, 8 is accepted by some speakers (including me) but not by most, that
4 is highly marked (i.e. strikes most speakers – including me – as weird),
and that the word order in the bracketed clause in 6 likewise strikes
many speakers (including me) as very odd. The notation *(it) in 11 means
‘the sentence is grammatical with it, but ungrammatical without it’; the
notation (*I wonder) in 12 means ‘the sentence is ungrammatical with I
wonder, but grammatical without it.’ The phrase ‘when the chips are down’
in 3 means ‘when things are going badly’.

Helpful hints
Concern yourself only with the structure of the peripheral projections in each
relevant clause, not with that of the TPs beneath these. Make the following
assumptions about the categorial status of various peripheral constituents, but
don’t concern yourself with their internal structure: what kind of food, who,
and how many other clubs are QPs; the homeless, Humpty Dumpty, the alibi
he gave, the main defence witness and life are DPs (headed by an overt or
null determiner); why, where, where else, despite being ill, and even after the
scandal are PPs (headed by an overt or null preposition); if you are diabetic
and when the chips are down are CONJPs/subordinate clause projections
headed by the CONJ/subordinating conjunction if/when. In addition, in
relevant sentences, consider the role played by one or more of the following:

13 Interrogative Generalisation (= 34)

(i) A peripheral interrogative XP that triggers Auxiliary Inversion moves
from spec‑FINP (with FIN attracting a finite auxiliary to adjoin to it)
into its criterial position in spec‑FOCP
(ii) The criterial position for a peripheral interrogative XP that does not
trigger Auxiliary Inversion is spec‑FORCEP
296 More peripheral constituents

14 Intervention Condition (= 18)

Likes cannot cross likes

15 Minimality Condition (from §1.3.2)

A moved constituent must move to the minimal/closest potential landing
site above it

16 Selection Condition (= 33)

A selector and its selectee (e.g. a verb and its complement) must be as
close as possible (ideally, immediately adjacent); the further apart they
are, the more degraded the structure becomes

EXERCISE 5.3

This exercise is designed to be attempted after reading Module 5.3 and the
modules and chapters preceding it.

Discuss the structure of the periphery of the interrogative clauses in the

sentences below (in the case of sentences 6–9, concerning yourself only
with the periphery of the bracketed complement clause):

1 During the demonstration, the police, did they arrest anyone? [M]
2 The president, if you need support, can you count on *(him)? [S]
3 At any point, did anyone threaten you? [S]
4 Did at any point did anybody ever point a gun at you? ([Link])
5 Crisp, yummy and delicious potato chips, do you want to eat? (pinterest)
6 He asked [if, the team captain, under no circumstances, at the end of the
season, would they sell to a rival]
7 I wonder [whether, at some stage, that Graham Potter might think the guile
and invention of Adam Lallana might pay dividends today] (Darren Fletcher,
Sky Sports TV) [S]
8 It worries me now [that Terry, does he really want to play for England?]
(Martin Keown, BBC Radio 5)
9 *It’s unclear [any of the suspects if the police arrested] [S]

Say what is unusual/noteworthy about some of the sentences and what

challenges (if any) they pose to the description of yes–no questions given
in Module 5.3: note that 4, 7, 8 are non-standard structures, and for me, 5
feels odd. Note also that speakers who treat Auxiliary Inversion as a root
phenomenon, or who find sentences with a complex periphery hard to
parse may find some of the sentences degraded. The notation *(him) in 2
means that the sentence is grammatical if him is used, but ungrammatical
without him.
 5.7 Workbook 297

Helpful hints
Concern yourself only with the structure of the peripheral projections
in the relevant clause, not with the internal structure of the TP beneath
these. Make the following assumptions about the categorial status of
various peripheral constituents, but don’t concern yourself with their
internal structure: in 2, take if you need support to be a CONJP/projection
of the subordinating CONJ/conjunction if; in 5, take crisp, yummy and
delicious potato chips to be a QP headed by a null quantifier meaning ‘an
unspecified quantity of’; in 6 take the team captain to be a DP; in 8 take
Terry to be a DP headed by a null determiner; and in 9 take any of the
suspects to be a QP. Bear in mind the following assumptions/conditions:

10 Intervention Condition (= 18)

Likes cannot cross likes

11 Path Containment Condition (= 19)

If two [movement] paths overlap, one must contain the other

12 Impenetrability Condition (= 22)

Anything below the head H of the highest projection HP in the periphery
of a complete clause is impenetrable to anything above HP

13 Selection Condition (= 33)

A selector and its selectee (e.g. a verb and its complement) must be as
close as possible (ideally, immediately adjacent); the further apart they
are, the more degraded the structure becomes

14 Auxiliary Inversion (= 57)

A null FIN head which has a focused interrogative/negative XP as its
specifier attracts a finite auxiliary in T to adjoin to FIN

EXERCISE 5.4

This exercise is designed to be attempted after reading Module 5.4 and the
modules and chapters preceding it.

Discuss the structure of the periphery of the exclamative/relative clauses in

the sentences below (in the case of sentences containing a bracketed clause,
concerning yourself only with the periphery of the bracketed clause):

1 We all know [what a job that Luis Suarez has done since he’s been at the club]
(Micky Gray, talkSPORT Radio) [M]
2 Look [how close, on the corners, that Valentino is] (Julian Ryder, BT Sports
TV = ‘Look how close moto GP champion Valentino Rossi is to the edge of the
track when he rides round corners’)
298 More peripheral constituents

3 Aaron Ramsey, again, what a season that he’s having! (Ray Parlour,
talkSPORT Radio)
4 *(To) how many people, nothing at all, at Xmas, did Scrooge give! [S]
5 It’s something [that, off the pitch, that we’ve got to help the players deal with]
(Football executive, BBC Radio 5)
6 They need to be solid at the back against a Spanish team [that what a goal
they scored!] (Ray Parlour, talkSPORT Radio) [S]
7 Joe Hart has come out to look at the wall, [which, Negredo, is he gonna join?]
(Clive Tyldesley, Sky Sports TV)
8 I am growing chillies this year which is something [never have I grown
before] ([Link]) [S]
9 As Liverpool chase the game, there may be more room [in which for
Manchester United to manoeuvre] (Clive Tyldesley, Sky Sports TV)

Say what is unusual/noteworthy about some of the sentences and what

challenges (if any) they pose to the assumptions made in Module 5.4.
Note that the use of that in 1–3 and the second occurrence of that in 5
is generally regarded as non-standard, as is the for-infinitive structure
in 9. Note also that speakers who treat Auxiliary Inversion as a root
phenomenon, or who find sentences with a complex periphery hard
to parse may find relevant sentences degraded. The notation *(to) in 4
means that the sentence is grammatical if to is used, but ungrammatical
without to.

Helpful hints
Concern yourself only with the structure of the peripheral projections in
the relevant exclamative or interrogative clauses, not with the internal
structure of the TP beneath these. Make the following assumptions about
the categorial status of various peripheral constituents, but don’t concern
yourself with their internal structure: the phrases what a job, what a
season, what a goal, and how many people are exclamative QPs, how close
is an exclamative AP, and nothing at all is a negative QP; proper names
like Aaron Ramsey and Negredo are DPs; never is a PP headed by a null
preposition, paraphrasable as ‘at no time’; and likewise again is a PP with
a more abstract structure paraphrasable as ‘on another occasion’. Bear in
mind the assumptions in 10 and 11 below made in Module 5.4, and the
further assumptions/conditions in 12–19:

10 Exclamative clauses are FORCEP constituents which involve an

exclamative XP moving from some position within TP into its criterial
position in spec‑FORCEP, where it serves to type the relevant clause as
exclamative in force
 5.7 Workbook 299

11 Relative clauses are RELP constituents whose specifier is an overt or

null relative XP (e.g. the overt DP which or the null relative DP operator
OpR L ) and whose head is an overt complementiser like that, or a null
complementiser. In resumptive relatives, the relative wh‑XP is generated
in situ in spec‑RELP; in gap relatives, the relative wh‑XP moves from a
position within TP to spec‑RELP, leaving a gap behind. Beneath RELP,
there is a FORCEP which marks the relative clause as for example,
interrogative, exclamative, imperative, or declarative etc. in force.

12 Auxiliary Inversion (= 57)

A null FIN head which has a focused interrogative/negative XP as its
specifier attracts a finite auxiliary in T to adjoin to it

13 Minimality Condition (from §1.3.2)

A moved constituent must move to the minimal/closest potential landing
site above it

14 Intervention Condition (= 18)

Likes cannot cross likes

15 Path Containment Condition (= 19)

If two [movement] paths overlap, one must contain the other

16 Impenetrability Condition (= 22)

Anything below the head H of the highest projection HP in the periphery
of a complete clause is impenetrable to anything above HP

17 Selection Condition (= 33)

A selector and its selectee (e.g. a verb and its complement) must be as
close as possible (ideally, immediately adjacent); the further apart they
are, the more degraded the structure becomes

18 Doubly Filled COMP Filter (revised version of 79)

At the end of a syntactic derivation, a structure is filtered out as
ungrammatical if it contains a clause whose periphery includes an overt
complementiser c‑commanded by some other overt constituent within
the periphery of the same clause

19 Distinctness Condition (from §4.1.3)

At PF, any sequence <α, α> of two adjacent non-distinct constituents is
degraded
6 The subperiphery

6.0 Overview

Under the traditional CP+TP+VP analysis of clause structure, the periphery

of the clause is analysed as a CP projection, and the subperiphery (i.e. that
part of the clause structure positioned between the CP periphery and the VP
core) is analysed as a TP projection. But in this chapter, we’ll look at evi-
dence that the subperiphery has a more articulated/complex structure, and
comprises a number of distinct projections, with each projection housing a
different kind of subperipheral constituent (e.g. an auxiliary, or subject, or
adverbial modifier, or floating quantifier, and so on). The term subperiphery
used here corresponds broadly to the term ‘middle field’ used in Germanic
Linguistics.

6.1 Auxiliary and subject projections

Under the classic CP+TP+VP analysis of clause structure, finite clauses can
contain an auxiliary occupying the head T position of TP, with the subject
of the auxiliary occupying spec‑TP. In this module, however, we’ll look at
evidence suggesting that auxiliaries and their subjects are housed in sep-
arate projections.

6.1.1 Auxiliary projections

Evidence from sentences such as the following suggests that the subperiph-
ery can contain other (italicised) auxiliaries in addition to (a bold-printed)
finite auxiliary in T:

(1) He might have been being held in solitary (confinement)

Here, the clause contains the modal auxiliary might, the perfect auxil-
iary have, the progressive auxiliary been, and the passive voice auxiliary
being. Superficially, the auxiliary might occupies the head T position of TP,
since (like a typical finite T auxiliary), it can undergo Inversion (i.e. T‑to‑C
movement) in sentences like:

(2) Might he have been being held in solitary?

 6.1 Auxiliary and subject projections 301

If we assume that, in the same way as might heads a TP projection of its

own, so too the perfect aspect auxiliary have heads a PERFP projection, the
progressive aspect auxiliary been heads a PROGP projection, and the pas-
sive voice auxiliary being heads a VOICEP projection, the subperiphery of
the sentence in (1) will have the following structure:

(3) TP

DP T′
he
T PERFP
might
PERFP PROGP
have
PROG VOICEP
been
VOICE VP
being held in solitary

(Since the focus here is on the subperiphery, for the sake of expository sim-
plicity, I generally omit any representation of the clause periphery in tree
diagrams in this chapter.)
Some evidence in support of the structure in (3) comes from ellipsis.
An auxiliary typically allows its complement to undergo ellipsis in an ap-
propriate discourse setting. And, just as the analysis in (3) predicts, the T
auxiliary might allows ellipsis of its complement (marked by strikethrough
below) in the utterance produced by the second speaker in (4a) below, so
does the PERF auxiliary have in (4b), the PROG auxiliary been in (4c), and
the VOICE auxiliary being in (4d):

(4) f i rs t s p eaker : Do you think he might have been being held in solitary?
s e c o nd s peaker : a. Yes, he might have been being held in solitary
b. Yes, he might have been being held in solitary
c. Yes, he might have been being held in solitary
d. Yes, he might have been being held in solitary

In accordance with Grice’s (1975) conversational maxim ‘Be concise’, a

more concise response like (4a) is preferred to a more repetitious response
like (4d) – and in addition, the string been being in (4d) feels awkward, per-
haps because it falls foul of an Anti-Iteration Filter which renders a string
of two or more words which are similar in form degraded (and the same
filter makes 1 ungainly as well).
However, an interesting issue which arises in relation to the structure in
(3) concerns the morphosyntax of the auxiliary might. This marks both mo-
dality (more specifically, possibility) and past tense. There are two reasons
for treating might as the past tense of may. Firstly, might ends in the same
302 The subperiphery

past tense suffix -t as irregular past tense verbs like sent/kept/bought/felt/

left/lost/meant etc. And secondly, when the present tense modal may is
used in direct speech (as in 5a below), it is transposed into the past tense
form might in indirect/reported speech after a past tense verb like said (as
in 5b):

(5) a. ‘I may be free tomorrow morning’, Sam said

b. Sam said that she might be free the following morning

We can therefore conclude that might marks both modality and past tense.
This is significant because Cinque & Rizzi (2008) argued for a featural
uniqueness principle which was outlined in Module 4.1 as follows (where
interpretable features are those contributing to meaning):

(6) One Feature One Head Principle

Each functional head carries only one interpretable feature

The principle in (6) entails that since tense and modality are different inter-
pretable features, they must originate on separate heads (tense on a T/tense
head, and modality on an M/modal head). In the case of our sentence (1),
this means that may originates as the head M constituent of an MP/modal
projection, and then (via Head Movement) adjoins to a past tense affix/Af
in T, in the manner shown by the arrow below (where the internal structure
of PERFP is not shown, because our concern here is with the modal):

(7) TP

DP T′
he
T MP
may+Af
M PERFP
may have been being
held in solitary

If the affix in T is present tense, the may+Af complex in T will be spelled

out as may in the PF component; if the affix is past tense, may+Af will be
spelled out as might.
The analysis in (7) has implications for other finite auxiliaries. For ex-
ample, consider the syntax of the auxiliaries italicised below:

(8) a. He has been held in solitary

b. He is being held in solitary
c. He was held in solitary

In (8a), has marks both present tense and perfect aspect. Thus, the One
Feature One Head Principle in (6) would suggest that have originates as
 6.1 Auxiliary and subject projections 303

a PERF head marking perfect aspect, and from there moves to adjoin to a
present-tense affix in T, in the manner shown by the arrow below:

(9) TP

DP T′
he
T PERFP
have+Af
PERF VOICEP
have been held in solitary

If (as here) the affix is present tense and has a third person singular subject,
it will be spelled out at PF as -s, and this affix will be attached to a con-
tracted stem form ha- of the perfect auxiliary h av e to form has. In much
the same way, the italicised progressive aspect auxiliary b e in (8b) will ori-
ginate as a PROG head which raises to attach to a (present) tense affix in
T; and the passive voice auxiliary b e in (8c) will originate as a VOICE head
which raises to attach to a (past) tense affix in T (and thus is spelled out as
was). So, on this view, finite auxiliaries have a complex (split projection)
structure, and originate as the head of a lower auxiliary projection MP/
PERFP/PROGP/VOICEP, and from there raise to adjoin to a (past or present)
tense affix in T. (A potential exception is the auxiliary do; if this marks only
present/past tense, it can be taken to originate in T.)
The split projection analysis of finite auxiliaries outlined above has inter-
esting implications for the syntax of need in negative sentences such as:

(10) a. He need not worry about anything ([Link])

b. Globalisation need not mean deregulation ([Link])
c. A conventional kitchen layout need not be boring (raymunnkitchens.
[Link])
d. He need not have been concerned ([Link])
e. Tax need not be taxing (Her Majesty’s Revenue and Customs)

In these sentences, need is a present tense modal auxiliary expressing neces-

sity, as we see from the fact that (like other modals such as may/can/must/
will), it does not carry the suffix -s when used with a third person singular
subject. But there are two other properties of need which are interesting
here. Firstly, need is a narrow scope modal, in the sense that it falls within
the scope of (i.e. is modified by) the negative adverb not in sentences like
(10): hence (10a) is paraphrasable as ‘It is not necessary for him to worry
about anything’ (and not as ‘It is necessary for him not to worry about any-
thing’). And secondly, need (when used as an auxiliary) is a polarity item
which must originate in a position where it is within the scope of a negative
304 The subperiphery

(or interrogative, or conditional) constituent – hence the ungrammaticality

of a sentence such as *He need worry about everything.
Such considerations suggest that the modal need originates in (10) as the
head M/modal constituent of an MP/modal phrase projection positioned
below the NEGP constituent containing not, and that the modal need raises
from M to adjoin to a present tense affix in T. On this view, (10a) will be
derived as follows. The verb worry first-merges with the PP about anything
and second-merges with the pronominal DP he to form the VP he worry
about anything. This VP is then merged with the M/modal head need to
form the MP/modal projection need he worry about anything. The resulting
MP is merged with a null NEG head which has the ADVP not as its specifier
to form the NEGP below (simplified by not showing the internal structure
of the VP, because this is not at issue here):

(11) NEGP

ADVP NEG′
not
NEG MP
ø
M VP
need he worry about anything

The negative ADVP not here c‑commands need, thereby accounting for not
having scope over need, and for not licensing the use of the polarity item
need (and also licensing the additional polarity item anything).
The NEGP in (11) is then merged with a present tense affix/Af in T. This
affix attracts the modal stem need to adjoin to it; and in addition, the EPP
feature on T attracts he to raise to spec‑TP, so deriving the structure below:

(12) TP

DP T′
he
T NEGP
need+Af
ADVP NEG′
not
NEG MP
ø
M VP
need he worry about anything

Since a present tense affix attached to a modal has a null spellout in English,
the need+Af complex in T will be spelled out as need at PF; and the overall
structure will be spelled out as the sentence 10a He need not worry about
 6.1 Auxiliary and subject projections 305

anything. The important point illustrated here is that we have to posit that
need originates in M and moves (across not) into T in order to account for not
having scope over need at LF in spite of need preceding not at PF. (A point of
detail to note is that I have assumed in 12 that need moves directly to adjoin
to the tense affix in T, and does not transit through NEG because NEG does
not contain any inflectional affix that the auxiliary needs to attach to.)
However, while the assumption in (12) that need originates in an MP
projection below NEGP holds for a narrow scope modal like need (i.e. one
interpreted as falling within the scope of not), it does not hold for a wide
scope modal (i.e. one interpreted as having scope over not) like may in:

(13) He may not have said anything

Here, the modal may has scope over not, as we see from the paraphrase ‘It
may be the case that he did not say anything.’ What this means is that the
modal may originates above not here, so that (13) has the derivation below:

(14) TP

DP T′
he
T MP
may+Af
M NEGP
may
ADVP NEG′
not
NEG PERFP
ø
PERF VP
have he said anything

Since the modal may originates in an M position c‑commanding not in (14),

it follows that may will have scope over not. The modal raises to adjoin to
a tense affix/Af in T, and will be spelled out as may if the tense affix marks
present tense, but as might if the tense affix marks past tense.
Overall, the key assumption underlying the analyses in (7, 9, 12, 14) is that
auxiliaries originate in a dedicated auxiliary projection below T and from
there (if finite) move to adjoin to a tense affix in T.

6.1.2 Subject projection

Under the classic CP+TP+VP analysis of clause structure, the criterial posi-
tion for subjects is spec‑TP: thus, subjects originate internally within VP,
and from there move to their criterial position in spec‑TP. This can be
illustrated by the reply given by speaker b in the dialogue below:
306 The subperiphery

(15) speaker a : What did John tell you?

speaker b : That he does not like garlic

If we adopt the classic CP+TP+VP model of clause structure, the clause

produced by speaker b in (15) above will have the structure shown below
(simplified by not showing the internal structure of NEGP):

(16) CP

C TP
that
DP T′
he
T NEGP
does not like garlic
Under the analysis in (16), the subject ends up in its criterial position as the
specifier of the tense head T (here containing the present tense auxiliary
does). Such an analysis provides a straightforward account of the subject
he immediately preceding does.
However, the spec‑TP analysis of subjects raises the question of how to
deal with sentences like those in (17) below where an italicised subject is
separated from its associated bold-printed auxiliary by some underlined
intervening constituent:

(17) a. Its polymorphic nature almost certainly has allowed the selection of
variants that have helped humankind survive the scourge of malaria
([Link])
b. They very occasionally can spread to either nearby lymph nodes or
internal organs ([Link])
c. My dad on occasions would send me to the co-op with an even
longer list ([Link])
d. The defendant during the trial had taken the stand in his defense
([Link])
e. The accused, when he took the oath of an American officer, was
not given a license to slaughter unarmed men, women and children
([Link])
f. And I am sure that everyone, although tired, had had a most
enjoyable day ([Link])
g. They almost all will suffer gearbox pinion bearing wear/failure at
some point ([Link])
h. The men both were charged and are awaiting trial (Gettysburg Times,
27 July 1988)

A range of different types of (underlined) constituent can intervene between

subject and auxiliary, including an ADVP in (17a, 17b), a PP in (17c, 17d), a
 6.1 Auxiliary and subject projections 307

full or reduced CONJP (i.e. clause headed by a subordinating conjunction/

CONJ) in (17e, 17f), and a floating QP in (17g, 17h). And indeed there can
be multiple constituents intervening between subject and auxiliary, whether
of the same or different types – as illustrated by the examples below:

(18) a. I simply eventually probably just wouldn’t care anymore (forum.

[Link])
b. I’ve learned that HR, historically, probably really has been in sore
need of improvement ([Link]: HR = Human Resources)
c. My guess is that it probably already has peaked ([Link])
d. I frankly really couldn’t care less about the political parties in Britain
([Link])
e. They clearly all can sing very well ([Link])
f. They definitely both should medal ([Link])

So, for example, the underlined intervenors are multiple ADVP constituents
in (18a–18d), and an ADVP followed by a floating QP in (18e, 18f). If the
subject and auxiliary are positioned on the edge of the same TP constitu-
ent (as with he and does in 16 above), how are we to account for constitu-
ents like those underlined in (17, 18) above being able to be positioned in
between a subject and its associated auxiliary?
At first sight, it might seem plausible to treat the underlined constituents
as adjuncts. Since adjuncts of the same kind can be recursively stacked
on top of each other, we could then suppose that the subperiphery of a
sentence like (18a) would have the simplified structure below (where n’t
originates as the head of NEGP, and encliticises onto/attaches to the end
of would):

(19) TP

DP T′
I
ADVP T′
simply
ADVP T′
eventually
ADVP T′
probably
ADVP T′
just
T NEGP
wouldn’t care anymore
Each adverbial would adjoin to a T‑bar constituent to project it into an
even larger T‑bar. The assumption made in (19) that the adverbs are ADVP
constituents (and not just ADV heads) is borne out by the observation that
they can be substituted by larger ADVPs (e.g. simply can be substituted by
308 The subperiphery

quite simply and probably by more likely than not). Analysing the relevant
adverbials as T‑bar adjuncts would enable us to maintain the traditional
assumption that a finite auxiliary and its subject are positioned on the edge
of the same TP projection, with the auxiliary being in head-TP (i.e. occupy-
ing the head T position of TP), and its subject in spec‑TP.
However, the T‑bar adjunct analysis runs into theoretical problems. This
is because it violates a putatively universal constraint on adjunction (dat-
ing back to Baltin 1978/1982 and Chomsky 1986b) which can be charac-
terised informally as follows:

(20) Like-Adjoins-to-Like Constraint

Adjunction can only adjoin like to like

The likeness constraint on adjunction in (20) allows one head X to be

adjoined to another head Y, or one maximal projection XP to be adjoined to
another maximal projection YP, but does not allow an ADVP to be adjoined
to a T‑bar (as in 19), because this involves illicitly adjoining a maximal
projection to an intermediate projection (and the two are not likes, because
they are different types of projection). Of course, the likeness violation is
even clearer where the status of the adjunct as a maximal projection is
self-evident – for example, where the supposed T‑bar adjunct is a PP like
during the trial in (17d), or a CONJP/subordinate clause like when he took
the oath of an American officer in (17e).
But if the T‑bar adjunction analysis of intervening adverbials in (19) is
ruled out by a universal constraint on adjunction, what alternative do we
have? A plausible answer is to reject the traditional assumption that the
auxiliary and its subject are contained within the same TP projection, and
to suppose instead that subject and auxiliary are contained in different
projections – and more specifically, to suppose that a tensed auxiliary is
contained within a TP/tense phrase projection, whereas its subject is con-
tained within a separate SUBJP/subject phrase projection. Let’s suppose
that spec‑SUBJP is the criterial position for a subject, and extend the Cri-
terial Freezing Condition outlined in Module 4.1 to apply not only to con-
stituents in the periphery, but also those in the subperiphery. This we can
do by adding the condition italicised below:

(21) Criterial Freezing Condition/CFC

(i) A peripheral or subperipheral constituent must occupy a criterial
position to be interpretable at LF
(ii) A constituent which occupies its criterial position is frozen in place

On one implementation of this idea, the subperipheral structure of the

clause in (18a) would not be as in (19) above, but rather as shown in
simplified form below:
 6.1 Auxiliary and subject projections 309

(22) SUBJP

DP SUBJ′
I
SUBJ TP
ø
ADVP TP
simply
ADVP TP
eventually
ADVP TP
probably
ADVP TP
just
T NEGP
wouldn’t care
anymore

Under the analysis in (22), the auxiliary wouldn’t is the head of a TP projec-
tion, and the subject I would be the specifier of a separate SUBJP projection
(with a null head). The adverbial constituents would each be analysed as TP
adjuncts which adjoin to a TP to form an even larger TP. Since in each case
adjunction involves one maximal projection adjoining to another (i.e. ADVP
adjoining to TP), the Like-Adjoins-to-Like Constraint in (20) is satisfied by
the analysis in (22). Hence, theoretical considerations relating to the likeness
constraint on adjunction in (20) favour the analysis in (22) over that in (19).
The assumption that subjects are contained in a separate SUBJP projec-
tion raises the question of what the function of SUBJP is. Rizzi (2005: 213)
argues that spec‑SUBJP is the criterial position for subjects, and that it is ‘a
position dedicated to a special interpretive property of the scope-discourse
kind (“aboutness”)’.
However, there is also a further potential rationale for the postulation
of separate tense and agreement projections from morphological consider-
ations. There is a widespread view that (in the same way as T is the locus
of tense), the head which is here labelled as SUBJ is the locus of agreement
with the clause subject; for this reason, in much work in the 1980s we find
the label AGRS (denoting subject agreement) used instead, to mark a head
which agrees in person and number with the clause subject.
One piece of evidence adduced in support of the claim that SUBJ is the
locus of agreement comes from north Italian dialects which use an (itali-
cised) subject clitic to ‘double’ (i.e. copy features of/mark agreement with)
the subject in a finite clause such as the following (from Belletti 2001: 499):

(23) La Maria la parla

The Mary she speaks
‘Mary is speaking’
310 The subperiphery

Subject clitic doubling is found in dialects spoken in Florence and Trento

(Brandi & Cordin 1989), Padua and other areas of northern Italy (Poletto
1993), and also in Franco-Provençal (Roberts 1993b). One way of handling
the phenomenon is to suppose that the subject DP lathe MariaMary occupies
spec‑SUBJP, and the clitic lashe (which agrees in person, number and gen-
der with the subject) occupies the head SUBJ position of SUBJP. Of course,
since English does not have clitic doubling, we can suppose that the head
of SUBJP in the English counterpart of this type of structure contains an
abstract person/number agreement affix which ends up attached to the
finite verb (in much the same way as a finite T carries a tense affix). This
would mean that SUBJ in finite clauses carries an agreement affix, and T a
tense affix. This approach can be justified on universalist grounds, in that
in more highly inflected languages, tense and agreement are marked by
morphologically distinct affixes (e.g. the French verb form chanterai ‘will
sing’ comprises the verb stem chant-, the future tense affix -er, and the first
person singular agreement affix -ai). There’s much more that could be said
about this, but I won’t pursue the issue any further here.
You should now be able to tackle Exercise 6.1.

6.2 Subperipheral adverbs

Although adverbial modifiers have been treated as adjuncts hitherto, in

this module we will see that there is evidence for treating them as specifiers
of dedicated functional heads (with each functional projection serving to
house a particular type of ADVP as its specifier).

6.2.1 Analysing ADVPs as specifiers

As we see from the analysis in (22) above, adverbs positioned between
a subject and auxiliary can be treated as adjuncts to the TP headed by
the auxiliary. However, in sentences such as those below, there is a close
semantic relationship between the italicised adverb and the bold-printed
auxiliary, in that they denote similar concepts:

(24) a. Ageing can be cured – and, in part, it soon will be ([Link])

b. He now is feeling empathy for her (E. McKinney, A Rose for Liliana,
iUniverse, 2008: 278)
c. But there is no reason why this necessarily must be so ([Link])
d. I think he possibly may have a storage unit somewhere too
([Link])
e. Authorities investigating a child abuse report legally may disclose to
parents that a report was made ([Link])
 6.2 Subperipheral adverbs 311

f. Experts say everyone should assume they already have been exposed
to the virus at this point ([Link])
g. I don’t think they yet have accepted the fact that there were only
Democrats voting with me in 1993 (Bill Clinton, Public Papers of the
Presidents of the United States, US Government Publishing Office,
June 13, 1998, p. 964, [Link])

So, for example, in (24a) the auxiliary will (denoting future time) is modi-
fied by the immediately preceding adverb soon (also denoting future time).
In (24b) the auxiliary is (denoting present tense) is modified by the immedi-
ately preceding adverb now (denoting present time). In (24c), the auxiliary
must (denoting necessity) is modified by the immediately preceding adverb
necessarily (also denoting necessity). In (24d), the auxiliary may (denoting
possibility) is modified by the immediately preceding adverb possibly (also
denoting possibility). In (24e), the auxiliary may (here denoting permis-
sion) is modified by the immediately preceding adverb legally (denoting the
kind of permission). In (24f/24g) the auxiliary have (used to denote present
tense and perfect aspect) is modified by an immediately preceding adverb
(already/yet) which is semantically related (because yet means ‘until now’
and already means ‘before now’). The close semantic relationship between
the auxiliary and the adverb modifying it in cases like (24) raises the ques-
tion of whether the two are on the edge of the same projection and in a
specifier-head relationship, and whether the adverb should be analysed as
occupying spec‑TP and functioning as the specifier of the auxiliary.
If we adopt this approach and treat the italicised adverbs in (24) as occu-
pying spec‑TP, the string it soon will be in (24a) would have the subperiph-
eral structure shown in simplified form in (25) below, where the VP headed
by cured undergoes ellipsis, marked by strikethrough; note that for the sake
of expository simplicity, here and elsewhere below, I set aside the claim
made in §6.1.1 that a finite auxiliary originates as the head of an auxiliary
projection below T and from there raises to adjoin to an affix in T:

(25) SUBJP

DP SUBJ′
it
SUBJ TP
ø
ADVP T′
soon
T VOICEP
will be cured
Since the ADVP soon is the specifier of will in (25) we can account for the close
relation between them, since both mark future time, and both are positioned
312 The subperiphery

on the edge of the same TP projection. On this view, each subperipheral adver-
bial is the specifier of a semantically related auxiliary head: for ease of ref-
erence, let’s refer to this as the spec‑AUX analysis of subperipheral adverbs.
However, the spec‑AUX analysis runs into a number of empirical prob-
lems. For one thing, the adverb modifying the auxiliary in sentences like
(24) can equally be positioned after the auxiliary – as illustrated below:
(26) a. It will soon be cured
b. He is now feeling empathy for her
c. But there is no reason why this must necessarily be so
d. I think he may possibly have a storage unit somewhere too
e. Authorities may legally disclose to parents that a report was made
f. Everyone should assume they have already been exposed to the virus
g. I don’t think they have yet accepted that only Democrats voted with
me in 1993

It is clear that the italicised adverb can’t (superficially, at least) be the speci-
fier of the bold-printed auxiliary in sentences like (26), because the adverb
follows the auxiliary (whereas specifiers precede their heads). We might
therefore be tempted to conclude that sentences like (26) undermine the
spec‑AUX analysis of subperipheral adverbs.
And yet, we can continue to maintain the spec‑AUXP analysis if we posit
that the ADVP+AUX order in a structure like (25) is reversed in sentences
like (26) because the auxiliary raises from T to SUBJ. On this (Auxiliary
Raising) view, the subperiphery of the sentence in (26a) It will soon be cured
will have the derivation shown below (where VOICEP is a projection headed
by the passive voice auxiliary be):
(27) SUBJP

DP SUBJ′
it
SUBJ TP
will+ø
ADVP T′
soon
T VOICEP
will
VOICE VP
be
V DP
cured it

Under the analysis in (27), the auxiliary will undergoes Head Movement
from T to SUBJ and thereby comes to move in front of the ADVP soon, so
 6.2 Subperipheral adverbs 313

reversing the underlying ADVP+AUX order. (A complication set aside here

arises if will originates in an AUXP below T, and then moves from AUX to
adjoin to a tense affix in T, and then the resulting will+Af complex adjoins
to the null SUBJ head.)
However, the assumption that a subperipheral adverb serves as the spe-
cifier of a semantically related auxiliary is called into question by the ob-
servation that (for example) an adverb like possibly is not restricted to
modifying a modal auxiliary like may expressing possibility, but rather can
modify a wide range of other auxiliaries with very different semantic prop-
erties, like those bold-printed below:

(28) a. The market possibly is feeling the absence of previous powerful

constructive speculative leadership (The San Francisco Examiner, 25
July 1934)
b. The volcano possibly has been active during the past 10,000 years
([Link])
c. Greg Jennings possibly will return to practice Sunday
([Link])
d. I think it possibly would have been helpful for this to have been a bit
higher ([Link])
e. I possibly shall start off on a trip the end of next week
(Letter from G. C. Marshall to Mrs J. J. Winn, May 24, 1943,
[Link])
f. This group possibly should be screened ([Link])
g. The second transaction possibly must wait until the first transaction
reaches syncpoint ([Link])
h. However, we have a sneaking suspicion that Levein will play as
defensively as he possibly dare ([Link])

Such sentences appear at first sight to undermine any attempt to claim that
the adverb serves as the specifier of a semantically related functional head
that is dedicated to licensing the adverb as its specifier.
However, Cinque (1999) outlines a way of defending the claim that
adverbs function as specifiers of dedicated functional heads. He argues
that in cases like (28), the adverb possibly does not serve as the speci-
fier of the auxiliary, but rather as the specifier of a separate functional
projection above the auxiliary: this functional projection has a null
modal head which itself expresses possibility, so that specifier and head
are semantically related to each other. More generally, Cinque (1999:
77) claims that ‘each adverb class enters into a special spec/head rela-
tion with one particular functional head, and vice-versa’; and Cinque
(2004) refers to this as the ‘functional specifier’ approach. On this view,
(28f) This group possibly should be screened will have the structure
314 The subperiphery

shown in simplified form below (where FMOD denotes a functional head

expressing modality):

(29) SUBJP

DP SUBJ′
this group
SUBJ FP
ø
ADVP F′
possibly
FMOD TP
ø
T VOICEP
should be screened
A structure like (29) enables us to maintain that a subperipheral adverb
functions as the specifier of a dedicated functional head/F. (An incidental
detail set aside in 29 is that under the analysis of finite auxiliaries in §6.1.1,
should will originate as the head of a lower modal projection and from
there raise into the T position that it occupies in 29.)
To illustrate how the analysis works, let’s look at some specific claims
about adverbs made by Cinque (1999: 84–107). Speech act adverbs like
frankly, honestly, sincerely etc. are generated as specifiers of a speech act
mood head. Evaluative adverbs like (un)fortunately, luckily, regrettably,
surprisingly, strangely/oddly (enough), (un)expectedly etc. are generated
in the specifier position of an evaluative mood head. Evidential adverbs
(like allegedly, reportedly, apparently, obviously, clearly, evidently, etc.)
are generated in the specifier position of an evidential mood head. Epis-
temic adverbs (like probably, presumably, supposedly etc.) are generated
as specifiers of an epistemic modal head, and perhaps as the specifier of
an irrealis mood head. Habitual adverbs like usually, habitually, generally,
regularly etc. are generated as specifiers of a habitual aspect head, frequen-
tative adverbs like often as specifiers of a frequentative aspect head, con-
tinuative adverbs like still as the specifiers of a continuative aspect head,
completive adverbs like completely as the specifiers of a completive aspect
head … and so on. In each case, the relevant mood/modal/aspectual heads
can be (and typically are) null in English.
But what is the motivation for treating ADVPs as functional specifiers
rather than adjuncts? There are two main reasons. One is theoretical in
nature: namely, that if we reanalyse adverbial modifiers as specifiers ra-
ther than adjuncts, we can eliminate phrasal adjuncts from our theory of
Universal Grammar, since all phrasal adjuncts will be reanalysed as spe-
cifiers, and this will lead to a more restrictive/tightly constrained theory.
 6.2 Subperipheral adverbs 315

The other reason is empirical in nature (relating to word order), in that the
functional specifier analysis provides a more principled account of word
order – for example, the relative ordering of ADVPs in clauses containing
multiple ADVPs (as we will see in §6.2.2 below), and the variable posi-
tioning of ADVPs with respect to auxiliaries on the one hand (discussed
in §6.3.1) and with respect to subjects on the other (discussed in §6.3.2).

6.2.2 Relative ordering of multiple ADVPs

Our discussion of adverbs in §6.2.1 was simplified by considering sen-
tences which contain only one subperipheral ADVP. However, sentences
can frequently contain two or more ADVPs within the subperiphery, as the
examples in (18) above illustrate. Nevertheless, there are restrictions on the
relative ordering of multiple subperipheral ADVPs in such cases, as we see
from the positions of the italicised ADVPs below:

(30) a. He already completely has annihilated the opposition

b. *He completely already has annihilated the opposition

(31) a. He still almost can’t stop himself

b. *He almost still can’t stop himself

(32) a. He once intentionally had crashed his car

b. *He intentionally once had crashed his car

(33) a. He frankly unfortunately has always been careless with money

b. *He unfortunately frankly has always been careless with money

(34) a. The car sometimes abruptly would come to a halt

b. *The car abruptly sometimes would come to a halt

How can we account for such restrictions on the relative ordering of ADVPs?
If the italicised adverbials were adjuncts (e.g. TP adjuncts as in 22 above),
we would expect that they could be stacked on top of each other in any or-
der, and hence the ungrammaticality of the (b) examples in (30–34) above
would remain unaccounted for. For example, under the adjunct analysis,
(34a) would have the structure shown in simplified form below:

(35) SUBJP

DP SUBJ′
the car
SUBJ TP
ø
ADVP TP
sometimes
ADVP TP
abruptly would come to a halt
316 The subperiphery

Under this analysis, the italicised adverbial modifiers sometimes and

abruptly are each analysed as TP adjuncts (i.e. as constituents which adjoin
to a TP to form a larger TP with the same head). Thus, the ADVP abruptly
adjoins to the TP would come to a halt to form the larger TP abruptly would
come to a halt; and the ADVP sometimes then adjoins to this larger TP to
form the even larger TP sometimes abruptly would come to a halt. However,
a problem which arises in relation to this analysis is that since adjuncts of
the same kind can be stacked on top of the constituents they modify in any
order, we would expect it to be equally possible to adjoin sometimes to the
TP would come to a halt to form the larger VP sometimes would come to a
halt, and then adjoin abruptly to this larger VP to form the even larger VP
abruptly sometimes would come to a halt. However, the resulting sentence
(34b) *‘The car abruptly sometimes would come to a halt’ is ungrammatical,
and the TP-adjunct analysis offers us no account of why this should be.
By contrast, under the functional specifier analysis, the relative ordering
of adverbial modifiers can be handled in terms of the familiar head-com-
plement selection mechanism by which heads impose restrictions on the
types of complement which they do and don’t select/allow. Let’s see how.
Cinque (1999: v) maintains that there is a ‘fixed universal hierarchy of
functional projections’, and that clausal projections housing adverbs occur
in the order specified below, where each italicised adverb functions as the
specifier of a capitalised functional head (with MOD denoting a modality
head, ASP an aspect head, and T a tense head, and some adverbials are
able to occupy two different positions in the hierarchy, denoted by the
subscripts (I) and (II) respectively):

(36) Adverb Hierarchy

[frankly MOODspeech act [fortunately MOODevaluative [allegedly MOODevidential
[probably MODepistemic [once Tpast [then Tfuture [perhaps MOODirrealis
[necessarily MODnecessity [possibly MODpossibility [usually ASPhabitual
[again ASPrepetitive(I) [often ASPfrequentative(I) [intentionally MODvolitional
[quickly ASPcelerative(I) [already Tanterior [no longer ASPterminative
[still ASPcontinuative [always ASPperfect(?) [just ASPretrospective
[soon ASPproximative [briefly ASPdurative
[characteristically(?) ASPgeneric/progressive [almost ASPprospective
[completely ASPcompletive(I) [well VOICE [fast/early ASPcelerative(II)
[again ASPrepetitive(II) [often ASPfrequentative(II) [completely ASPcompletive(II)

What the Adverb Hierarchy specifies is that (e.g.) a functional head denot-
ing speech act mood (and allowing an ADVP like frankly as its specifier) can
select as its complement a functional projection whose head denotes eval-
uative mood (and which allows an ADVP like fortunately as its specifier) …
and so on.
 6.3 Word order variation 317

How this works in more concrete terms can be illustrated for (34a) above,
which will have the structure in (37) below under the functional specifier
analysis of ADVPs, where FFREQ denotes a functional head which licenses a
frequency adverb like sometimes as its specifier, and FMANN denotes a func-
tional head which licenses a manner adverb like abruptly as its specifier:

(37) SUBJP

DP SUBJ′
the car
SUBJ FP2
ø
ADVP F′
sometimes
FFREQ FP1
ø
ADVP F′
abruptly
FMANN TP
ø
T VP
would come to a halt

Adverb order in (34/37) then follows if FFREQ can select a complement headed
by FMANN, but FMANN cannot select a complement headed by FFREQ. I leave you
to verify for yourself that the ordering of adverbs in (30–33) above likewise
conforms to the ordering in the Adverb Hierarchy in (36) above.
The analysis outlined in this section presupposes that ‘ADVPs occupy
the specifier position of distinct functional projections’ (Cinque 1999: 29).
Cinque (1999: 44) notes that an important theoretical advantage of the
functional specifier analysis of adverbials over the more traditional adjunct
analysis outlined in (22) above is that it enables us to dispense with phrasal
adjuncts (i.e. XPs which adjoin to YPs to form even larger YPs), since all
phrasal adjuncts (e.g. sometimes and abruptly in 37 above) are reanalysed
as specifiers.
You should now be able to tackle Exercise 6.2.

6.3 Word order variation

In §6.2.2, we saw that the relative ordering of adverbial modifiers with

respect to each other can be accounted for in a principled fashion under the
functional specifier analysis. In this module, we will see that the analysis
can also account for the variable order which ADVPs show with respect to
other constituents in the clause – including auxiliaries and subjects.
318 The subperiphery

6.3.1 ADVPs and auxiliaries

The variable position of ADVPs with respect to auxiliaries can be illustrated
by the following examples (which are the English counterparts of a set of
Italian examples used to illustrate this variability in Cinque 1999: 49):

(38) a. Gianni then perhaps wisely had decided not to go

b. Gianni then perhaps had wisely decided not to go
c. Gianni then had perhaps wisely decided not to go
d. Gianni had then perhaps wisely decided not to go

If we adopted the traditional adjunct analysis of adverbial modifiers, the

three adverbs then, perhaps and wisely that precede had in (38a) would be
treated as TP adjuncts, adjoined to a TP headed by had. And yet, if they are
all TP adjuncts, we would predict that the auxiliary had cannot be posi-
tioned in front of any of them. But this prediction is false, as we see from
the grammaticality of (38b–38d).
By contrast, under the functional specifier analysis of adverbial modi-
fiers, (38a) will have the subperipheral structure shown in simplified form
below (where F1, F2 and F3 are functional heads which each license a differ-
ent kind of adverbial specifier):

(39) SUBJP

DP SUBJ′
Gianni
SUBJ FP3
ø
ADVP F′
then
F3 FP2
ø
ADVP F′
perhaps
F2 FP1
ø
ADVP F′
wisely
F1 TP
ø had decided
not to go
In (39), had occupies the head T position of TP, and each modifying ADVP
is housed in a separate dedicated functional projection above TP. But how
do we account for the different positions occupied by had in (38b–38d)?
The answer given by Cinque is that the auxiliary had can optionally
raise to adjoin to any of the functional head positions between T and its
subject.
 6.3 Word order variation 319

On this view, (38b) Gianni then perhaps had wisely decided not to go
involves the auxiliary had raising to adjoin to the F2 head between per-
haps and wisely in (39); sentence (38c) Gianni then had perhaps wisely
decided not to go involves the auxiliary had raising to adjoin to the F3 head
between then and perhaps; and sentence (38d) Gianni had then perhaps
wisely decided not to go involves the auxiliary had raising to adjoin to the
SUBJ head between Gianni and then. If we assume that locality constraints
require a moved head to adjoin to the closest accessible head above it each
time it moves, this means that (38d) will involve the four separate Auxiliary
Raising operations arrowed below:

(40) SUBJP

DP SUBJ′
Gianni
SUBJ FP3
had+ø
ADVP F′
then
F3 FP2
had+ø
ADVP F′
perhaps
F2 FP1
had+ø
ADVP F′
wisely
F1 TP
had+ø
T VP
had decided
not to go
On these assumptions, the subperipheral adverbs (and the subject) must be
specifiers of (null) functional heads in order to account for how the auxil-
iary had can come to be positioned between them. The auxiliary had will
remain in situ in T in (38a), but raise up to F2 in (38b), up to F3 in (38c) and
up to SUBJ in (38d).

6.3.2 ADVPs and subjects

In the same way as adverbs can occupy variable positions with respect to
auxiliaries (as illustrated in 38 above), so too they can occupy variable
positions with respect to subjects – as illustrated by Cinque (1999: 109) in
relation to the variable position occupied by the adverb probably in relation
to the subject George in the following sentences:

(41) a. George probably will have read the book

b. Probably George will have read the book
320 The subperiphery

Rather than assume that the adverb probably occupies two different posi-
tions in each of these sentences (on the edge of a functional projection
immediately above TP in 41a, and above SUBJP in 41b), Cinque suggests
that the adverb occupies a fixed position, and that the variant word orders
‘involve the movement of the subject DP … around the ADVP’. He does not
add further detail, but a plausible implementation of this idea is as follows.
Let us suppose that (41a) has the subperipheral structure below (simplified
for expository purposes by setting aside the possibility that will originates as
the head of a lower auxiliary projection and from there raises to adjoin to T):

(42) SUBJP

DP SUBJ′
George
SUBJ FP
ø
ADVP F′
probably
F TP
ø
T PERFP
will have read the book
Given the VP-Internal Subject Hypothesis, the DP George will originate as
the thematic subject of read the book, and from there raise to the specifier
position of SUBJP. Under the Auxiliary Raising analysis in §6.1.1, the aux-
iliary will originates as the head of a lower auxiliary projection (not shown
above), and from there raises to T. The ADVP probably will serve as the spec-
ifier of a functional projection/FP with an abstract epistemic modal head
(see 36 above), and occupy an intermediate position between SUBJP and TP.
Now consider (41b). Let’s suppose that once again the adverb probably
is housed in a functional projection positioned between SUBJP and TP, but
that this time, the subject George (which originates as the specifier of read
the book) raises only as far as spec‑TP, resulting in the structure below:

(43) SUBJP

SUBJ FP
ø
ADVP F′
probably
F TP
ø
DP T′
George
T PERFP
will have read the book
 6.3 Word order variation 321

The structure (43) will then be spelled out at PF as the sentence (43b)
Probably George will have read the book. On this view, the relative order of
subject and adverb is accounted for not in terms of variation of the position
of the adverb, but rather in terms of variation in the position of the subject
(i.e. whether the subject raises out of its initial VP‑internal position to the
edge of TP, or to the edge of SUBJP).
One consequence of this kind of approach is that it assumes that the
subject can occupy variable positions within the clause – for example,
spec‑SUBJP in (42) but spec‑TP in (43). One way of accounting for this
might be to suppose that either T or SUBJ can carry an EPP feature allowing
it to attract a subject to become its specifier. Still, the assumption that the
subject only raises as far as spec‑TP in a structure such as (43) is potentially
problematic in a framework which posits that spec‑SUBJP is the criterial
position for a subject. After all, since George is in spec‑TP in (43), how can
it be interpreted as the clause subject if spec‑SUBJP is the criterial position
for a subject? And if SUBJP does not house the subject in (43), what func-
tion does the SUBJP projection serve? One answer might be to suppose that
the SUBJ head serves the PF function of housing a subject agreement affix
(which lowers onto T in structures containing a finite auxiliary in T), and
SUBJ serves the LF function of attracting the subject to move from spec‑TP
to spec‑SUBJP in the semantic/LF component, thereby enabling George to
be interpreted as the clause subject at LF in (43) because it is in spec‑SUBJP
at LF. Note, however, that this involves positing that movement operations
can apply not only in the syntax but also in the LF component (e.g. raising
a constituent from a non-criterial to a criterial position).

6.3.3 Postmodifying ADVPs

The overall conclusion which emerges from our discussion in this module
so far is that subperipheral ADVPs can be analysed (within a more articu-
lated model of syntax) as specifiers of dedicated functional heads. However,
since specifiers are always the leftmost constituents within the projection
housing them, what this predicts is that subperipheral modifiers will always
be positioned to the left of the constituents that they modify. And yet a
complication which arises from this assumption is that some subperipheral
modifiers can not only precede the constituents they modify, but can alter-
natively follow them – as illustrated by the dual position of the adverb once
below:

(44) a. He probably once had been in love with her

b. He probably had been in love with her once

In (44a), once is a premodifying ADVP in the sense that it precedes the itali-
cised TP that it modifies (had been in love with her). By contrast, in (44b)
322 The subperiphery

once is a postmodifying ADVP, in that it follows the TP that it modifies.

Under the traditional TP adjunct analysis of subperipheral adverbs (illus-
trated in 22 above), the dual position of once can be accounted for straight-
forwardly, since adjuncts can (in principle) be adjoined either to the left or
right of constituents they modify: hence (under the adjunct analysis), once
is adjoined to the left of the TP had been in love with her in (44a), but to the
right of it in (44b). But how can we account for the dual position of once
under the functional specifier analysis?
Sentence (44a) in which once precedes the TP had been in love with
her can be analysed under the functional specifier analysis of ADVPs as
follows:

(45) SUBJP

DP SUBJ′
he
SUBJ FP2
ø
ADVP F′
probably
F2 FP1
ø
ADVP F′
once
F1 TP
ø had been in
love with her
Here, F1 is a functional head which allows a temporal adverb such as once
as its specifier, and F2 is a functional head which allows a modal adverb
like probably as its specifier. As we would expect from their specifier status,
each of the two adverbs precedes the TP that they modify (had been in love
with her).
But now consider how we can account for the clause-final position of
once in (44b) He probably had been in love with her once. Since speci-
fiers are always the leftmost constituents of the projections containing
them, how does once end up to the right of the TP had been in love with
her in (44b)? The answer suggested by Cinque is that adverb-final order
is the result of moving the constituent which is the complement of the
projection housing the adverb to a position where it is the specifier of a
functional projection immediately above that housing the adverb. More
concretely, in the case of a structure like (45) above, the TP had been in
love with her moves to become the specifier of a functional projection
 6.3 Word order variation 323

(denoted as FP2 below) which is immediately above the FP1 projection

containing the adverb once. The relevant movement operation is shown
by the arrow below:

(46) SUBJP

DP SUBJ′
he
SUBJ FP3
ø
ADVP F′
probably
F3 FP2
ø
TP F′
had been
in love F2 FP1
with her ø
ADVP F′
once
F1 TP
ø —

We can refer to this type of movement operation as Complement Raising,

because it involves raising the TP complement of the functional projection
FP1 housing the adverb once to become the specifier of a higher functional
projection FP2 positioned immediately above FP1.
As our discussion here illustrates, the functional specifier analysis of
subperipheral adverbs brings with it the potential complication that adver-
bial postmodifiers involve positing an additional movement operation and
an additional functional projection (to house the moved constituent). Clear-
ly, even more abstract structures and even more movement operations will
be required in sentences involving multiple postmodifying constituents.
Moreover, a further potential complication posed by the Complement
Raising analysis in (46) can be illustrated by a sentence such as the
following:

(47) Which of the two sisters had he probably been in love with once?

Here, the italicised interrogative wh‑phrase which of the two sisters originates
as the complement of the preposition with, and from there moves through
spec‑FINP into spec‑FOCP in the clause periphery via Wh‑Movement/A‑bar
Movement. However, if the TP had been in love with which of the two sisters
moves into the spec‑FP2 position in (46), and Wh‑Movement subsequently
324 The subperiphery

extracts which of the two sisters out this fronted TP, the resulting derivation
will violate the following two constraints (outlined in §1.5.1):

(48) Constraint on Extraction Domains/CED

Only complements allow material to be extracted out of them, not
specifiers or adjuncts (Cattell 1976, Cinque 1978, and Huang 1982)

(49) Freezing Principle/FP

The constituents of a moved phrase are frozen internally within (and so
cannot be extracted out of) the moved phrase (Wexler & Culicover 1980:
119)

The CED violation arises because which of the two sisters is extracted out
of a TP which occupies a specifier position (spec‑FP2) in (46); and the freez-
ing violation arises because which of the two sisters is extracted out of a
TP which has already undergone the movement arrowed in (46). Thus, the
Complement Raising analysis in (46) would seemingly make the incorrect
prediction that sentences such as (47) should be ungrammatical – unless
constraints like (48) and (49) are modified in some way (see Bošković 2020
for discussion of this kind of problem).
To summarise: in Module 6.3, we have looked at the variable pos-
ition of adverbs, auxiliaries and subjects. I noted Cinque’s claim that
this variation can be accounted for by supposing that adverbs occupy
a fixed position within the clause, and that word order variation can be
attributed to movement of other constituents – for example, auxiliaries
raising to a range of functional head positions above T, subjects raising
either to spec‑TP or to spec‑SUBJP, and the complement of a functional
head raising to become the specifier of a superordinate/higher functional
projection.
You should now be able to tackle Exercise 6.3.

6.4 Other subperipheral constituents

Although our discussion in Module 6.2 and Module 6.3 dealt specifically
with subperipheral ADVPs, in this module, we will see that it can be gen-
eralised from these to other types of subperipheral XP which function as
modifiers.

6.4.1 PPs and subordinate clauses

As we see from the sentences below (repeated from 17c–17h above), in
addition to ADVPs, a range of other types of XP (underlined below) can be
 6.4 Other subperipheral constituents 325

positioned in the subperiphery between the italicised subject and the bold-
printed auxiliary:

(50) a. My dad on occasions would send me to the co-op with an even

longer list ([Link])
b. The defendant during the trial had taken the stand in his defense
([Link])
c. The accused, when he took the oath of an American officer, was
not given a license to slaughter unarmed men, women and children
([Link])
d. And I am sure that everyone, although tired, had had a most
enjoyable day ([Link])
e. They almost all will suffer gearbox pinion bearing wear/failure at
some point ([Link])
f. The men both were charged and are awaiting trial (Gettysburg Times,
27 July 1988)

In (50a, 50b), the underlined subperipheral constituent is a PP, in (50c, 50d)

it is a CONJP (i.e. clause headed by the subordinate conjuction/CONJ when/
although), and in (50e, 50f) it is a floating QP. A range of different constitu-
ents can co-occur within the subperiphery of a single clause, as we see from
the highlighted constituents below:

(51) They probably, while at uni, at some point had done drugs

Thus in (51) we find the adverb probably, the reduced CONJP/subordinate

clause while at uni, and the PP at some point all positioned between the
subject they and the auxiliary had. There are parallels between PPs and
ADVPs which suggest that the two should be treated in the same way (i.e.
as specifiers of dedicated functional heads): for example, the ADVP prob-
ably in (51) can be substituted by the PP in all probability, and conversely
the PP on occasions in (50a) can be substituted by the ADV occasionally;
and many of the items treated as ADVPs in our earlier discussion could be
given a more abstract analysis as PPs headed by a null preposition (e.g.
then in 40 is paraphrasable as ‘at that time’). Moreover, CONJP constituents
(i.e. clauses introduced by a subordinating conjunction/CONJ) are similar
in many ways to PPs, and indeed prepositions like before/after/since/until
can be used to introduce subordinate clauses. If we treat all subperipheral
ADVPs, PPs and CONJPs in the same way (as specifiers of dedicated func-
tional heads), a sentence like (51) above will have the subperipheral struc-
ture below (if while at uni is a projection of the subordinating conjunction/
CONJ while and is perhaps a reduced form of while they were at uni):
326 The subperiphery

(52) SUBJP

DP SUBJ′
they
SUBJ FP3
ø
ADVP F′
probably
F3 FP2
ø
CONJP F′
while
at uni F2 FP1
ø
PP F′
at some
point F1 TP
ø
T VP
had done drugs
Such an analysis enables us to generalise Cinque’s functional specifier
analysis of adverbs as follows:

(53) Functional Specifier Hypothesis/FSH

Each XP in the periphery or subperiphery of a clause is the specifier of a
dedicated functional head

In conformity with FSH, each XP/maximal projection in the subperiph-

ery in (52) is housed in a separate projection: DP they is in spec‑SUBJP,
ADVP probably is in spec‑FP3, CONJP while at uni is in spec‑FP2, and PP at
some point is in spec‑FP1. Under FSH, XPs previously analysed as adjuncts
are reanalysed as specifiers (so eliminating operations which adjoin one
maximal projection to another). We might further suppose that each type
of (sub)peripheral XP occupies a fixed position within the (sub)periphery,
determined by the kind of selectional requirements embodied in a template
like the Adverb Hierarchy in (36) above.
FSH (53) can arguably be subsumed within a broader generalisation
along the following lines:

(54) Merge Hypothesis

A maximal projection/XP can only be introduced into a structure
containing it via (first- or second-) merge with a head

It follows from (54) that an XP can either be a complement (first-merged

with a head) or a specifier (second-merged with a head), but cannot be an
adjunct. This in turn would mean that the only kind of adjunction operation
sanctioned by linguistic theory is adjunction of one head to another: this
 6.4 Other subperipheral constituents 327

type of head-to-head adjunction appears to be independently required in

order to handle cliticisation (e.g. of n’t to could in couldn’t).

6.4.2 Floating QPs: Stranded under movement

As illustrated in (50e, 50f) above, the subperiphery can also contain float-
ing QPs. Interestingly, these can occupy a range of different positions in the
subperiphery – as illustrated below:

(55) a. The players should have both been fined

b. The players should both have been fined
c. The players both should have been fined

One way of accounting for the multiple positions they occupy is to sup-
pose that floating quantifiers can be stranded in intermediate positions
through which subjects transit: for succinctness, let’s call this the stranding
analysis. We can illustrate how it works in terms of the sentences in (55).
Let’s suppose that both originates as the head Q/quantifier of the QP both
the players, and that this QP originates as the complement of the passive
participle fined. Let’s further suppose that in (55a), the QP both the players
moves to become the specifier of the passive voice auxiliary been, and then
the DP the players subsequently moves on its own to spec‑SUBJP, leaving
both stranded as the specifier of the VOICEP headed by been. Likewise, let’s
suppose that in (55b), the QP both the players moves to become the specifier
of the perfect aspect auxiliary have, and then the DP the players moves on
its own to spec‑SUBJP, leaving both stranded as the specifier of the PERFP
constituent headed by have. And finally, let’s suppose that in (55c), the QP
both the players moves to become the specifier of the auxiliary should in T,
and then the DP the players moves on its own to spec‑SUBJP, leaving both
stranded as the specifier of the TP headed by should.
These assumptions would mean that the sentences in (55a/55b/55c)
above would have the respective structures shown in highly simplified form
in (56a/56b/56c) below (with the structures simplified by showing only
maximal projections, and only overt items):

(56) a. [SUBJP the players [TP should [PERF have [VOICEP both been [VP fined]]]]]
b. [SUBJP the players [TP should [PERF both have [VOICEP been [VP fined]]]]]
c. [SUBJP the players [TP both should [PERF have [VOICEP been [VP fined]]]]]

On this view, the floating quantifiers are stranded under movement of the
subject through the edge of the various auxiliary projections, and end up as
specifiers of the relevant auxiliaries. (Note that, for the sake of expository
simplicity, I set aside the possibility that should originates as the head of a
modal projection below TP and from there moves to adjoin to a past tense
affix in T.)
328 The subperiphery

Although the stranding analysis of floating quantifiers seems attractive

at first sight, closer reflection suggests that it is potentially problematic in
certain respects. For example, on the classic variant of the stranding analy
sis, ‘leftward movement undergone by the subject over the quantifier pro-
ceeds through the specifier of QP’ Shlonsky (1991: 159). This would mean
that the derivation of (55c) The players both should have been fined would
include the movement operations arrowed in the structure shown below
(simplified by not showing the internal structure of T‑bar):

(57) SUBJP

DP SUBJ′
the players
SUBJ TP
ø
QP T′
should have been fined
DP Q′
the players
Q DP
both the players

On this view, the DP the players moves from comp‑QP (i.e. from being
the complement of Q‑both) to spec‑QP, as shown by the lower/dotted
arrow in (57); and subsequently the DP the players moves from spec‑QP to
spec‑SUBJP, as shown by the higher/dashed arrow.
Unfortunately, however, the movements arrowed in (57) violate a num-
ber of constraints on movement. For example, movement of the DP the
players from comp‑QP to spec‑QP violates an antilocality constraint which
Boeckx (2007: 110) formulates as follows:

(58) Antilocality Constraint

‘Movement internal to a projection counts as too local, and is banned’

The antilocality violation arises because the DP moves from complement

to specifier position internally within the same QP projection. In addition,
the second movement of the DP the players from spec‑QP to spec‑SUBJP
violates the Constraint on Extraction Domains/CED (48), because CED bars
extraction out of a specifier, and the QP out of which the players is extracted
in (57) occupies spec‑TP. Indeed, given the assumption that the relevant QP
originates in spec‑VP and from there moves to spec‑TP, the DP movement
operation arrowed in (57) will also violate the Freezing Principle/FP (49),
because the DP the players is extracted out of a QP both the players that has
moved from spec‑VP to spec‑TP.
 6.4 Other subperipheral constituents 329

Our discussion here leads to the conclusion that the particular implemen-
tation of the quantifier stranding analysis outlined in (57) runs into the-
oretical problems. However, these can potentially be overcome if we adopt
an alternative implementation of the stranding analysis which makes use
of the split spellout mechanism outlined in §2.3.3. On this alternative ap-
proach to stranding, all the variants of the sentences in (55) would involve
fronting the whole QP both the players, and the differences between them
would lie in whether the quantifier all is spelled out on the highest copy
of the movement chain, or on some lower copy in the chain. On this split
spellout account, (55c) would have the syntactic derivation below (simpli-
fied for expository purposes by ignoring the possibility that should origin-
ates as the head of a modal projection and from there moves to adjoin to
a tense affix in T):

(59) SUBJP
QP5 SUBJ′
both the
players SUBJ TP
ø T′
QP4
both the
players T PERFP
should QP3 PERF′
both the
players PERF VOICEP
have QP VOICE′
2
both the
players VOICE VP
been
V QP1
fined both the
players

As indicated by the arrows in (59), the QP both the players originates as

the complement of the passive participle fined, and then moves in suc-
cessive-cyclic fashion through spec‑VOICEP, spec‑PERFP, and spec‑TP
into spec‑SUBJP. In the PF component, only the highest copy of the DP
the players (inside QP5) can be spelled out overtly, in accordance with the
Default Spellout Rule which was given the following formulation in §1.4.1:

(60) Default Spellout Rule/DSR

For a constituent whose spellout is not determined in some other way,
the highest copy of the constituent is overtly pronounced, and any lower
copies are silent
330 The subperiphery

But let’s suppose that floatable quantifiers like all/both/each are sub-
ject to a special spellout rule of their own, which can be formulated as
follows:

(61) Floating Quantifier Spellout Rule/FQSR

Floatable quantifiers can be spelled out on any non-initial link of an
A‑Movement chain in standard varieties of English

The qualification ‘non-initial’ is added to (61) in order to account for the

ungrammaticality of sentences like that below:

(62) *The players should have been fined both

In (62), the quantifier both is illicitly spelled out in its initial position (as
the complement of fined), and this accounts for the ungrammaticality of the
resulting sentence. A point to note in passing is that there is considerable
variation in non-standard varieties of English with respect to the positions
in which floating quantifiers can be spelled out: see Henry (2012) for an
interesting discussion.
Under the spellout account in (61), we can make the following assump-
tions about where in the tree (59) Q‑both is spelled out. In (55a) ‘The play-
ers should have both been fined’, Q‑both is spelled out in spec‑VOICEP;
in (55b) ‘The players should both have been fined’, Q‑both is spelled out
in spec‑PERFP; and in (55c) ‘The players both should have been fined’,
Q‑both is spelled out in spec‑TP. And in (63) below, Q‑both is spelled out
in spec‑SUBJP:

(63) Both the players should have been fined

So, the split spellout account seemingly provides us with a mechanism for
maintaining the stranding analysis of floating quantifiers.
However, although the split spellout variant of the stranding analysis
accounts for the (un)grammaticality of quantifier stranding in sentences
like those in (55) and (62) above, it proves problematic for sentences like
those below:

(64) a. They have both been vaccinated

b. Peter, Paul and Mary have all been vaccinated
c. The patients have all three been vaccinated

The reason is that the putative source QPs under the stranding analysis,
(highlighted below) are ungrammatical:

(65) a. *Both they have been vaccinated

b. *All Peter, Paul and Mary have been vaccinated
c. *All three the patients have been vaccinated
 6.4 Other subperipheral constituents 331

Nor can the source QP be both of them/all of Peter, Paul and Mary/all three
of the patients, since this would raise questions about where of has gone,
and why in (64a) the subject is they and not them.
A further problem for the stranding account arises in clauses with a plu-
ral subject where the stranded quantifier is each, as in:

(66) We have each been goading the other

Any suggestion that the underlying subject of goading the other is the QP
*each we poses the problem that this is ungrammatical; and to assume that
the underlying subject is each of us raises not only the question of where
of has gone, but also how the auxiliary have comes to be first person plural
in (66), when a phrase like each of us requires third person singular agree-
ment, as in (67) below:

(67) Each of us has been goading the other

Moreover, the further question arises of how the subject comes to be nom-
inative we in (66) when its supposed source in (67) contains accusative us.
An additional potential problem for the stranding analysis is posed by
sentences where the (bold-printed) stranded quantifier has an (italicised)
complement of its own, as in:

(68) a. But they were, all of them, deceived, for another Ring was made (J.
R. R. Tolkien, The Fellowship of the Ring, George Allen & Unwin,
London, 1954)
b. They have, each of them been allowed a bottle of beer a day
(George Washington to William Pearce, 23 December 1793,
[Link])
c. They would neither of them have opposed that (J. Shattock & E.
Jay, The Selected Works of Margaret Oliphant, Part VI, London,
Routledge, 2021: 138)
d. They should none of them have the tag in the query ([Link])
e. They didn’t either of them appreciate what had happened ([Link])
f. They have several of them been widely reported both in the East and
the West ([Link])

In cases like (68), it is hard to see how the quantifiers all/each/neither/none/

either/several could be stranded by movement of the subject they, since the
floating quantifier has a complement of its own – namely of them.
Nonetheless, care needs to be taken in evaluating sentences like (68), since
they may not be instances of floating quantifiers, but rather may represent
a different kind of structure in which a highlighted QP is right-dislocated.
Some evidence in support of this view is that (like right-dislocated con-
stituents but unlike typical floating quantifiers) the highlighted QPs in (68)
332 The subperiphery

can also occur at the end of the sentence. Moreover, the structure in (68)
can contain quantifiers like either/neither/several/none which are otherwise
unable to float, and the highlighted constituents seem to form a separate
intonation group of their own (which can be marked by the use of a comma
preceding and following the relevant constituent, as in 68a): this marks
them out as different from typical floating quantifiers. If so, we should
probably not treat sentences like (68) as providing us with evidence about
the syntax of floating quantifiers.
Yet another potential problem which arises with the stranding analysis is
the following. On the stranding account, floating QPs always occupy the spe-
cifier position of an auxiliary projection which the subject transits through.
However, this assumption proves problematic for sentences like the following:

(69) They allegedly both rarely do heed advice

It is difficult to argue that both is stranded as the specifier of an auxil-

iary in (69), since there is no auxiliary between the adverbs allegedly and
frequently.
Overall, then, the considerations outlined in relation to (55–69) above
call into question the twin ideas that (i) a floating quantifier is stranded
under movement, and that (ii) superficially a floating quantifier occupies
the specifier position of an auxiliary projection through which the subject
transits. Consequently, in the next section, I explore the alternative possi-
bility that a floating quantifier is directly generated in situ as the specifier
of a dedicated functional projection which can occupy a variety of posi-
tions within the subperiphery.

6.4.3 Floating QPs: In a dedicated projection

A traditional assumption made in work in the 1980s was that floating
quantifiers occupy much the same range of positions as adverbs – as we see
from the potential parallel between the positions occupied by the floating
quantifier all and the adverb quietly below:

(70) a. The students all entered the room

b. The students quietly entered the room

Moreover, as observed by Sag (1978), floating quantifiers pattern like

adverbs (and unlike not/n’t) in respect of not permitting VP ellipsis in struc-
tures like:

(71) a. Otto has read this book, and my brothers have (all/certainly) read it, too
b. Otto has read this book, and my brothers have (*all/*certainly) read it,
too
c. Otto has read this book, but my brothers have (n’t/not) read it
 6.4 Other subperipheral constituents 333

This led to proposals that floating quantifiers should be treated in the same
way as adverbs: on one implementation of this idea, both all and quietly
would serve as adjuncts to the VP entered the room in (70). This is far from
implausible at first sight, since a sentence like (72a) below containing the
quantifier most can be closely paraphrased by a sentence like (72b) con-
taining the adverb mostly:

(72) a. Most students hate taking exams

b. Students mostly hate taking exams

However, closer reflection suggests that floating quantifiers aren’t adver-

bial in nature. For example, unlike mostly in (72b), floating quantifiers such
as both/all/each don’t allow the adverbial affix -ly to be attached to them;
and conversely, mostly in (72b) has no floating quantifier counterpart (cf.
*Students most hate taking exams). Moreover (unlike adverbs) floating
quantifiers have inherent number properties, and can only quantify sub-
jects with matching number properties – as we see from:

(73) a. The two men were each/both/*all feeling tired

b. The three men were each/all/*both feeling tired

As these examples show, the quantifier each requires a subject denoting

more than one entity, both requires a subject denoting two entities, and all
requires a subject denoting more than two entities. By contrast, when used
as a genuine adverb (paraphrasable as ‘entirely’), the word all has no num-
ber properties and hence can be used with a singular subject, as in:

(74) The dog was all wet

The contrast between (73) and (74) thus casts doubt on the assumption
that floating quantifiers are adverbs. In addition, floating quantifiers and
adverbs differ in their distribution, as illustrated below:

(75) a. The students entered the room quietly

b. *The students entered the room all

Thus, the adverb quietly in (75a) above can follow the VP entered the room,
but the floating quantifier all in (75b) cannot.
The considerations outlined above lead to the conclusion that while
floating quantifiers resemble adverbs in some respects, they differ from
them in others. This raises the possibility of treating them in much the same
way as Cinque treats adverbial modifiers (namely as specifiers of dedicat-
ed functional heads), but with the difference that the relevant dedicated
functional head licenses a QP rather than an ADVP as its specifier. On this
functional specifier analysis, the subperiphery of a sentence like (69) ‘They
allegedly both rarely do heed advice’ would have the structure shown in
334 The subperiphery

(76) below (where F1/F2/F3 denote dedicated functional heads which each
license a different kind of specifier):

(76) SUBJP

DP SUBJ′
they
SUBJ FP3
ø
ADVP F′
allegedly
F3 FP2
ø
QP F′
both
F2 FP1
ø
ADVP F′
rarely
F1 TP
ø
T VP
do heed advice

An incidental point of detail to note in relation to (76) is that I have made

the simplifying assumption that the auxiliary do marks present tense but
not aspect/modality, and hence can be taken to originate in T; but if we
take do to mark habitual aspect as well as tense, then do will originate as
the head of a habitual aspect projection/HABP below TP, and from there
raise to adjoin to a present tense affix in T.
As for the range of positions which floating quantifiers can or can’t oc-
cupy in sentences like (55/62), we can handle this by supposing that a func-
tional projection housing a floating quantifier can in principle occupy any
position in the subperiphery below the subject (i.e. any position below the
edge of SUBJP but above the edge of VP). Moreover, we can handle struc-
tures in which a floating quantifier is positioned between two adverbs, as
we see from (76) above. Furthermore, if we make the plausible assumption
that a floating quantifier is construed as quantifying a subject with specific
number properties, we can account for each requiring a subject denoting
more than one entity, both requiring a dual subject (i.e. one denoting two
entities), and all requiring a subject denoting three or more entities. I shall
therefore assume henceforth that floating quantifiers are QPs which serve
as specifiers of dedicated functional heads.
Although (as noted above) floating quantifiers behave like adverbs in cer-
tain respects, they also differ from them in other respects. One significant
difference is that (as illustrated below) an adverb like sometimes can typically
 6.4 Other subperipheral constituents 335

either precede or follow a constituent that it modifies, whereas a floating

quantifier like both cannot function as a postmodifier – as illustrated below:

(77) a. They sometimes both would order lobster

b. They both would order lobster sometimes
c. *They sometimes would order lobster both

In (77a), the ADVP sometimes and the floating QP both serve as premodi-
fiers of the TP would order lobster. But while sometimes can also serve as
a postmodifier in (77b), the floating QP both cannot – as we see from the
ungrammaticality of (77c). Why should this be?
Under the functional specifier analysis of subperipheral modifiers out-
lined here, (77a) will have the structure shown in simplified form below:

(78) SUBJP

DP SUBJ′
they
SUBJ FP2
ø
ADVP F′
sometimes
F2 FP1
ø
QP F′
both
F1 TP
ø would order lobster

Given this assumption, (77b) will be derived by moving the FP1 constituent
both would order lobster to become the specifier of an additional functional
projection above the FP2 constituent housing sometimes – as shown below:

(79) SUBJP

DP SUBJ′
they
SUBJ FP3
ø
FP1 F′
both would
order lobster F3 FP2
ø
ADVP F′
sometimes
F2 FP1
ø —
336 The subperiphery

Although (as we see from 78 above) the ADVP sometimes originates as a

premodifier to FP1 both would order lobster, it comes to be positioned after
FP1 and stranded at the end of the sentence as a result of the Complement
Raising operation arrowed in (79).
By contrast, the quantifier both cannot be stranded at the end of the sen-
tence in the same way, as we see from the ungrammaticality of (77c). What
this means in terms of the functional specifier analysis is that a floating
quantifier doesn’t allow the Complement Raising operation arrowed below:

(80) SUBJP

DP SUBJ′
they
SUBJ FP3
ø
ADVP F′
sometimes
F3 FP2
ø
TP F′
would
order F2 FP1
lobster ø
ADVP F′
both
F1 TP
ø —

This means that an additional complication posed by the functional spe-

cifier analysis of modifiers is that some premodifiers allow a Complement
Raising operation which in effect converts them into postmodifiers, whereas
others do not. We can account for this by positing that the kind of func-
tional head that allows a frequency adverbial like sometimes as its specifier
can trigger raising of its complement, whereas the kind of functional head
which allows a floating QP as its specifier cannot. Still, this leaves open the
question of why this should be so.
To sum up: I have argued in Module 6.4 that, like adverbial modifiers,
floating quantifiers also function as specifiers of dedicated functional heads,
and I have shown that such an analysis accounts for how a floating quanti-
fier can end up positioned between two adverbs. I noted, however, that the
type of functional head which licenses a floating quantifier as its specifier
does not trigger the Complement Raising operation by which functional
heads which license an adverb as their specifier can trigger movement of
their complement to become the specifier of a higher functional projection
 6.5 Summary 337

(contrast 79 and 80 above). The consequence of this is that floating quanti-

fiers in English can only function as premodifiers, whereas adverbs can be
either premodifiers or postmodifiers.
You should now be able to tackle Exercise 6.4.

6.5 Summary

This module has been concerned with the structure of the subperiphery of
clauses – that is, that part of the structure positioned between the CP per-
iphery and the VP core. In the classic CP+TP+VP model of clause structure,
the subperiphery was analysed as a unitary TP constituent; however, in this
module we have looked at evidence for the subperiphery being split into a
number of separate projections, resulting in it having a more articulated/
complex structure than traditionally assumed.
In Module 6.1, I noted that the subperiphery can contain multiple auxil-
iary projections, including not just a TP/tense projection, but also an MP/
modal projection, PERFP/perfect aspect projection, PROGP/progressive as-
pect projection, and VOICEP/passive voice projection, in each case headed
by a corresponding type of auxiliary. I argued that finite auxiliaries typi-
cally originate as the head of an auxilary projection (MP/PERFP/PROGP/
VOICEP) below TP, and from there raise to adjoin to a (past or present)
tense affix in T – though I noted that the auxiliary do (if we take it to mark
tense alone) can be taken to be directly generated in T. I went on to argue
that the traditional assumption that subjects are contained in the same
TP projection as finite auxiliaries proves problematic for sentences which
contain constituents intervening between subject and auxiliary, like the
intervening ADVP most definitely in ‘He most definitely did apologise.’ I
noted that earlier work had treated such intervening adverbials as T‑bar
adjuncts, but that such an analysis would violate a likeness constraint on
adjunction proposed by Baltin (1978) to the effect that a maximal projec-
tion like most definitely can only adjoin to another maximal projection. I
argued that this problem can be overcome if the auxiliary did is treated as
heading the TP did apologise, and if the subject he is analysed as contained
in a separate SUBJP projection; under the traditional analysis of adverbial
modifiers as adjuncts, the ADVP most definitely can then be analysed as
a maximal projection which adjoins to another maximal projection (i.e. to
the TP did apologise).
However, in Module 6.2, I outlined Cinque’s claim that adverbial modi-
fiers should be reanalysed as specifiers of functional projections, each with
a dedicated functional head that licenses a specific type of adverb as its
specifier. I reported Cinque’s claim that such a functional specifier analysis
338 The subperiphery

is preferable to the traditional adjunct analysis, in part because it enables

us to dispense with phrasal adjuncts altogether, and in part because it en-
ables us to account for the relative ordering of multiple subperipheral ad-
verbials in terms of the familiar head-complement selection mechanism. I
also noted Cinque’s hypothesis that adverbs of a given type occupy a fixed
position within the clause, and that the relative positions of different types
of adverb can be captured by the putatively universal Adverb Hierarchy
in (36).
In Module 6.3, I looked at the variable position of adverbs, auxiliaries
and subjects. I noted Cinque’s claim that this variation can be accounted
for by supposing that adverbs occupy a fixed position within the clause,
and that word order variation can be attributed to movement of other con-
stituents – for example, auxiliaries raising to a range of functional head
positions above T, subjects raising to either spec‑TP or spec‑SUBJP, and
the complement of a functional head raising to become the specifier of a
superordinate/higher functional projection.
In Module 6.4, I argued that other subperipheral constituents (includ-
ing PPs, subordinate clauses/CONJPs, and floating QPs) can be analysed
(within a more articulated model of syntactic structure) as specifiers of
dedicated functional heads. But I noted that floating quantifiers differ from
other subperipheral constituents in that they can occupy a wide range of
subperipheral positions between the SUBJ head and the VP. I outlined the
stranding analysis of floating quantifiers whereby they are stranded un-
der movement of the subject through intermediate auxiliary projections,
and I presented a split spellout account of this. However, I argued that the
stranding analysis of floating quantifiers proves problematic in relation
to sentences where a floating quantifier is stranded between two adverbs,
and I outlined an alternative account under which floating quantifiers are
housed within a dedicated functional projection that can occupy a range
of positions within the subperiphery (between SUBJP and VP). I noted that
floating quantifiers differ from adverbs in that they do not allow Comple-
ment Raising, and thus cannot function as postmodifiers.
Key constructs in our discussion this chapter include the following:

(6) One Feature One Head Principle

Each functional head carries only one interpretable feature

(20) Like-Adjoins-to-Like Constraint

Adjunction can only adjoin like to like

(36) Adverb Hierarchy

[frankly MOODspeech act [fortunately MOODevaluative [allegedly MOODevidential
[probably MODepistemic [once Tpast [then Tfuture [perhaps MOODirrealis
 6.6 Bibliographical notes 339

[necessarily MODnecessity [possibly MODpossibility [usually ASPhabitual

[again ASPrepetitive(I) [often ASPfrequentative(I) [intentionally MODvolitional
[quickly ASPcelerative(I) [already Tanterior [no longer ASPterminative
[still ASPcontinuative [always ASPperfect(?) [just ASPretrospective [soon ASPproximative
[briefly ASPdurative [characteristically(?) ASPgeneric/progressive [almost ASPprospective
[completely ASPcompletive(I) [well VOICE [fast/early ASPcelerative(II)
[again ASPrepetitive(II) [often ASPfrequentative(II) [completely ASPcompletive(II)

(53) Functional Specifier Hypothesis/FSH

Each XP in the periphery or subperiphery of a clause is the specifier of a
dedicated functional head

(54) Merge Hypothesis

A maximal projection/XP can only be introduced into a structure
containing it via (first- or second-) merge with a head

(61) Floating Quantifier Spellout Rule/FQSR

Floatable quantifiers can be spelled out on any non-initial link of an
A‑Movement chain in standard varieties of English

6.6 Bibliographical notes

For more detailed textbook accounts of the claim made here that auxil-
iaries originate as the head of an auxiliary projection below T and from
there (if finite) move to adjoin to a tense affix in T, see the discussion of
Auxiliary Raising in Radford (2020: 236–60), or Radford (2004a, 2004b,
2009a, 2009b, 2016).
The split TP analysis (splitting TP into two separate projections, one
which is the locus of tense, and another which is the locus of subjects/
subject agreement) was known in earlier work as the split INFL hypothesis
(in an era when INFL was treated as the locus of tense and agreement fea-
tures in finite clauses). It dates back to work by Pollock (1989), Chomsky
(1989, 1993), Belletti (1990, 2001), Speas (1994), Watanabe (1994), Giorgi &
Pianesi (1997), Aygen-Tosun (1998), and Griffin (2003). It has been claimed
that there is evidence supporting the hypothesis from language acquisition
(Malamud-Makowski 1994; Ingham 1998; Hanson 1999; Guasti & Rizzi
2002; Deen 2002), and language disorders (Schaeffer 1994; Friedemann
1998; Friedmann & Grodzinsky 1997, 2000; Ferreiro 2003). For discussion
of SUBJP, see Cardinaletti (1997, 2004), Rizzi & Shlonsky (2006, 2007),
Haegeman (2002a, 2002b, 2007b, 2013, 2017, 2019, 2021), Haegeman &
Danckaert (2017), and Haegeman & Starke (2021).
340 The subperiphery

On the treatment of adverbs as specifiers of dedicated functional pro-

jections, see Cinque (1999, 2004, 2017a, 2023), and Rizzi & Cinque (2016):
for a critical perspective, see Bobaljik (1999), Nilsen (2003), Meira (2004),
Manninen (2005), Holmer (2013), Smit (2014), and Forsmo (2020). On sim-
ilar approaches to Cinque’s, see Laenzlinger (1993, 1996, 1999, 2000),
and Alexiadou (1994, 1997). For a defence of the alternative (traditional)
adjunct analysis of adverbs, see Costa (2000, 2004), Haider (2000, 2004),
Rosengren (2000), Maienborn (2001), Ernst (2002, 2004), Williams (2002),
and Svenonius (2002b).
On floating quantifiers, see Maling (1976), Klein (1977), Sag (1978),
Haig (1980), Vater (1980), Jaeggli (1982), Nakamura (1983), Dowty & Bro-
die (1984), Sportiche (1986, 1988), Giusti (1990, 1991b), Junker (1990),
Yatabe (1990), Fukushima (1991a, 1991b), Shlonsky (1991), Doetjes (1992,
1997), Hasegawa (1993), Akiyama (1994), Baltin (1995), Hoeksema (1996),
Merchant (1996), Tonoike (1996), Torrego (1996), Hamamo (1997), Pau-
novic (1997), Brisson (1998, 2000), Abeillé & Godard (1999), Benmamoun
(1999), Fox & Nissenbaum (1999), Haegeman & Guéron (1999), Ishii (1999),
De Cat (2000), McCloskey (2000), Bošković (2001, 2004), Yoo (2001, 2002),
Kang (2002), Bobaljik (2003), Kobuchi-Philip (2003, 2007), Puskás (2003),
Terada (2003), Tsoulas (2003), Nakanishi (2004, 2008), Janke & Neeleman
(2005), Fitzpatrick (2006), Miyagawa (2006b, 2017b), Endo (2007b), Mi-
yagawa & Arikawa (2007), Roehrs (2008), Spector (2008, 2009), Tanaka
(2008), Valmara (2008), Cirillo (2009, 2012), Kim & Kim (2009), Heejeong
Ko (2010), Koopman (2010), Reed (2010), Rezak (2010), Rochman (2010),
Simpson (2011), Harwood (2012), Henry (2012), Ko & Oh (2012), Ott (2012b),
Jenks (2013), Osborne (2013), Fukuda & Polinsky (2014), Lacerda (2016),
Otoguro & Snijders (2016), Chaiphet (2017), Kostopoulos (2017), Kitagawa
(2018), Zyman (2018b), Al Khalaf (2019), Kim (2020), and Netkachev (2020).

6.7 Workbook

In the exercises below, an [M] after an example sentence indicates that a

model answer is provided in the free-to-download Students’ Answerbook
(and in the Teachers’ Answerbook as well). You will find it helpful to look at
this model answer before attempting to tackle other examples in the exer-
cise. Some other exercise examples have an [S] after them, indicating that
they are for self-study: this means that after reading the relevant module
(and those preceding it), and looking at the helpful hints and model answer
associated with the exercise, students should be able to analyse these self-
study examples on their own, and then compare their answers with those
provided in the Students’ Answerbook. The remaining exercise examples
 6.7 Workbook 341

(i.e. those not marked with [M] or [S] after them) are intended for teachers
to use as the basis for hands-on problem-solving work in seminars, classes,
or assignments: answers to all of these non-self-study examples are pro-
vided in the Teachers’ Answerbook. Note that Exercise 6.1 tests you on
Module 6.1, Exercise 6.2 on Module 6.2, and so on; and that each exercise
in a given chapter presupposes familiarity with material covered in earlier
chapters and in earlier modules of the same chapter. Unparenthesised num-
bers like 4 refer to material in the exercises, whereas parenthesised numbers
like (22) refer to examples and conditions in the main text of the book.

EXERCISE 6.1

This exercise is designed to be attempted after reading Module 6.1 and the
chapters preceding it.

Discuss the syntax of the subperiphery of the sentences below (concerning

yourself only with the bracketed complement clause in 2):

1 He probably need not worry about that ([Link]) [M]

2 I know [what song I definitely really truly do not want played at my funeral],
and that is Amazing Grace ([Link]) [S]
3 He clearly couldn’t do anything
4 He potentially may not have been telling the truth
5 What else could he possibly have done? [S]
6 What else could potentially he be doing? (Bret Baier, Fox News)

For the interrogative clauses in 2, 5, 6, consider also the structure of the

periphery.

Helpful hints
For the purposes of this exercise, follow Module 6.1 and adopt the
traditional analysis of adverbial modifiers as adjuncts to the constituent
they modify. Don’t concern yourself with the internal structure of verb
phrases in any of the sentences. In 2, assume that what song is an
interrogative QP, and that not want played at my funeral is a NEGP,
but don’t concern yourself with the internal structure of either. In 3,
assume that n’t is a clitic which originates as the head NEG of NEGP,
and encliticises/attaches to the end of a tensed auxiliary (i.e. an auxiliary
carrying an overt or null tense affix) in T: for present purposes, treat
could as the past tense of can and assume that could spells out a can+Af
structure, where Af is a past tense affix in T. In 4, treat not as an ADVP
which is the specifier of a NEGP with a null NEG head; in addition, treat
342 The subperiphery

have been telling the truth as a PERFP/perfect aspect phrase, but don’t
concern yourself with its internal structure. In 5 and 6 treat what else as
an interrogative QP, but don’t concern yourself with its internal structure.
Discuss what role is played by the following (where relevant):

7 One Feature One Head Principle (= 6)

Each functional head carries only one interpretable feature

8 Like-Adjoins-to-Like Constraint (= 20)

Adjunction can only adjoin like to like

9 Minimality Condition (from §1.3.2)

A moved constituent must move to the minimal/closest potential landing
site above it

10 Interrogative Generalisation (from §5.2.1)

(i) A peripheral interrogative XP that triggers Auxiliary Inversion moves
from spec‑FINP (with FIN attracting a finite auxiliary to adjoin to it)
into its criterial position in spec‑FOCP
(ii) The criterial position for a peripheral interrogative XP that does not
trigger Auxiliary Inversion is spec‑FORCEP

EXERCISE 6.2

This exercise is designed to be attempted after reading Module 6.2 and the
module and chapters preceding it.

Discuss the syntax of the subperiphery of the sentences below:

1 He perhaps already has hired someone [M]

2 He frankly usually doesn’t completely trust anyone
3 They unfortunately couldn’t afford it [S]
4 They probably will soon have finished it
5a She doesn’t often have time for lunch
5b She often doesn’t have time for lunch
6 He sadly wasn’t fully paying attention [S]

Helpful hints
Compare the traditional CP+TP+VP model of clause structure in which
subperipheral adverbs are treated as adjuncts with the alternative
functional specifier analysis outlined in Module 6.2. In negative
sentences, treat n’t as the head NEG constituent of NEGP, and take it to
be a clitic which (in the syntax) encliticises to (i.e. attaches to the end/
 6.7 Workbook 343

right of) a tensed auxiliary in T (i.e. an auxiliary inflected for present/

past tense). Concern yourself only with the structure of the subperiphery,
and do not attempt to analyse the internal structure of verb phrases. In
developing your analyses, discuss the role played by the following (where
relevant):

7 One Feature One Head Principle (= 6)

Each functional head carries only one interpretable feature

8 Like-Adjoins-to-Like Constraint (= 20)

Adjunction can only adjoin like to like (i.e. one head to another, or one
maximal projection to another)

9 Adverb hierarchy (= 36)

[frankly MOODspeech act [fortunately MOODevaluative [allegedly MOODevidential
[probably MODepistemic [once Tpast [then Tfuture [perhaps MOODirrealis
[necessarily MODnecessity [possibly MODpossibility [usually ASPhabitual
[again ASPrepetitive(I) [often ASPfrequentative(I) [intentionally MODvolitional
[quickly ASPcelerative(I) [already Tanterior [no longer ASPterminative
[still ASPcontinuative [always ASPperfect(?) [just ASPretrospective [soon ASPproximative
[briefly ASPdurative [characteristically(?) ASPgeneric/progressive [almost ASPprospective
[completely ASPcompletive(I) [well VOICE [fast/early ASPcelerative(II)
[again ASPrepetitive(II) [often ASPfrequentative(II) [completely ASPcompletive(II)

10 Functional Specifier Hypothesis/FSH (= 53)

Every XP in the periphery or subperiphery of a clause is the specifier of
a dedicated functional head

11 Merge Hypothesis (= 54)

A maximal projection/XP can only be introduced into a structure
containing it via (first- or second-) merge with a head

EXERCISE 6.3

This exercise is designed to be attempted after reading Module 6.3 and the
modules and chapters preceding it.

Account for word order variation in the position of the italicised ADVPs/
PPs in the following sets of sentences:

1a I should have probably reported it immediately [M]

1b I should probably have reported it immediately [M]
1c Probably I should have reported it immediately [M]
1d I probably should have reported it immediately [M]
344 The subperiphery

2a This is possibly the best outcome

2b Possibly this is the best outcome
2c This possibly is the best outcome
3a He must surely still be in his room
3b He surely must still be in his room
3c Surely he must still be in his room
4a He usually doesn’t at any point interrupt her [S]
4b Usually he doesn’t at any point interrupt her [S]
4c He doesn’t usually interrupt her at any point [S]
5a They could potentially entirely reject the deal
5b They could potentially reject the deal entirely
5c Potentially they could reject the deal entirely
5d They potentially could entirely reject the deal
5e They potentially could reject the deal entirely

Helpful hints
Compare the adjunct analysis of ADVP/PP modifiers in the traditional
CP+TP+VP model of clause structure with the alternative functional
specifier account outlined in Module 6.3. Concern yourself only with the
structure of the subperiphery (not with the periphery), and do not attempt
to show the internal structure of verb phrases/VPs in any of the examples.
In 2, bear in mind the traditional assumption that b e (in its use as a
copula/linking verb) originates in V and from there moves to adjoin to a
tense affix in T in indicative clauses. In 4, take n’t to originate as the head
NEG constituent of a NEGP constituent, and assume that it encliticises to
(i.e. adjoins to the end/right of) a finite auxiliary in T. In developing your
analyses, discuss what role (if any) is played in relevant examples by the
principles/hypotheses/hierarchy outlined in 7–11 of Exercise 6.2.

EXERCISE 6.4

This exercise is designed to be attempted after reading Module 6.4 and the
modules and chapters preceding it.

Discuss the syntax of the subperiphery in the sentences below:

1 The Brownlee brothers, in spite of leading the race, both were eventually
overtaken [M]
2 They undoubtedly had both deeply regretted the decision
3 The men had evidently all blatantly disobeyed orders [S]
4 The fighters, throughout, although tired, had both gamely swapped punches
 6.7 Workbook 345

5 They allegedly each are still blaming the others

6 She probably hasn’t at all (In reply to ‘Has she heard from him?’) [S]

Helpful hints
Compare the adjunct analysis of peripheral modifiers in the traditional
CP+TP+VP model with the functional specifier account. Concern yourself
only with the structure of the subperiphery of the sentences, and do not
analyse the internal structure of the verb phrases they contain. In 1–5,
consider whether floating quantifiers are (i) adjuncts, or (ii) stranded in
the specifier position of an intermediate auxiliary projection through
which the subject transits, or (iii) specifiers of a dedicated functional
head which licenses a floating QP as its specifier. In 1, treat the Brownlee
brothers as a DP, and in spite of leading the race as a PP but don’t
concern yourself with the internal structure of either phrase. In 4 treat
throughout as a PP (paraphrasable as ‘throughout the fight/event’), and
treat although exhausted as a CONJP/subordinate clause (a reduced form
of although they were exhausted, perhaps) – but do not concern yourself
with the internal structure of these constituents. Treat 6 as involving a VP
ellipsis operation whereby a VP present in the syntax can be given a null
spellout in the PF component if its contents are recoverable (e.g. from an
antecedent). Discuss the role played by the following (where relevant):

7 One Feature One Head Principle (= 6)

Each functional head carries only one interpretable feature

8 Like-Adjoins-to-Like Constraint (= 20)

Adjunction can only adjoin like to like (i.e. one head to another, or one
maximal projection to another)

9 Constraint on Extraction Domains/CED (= 48)

Extraction is only possible out of a complement, not out of a specifier or
adjunct

10 Freezing Principle (= 49)

The constituents of a moved phrase are frozen internally within (and so
cannot be extracted out of) the moved phrase

11 Adverb Hierarchy (= 36)

[frankly MOODspeech act [fortunately MOODevaluative [allegedly MOODevidential
[probably MODepistemic [once Tpast [then Tfuture [perhaps MOODirrealis
[necessarily MODnecessity [possibly MODpossibility [usually ASPhabitual
[again ASPrepetitive(I) [often ASPfrequentative(I) [intentionally MODvolitional
[quickly ASPcelerative(I) [already Tanterior [no longer ASPterminative
[still ASPcontinuative [always ASPperfect(?) [just ASPretrospective [soon ASPproximative
[briefly ASPdurative [characteristically(?) ASPgeneric/progressive [almost ASPprospective
346 The subperiphery

[completely ASPcompletive(I) [well VOICE [fast/early ASPcelerative(II)

[again ASPrepetitive(II) [often ASPfrequentative(II) [completely ASPcompletive(II)

12 Functional Specifier Hypothesis/FSH (= 53)

Every XP in the periphery or subperiphery of a clause is the specifier of
a dedicated functional head

13 Merge Hypothesis (= 54)

A maximal projection/XP can only be introduced into a structure
containing it via (first- or second-) merge with a head

14 Default Spellout Rule/DSR (= 60)

For a constituent whose spellout is not determined in some other way,
the highest copy of the constituent is overtly pronounced, and any lower
copies are silent

15 Floating Quantifier Spellout Rule/FQSR (= 61)

Floatable quantifiers can be spelled out on any non-initial link of an
A‑Movement chain in standard varieties of English
7 Abbreviated registers

7.0 Overview

In this module, we look at abbreviated registers of (spoken or written)

English which allow certain types of constituent to be dropped (i.e. not
overtly expressed), resulting in sentences which lack overt subjects, or
auxiliaries, or articles, or objects, and so on. We will analyse different
kinds of constituent drop found in different registers of English, and
ask whether this is a syntactic phenomenon (whereby certain functional
projections are missing in the syntax), or a phonological one (whereby
material present in the syntax is given a null spellout in the phonology).
We deal with Subject Drop in Module 7.1, Auxiliary Drop in Module 7.2,
Article Drop in Module 7.3, e Drop in Module 7.4, and Object Drop in
Module 7.5.

7.1 Subject Drop

In abbreviated registers/styles of spoken English, we find a phenomenon

which I will refer to as Sentence-Initial Ellipsis, whereby sentences can
be abbreviated by allowing one or more constituents at the beginning of
a sentence to be dropped. In this module, we examine one phenomenon
of this kind (referred to here as Subject Drop) by which a subject can be
dropped when it is the first word in a sentence.

7.1.1 The nature of Subject Drop

The phenomenon of Subject Drop can be illustrated by sentences like those
below, in which the subject in <angle brackets> is ‘dropped’ in the sense
that it is unpronounced:

(1) a. s p ea k er a :
Would you care for some tea?
s p ea k er b : <I> wouldn’t say no (Mackenzie 1998: 288)

b. s p ea k er a : Am I invited to the party?

s p ea k er b : <You> must be, surely (Weir 2012: 106)

c. <I> just saw Bert. <He> looks like death warmed over (Zwicky &
Pullum 1983: 159)
348 Abbreviated registers

d. M onica : Okay, everybody relax. This is not even a date. It’s just
two people going out to dinner and not having sex.
hand er : <It> sounds like a date to me (TV show Friends, Ozaki

2010: 39)
e. speaker a : I saw Rachel yesterday. I brought her flowers and we
had a chat
speaker b : <It> clearly did her a lot of good (Nariyama 2004: 252)

f. <There> isn’t much we can do about it (Thrasher 1977: 44).

g. <It> turns out you can’t do that in Texas (Schmerling 1973: 582).

As these examples illustrate, constituents which undergo Subject Drop are

typically personal pronouns (like those bracketed in 1a–1e), or expletive
pronouns like those bracketed in (1f, 1g). This may be because such pro-
nouns have little or no lexical semantic content, and instead are func-
tors (i.e. function words) essentially comprising bundles of grammatical
features.
There is empirical evidence in support of positing that sentences like
those in (1) have a <bracketed> subject which is present in the syntax, but
not pronounced in the phonology. One such piece of evidence comes from
the agreement morphology on the (bold-printed) finite auxiliary/verb in
sentences such as the following:

(2) a. <I> am not feeling too good this morning

b. <You> weren’t the only one, were you?
c. speaker a : How’s the veal today?
speaker b : <It> has been better (Napoli 1982: 95)
d. He’s a wild card. <He> flies by the seat of his pants (dialogue from
the film Top Gun; Ito & Kashirara 2010: 21)
e. speaker a : You want to use that?
speaker b : <It> depends on how big a news day it is (dialogue

from film Broadcast News; Ito & Kashihara 2010: 28)

Since finite auxiliaries/verbs agree in person and number with their sub-
jects, it’s hard to see how we can (for example) account for use of the
first person singular form am in (2a) unless we posit that am has (a silent
counterpart of) the first person subject pronoun I as its subject; and similar
considerations hold for (2b–2e).
A second piece of evidence in support of positing that sentences like
those in (1) have syntactic subjects comes from tag sentences like those
below:

(3) a. <We> mustn’t jump to conclusions, must we?

b. <You> blew it, didn’t you? (Kay 2002)
c. <He> can’t sing a note, can he? (Napoli 1982: 91)
 7.1 Subject Drop 349

d. <They> can’t take everybody, can they? (Bailey 2011: 24)

e. <It> feels good to have them on again, doesn’t it? (Shibata 2017: 2)
f. <There> isn’t any pizza left by any chance, is there?

Tag sentences have the property that the tag (following the comma) con-
tains a (bold-printed) pronominal copy of the main clause subject. Thus,
the occurrence of the bold-printed pronoun I in the tag in (3a) provides
evidence for claiming that the main clause must have a first person singu-
lar subject (albeit unpronounced). And a parallel logic can be applied to the
examples in (3b–3f ).
A third piece of evidence in support of positing that reduced sentences
have syntactic subjects comes from sentences containing anaphors (and
other expressions requiring a local antecedent) like those bold-printed be-
low:

(4) a. <I> accidentally cut myself while shaving

b. <We/You/They> took each other/one another for granted
c. <She> ought to mind her own business
d. <They> shouldn’t lose their cool in front of the children

Since the bold-printed expressions in (4) require a local antecedent with

matching person/number/gender features, the grammaticality of the rele-
vant sentences suggests that they have a syntactic subject with matching
features. Moreover, a subject is required in order to bear the theta role
assigned to it by the relevant predicate: for example, <I> is the e x pe r i e nc e r
argument of cut in (4a).
A claim implicit in the discussion of the examples in (1–4) above is that
the missing subject is a pronoun, and not a full DP. Evidence that this is
indeed the case comes from the following dialogue, where <?> indicates a
missing subject whose precise identity is in question:

(5) s p ea k e r a : What do you think of Dubai?

s p ea k e r b : <?>’s a great place for a holiday

The use of the contracted third person singular copula ’s here provides evi-
dence that the missing subject is third person singular. But is the missing
subject the pronominal DP it as in (6a) below, or the lexical DP Dubai as
in (6b)?

(6) a. <It>’s a great place for a holiday

b. <Dubai>’s a great place for a holiday

A clue to the answer comes from the observation that the string ’s a in
(5 b ) has the phonetic spellout [sǝ], not [zǝ]. This can be accounted for in a
straightforward fashion if the missing subject in (5 b ) is it as in (6a), since
350 Abbreviated registers

the word it ends in the voiceless consonant [t], and a contracted auxiliary
assimilates to the last segment of the immediately preceding word it cliti-
cises to in respect of voicing and hence is realised as voiceless [s] after the
voiceless consonant [t]. By contrast, if the missing subject had been Dubai
as in (6b), we would (wrongly) have predicted that the auxiliary would have
been realised as the voiced segment [z], since this is the form of the con-
tracted auxiliary required when it cliticises onto a word ending in a vowel/
diphthong, like the diphthong [ai] in Dubai. Such considerations provide
evidence that the ‘missing’ subject is a pronoun (like it) which has phonetic
features of its own, and is not inherently null. In this respect, it differs from
the inherently null subject of a coordinate clause like the and-clause in:

(7) Muscat is in the Gulf of Oman and’s only a stone’s throw from the desert

In (7), the contracted auxiliary’s is pronounced [z], consistent with the

view that the subject of ’s is an empty category with no phonetic features
of its own, with the result that ’s undergoes progressive voice assimilation
to the voiced final consonant [d] of and. (See Wilder 1994a, 1994b, and
Haegeman & Starke 2021 for arguments that the subject of a coordinate
clause is an empty category.)
Further evidence that the missing subject in Subject Drop structures is
a pronoun comes from ellipsis in clausal idioms. Huang & Mendes (2019)
note that where a clausal idiom undergoes VP Ellipsis, its subject can only
be pronominal, not nominal – as illustrated by the following examples:

(8) a. speaker a : When the news got out, the shit hit the fan
speaker b : No it/*the shit didn’t

b. speaker a : It looks like the chickens have come home to roost

speaker b : Yes, I’m afraid they/*the chickens have

Significantly, a VP-ellipsed clausal idiom allows a silent subject in (9b )

below, where <?> denotes a mystery missing subject:

(9) speaker a : Did the shit hit the fan?

speaker b : <?> Sure did!

Given that (8a, 8b) tell us that a clausal idiom only allows VP Ellipsis if it
has a pronominal subject, it follows that the mystery missing subject <?>
in (9) must be the pronoun it, and not the nominal the shit.
Our discussion so far has led to the conclusion that Subject Drop involves
sentences that have a pronominal subject which is present in the syntax,
but is given a null/silent pronunciation in the phonology. A question which
this assumption poses is: under what conditions/in what contexts can a
 7.1 Subject Drop 351

subject be dropped? Two main conditions have been proposed governing

the use of Subject Drop in spoken English. Firstly, it only applies in root/
main clauses like those in (1) above, not in non-root clauses like those in
square brackets below:

(10) a. *I know [<you> gotta leave now] (Zwicky & Pullum 1983: 159)
b. *[Because <it> was raining hard], we decided to stay at home
c. *She is someone [that <we> know well]
d. *There is more snow today [than <there> was yesterday]
e. *The pitch was so flooded [that <they> had to call off the game]
f. *[Clever though <he> is], he can’t beat his 9-year-old son at chess
g. *[Had <you> been there], you would really have enjoyed it

Thus, the subject cannot undergo ellipsis in a bracketed clause which is a

complement clause as in (10a), an adverbial clause as in (10b), a relative
clause as in (10c), a comparative clause as in (10d), a consecutive clause
as in (10e), a concessive clause as in (10f), or a conditional clause with an
inverted auxiliary as in (10g). Numerous studies have observed this restric-
tion holding in spoken English: for example, Nariyama (2004: 245) reports
that in his empirical study ‘No subject ellipsis was found for subordinate
clauses.’
A second widely reported condition on Subject Drop in spoken English is
that it only applies where the missing subject is the first word in the sen-
tence containing it. Accordingly, it is permitted in sentences like (1) above
where this ‘first word’ requirement is met, but not in sentences like those
in (11) below:

(11) a. *Which pub did <you> go to?

b. *Not a single word did <he> say
c. *Have <you> seen any good movies lately? (Napoli 1982: 93)
d. *What a great time <we> had!
e. *Crying like a baby <he> was

The dropped subject is preceded by an (italicised) preposed interrogative

wh‑phrase and inverted auxiliary in (11a), a preposed negative phrase
and inverted auxiliary in (11b), an inverted auxiliary in (11c), a preposed
exclamative wh‑phrase in (11d), and a fronted VP in (11e). Sentences like
(11) are ruled out by the constraint that a subject can only be dropped when
it is the first word in its sentence. (For discussion of apparent – but not
real – exceptions to this first word condition, see Haegeman 2019.) Why
should it be that Subject Drop in spoken English only affects sentence-
initial subjects in root clauses? In the next section, we’ll look an interesting
answer provided by Luigi Rizzi, involving a mechanism termed Truncation.
352 Abbreviated registers

7.1.2 A Truncation account of Subject Drop

Rizzi (1994, 1998, 2000a) argues that Subject Drop sentences have a
syntactic structure which is truncated above SUBJP (in the sense that
peripheral projections above SUBJP are not present in the syntax), leav-
ing the subject as the highest constituent in the structure, and vulner-
able to being deleted (i.e. given a silent spellout) in the phonology. More
specifically, Rizzi argues that Subject Drop in spoken English can be
handled in terms of a Truncation analysis which involves the assump-
tions below:

(12) Syntactic truncation account of Subject Drop

(i) A sentence can have a truncated syntactic structure whose root is not
FORCEP, but some lower projection such as SUBJP
(ii) The edge of the highest projection in a sentence structure can be
unpronounced in the PF component, provided that it contains only
phonetically weak material with recoverable semantic content

The edge of a given projection XP comprises its head X, and any other
constituent of XP that c‑commands X: in typical cases, this means that
the edge of a projection XP includes the head of XP and any specifier
that it may have. The ‘provided that’ condition in (12ii) is a reflection of
the Recoverability Condition on deletion proposed in Chomsky (1964b: 41)
and Katz & Postal (1964: 79) to the effect that material can only be unpro-
nounced if its content is recoverable. In consequence of this condition, only
functors (i.e. functional categories with no substantive lexical semantic
content) can be unpronounced.
Having sketched the assumptions underlying the Truncation account,
let’s look at how this approach handles Subject Drop in colloquial English.
Rizzi argues that this is the result of the root sentence structure being trun-
cated above SUBJP, leaving SUBJP as the root/highest projection in the
structure. In the light of this, consider the structure of speaker b’s sentence
in the dialogue below (slightly adapted from 1e above):

(13) speaker a : I saw Rachel yesterday. I brought her flowers and we had a
chat
speaker b : <It> clearly has done her a lot of good

On Rizzi’s account, (13 b ) will have a truncated structure along the lines
shown in simplified form below (with the functional head F being an evi-
dential mood head under the analysis of adverbs like clearly in Cinque
1999, and the perfect aspect auxiliary has originating as a PERF head and
from there raising to T):
 7.1 Subject Drop 353

(14) SUBJP

DP SUBJ′
it
SUBJ FP
ø
ADVP F′
clearly
F TP
ø
T PERFP
has done her a lot of good
Because the pronoun it is on the edge of the root projection SUBJP, it can
be unpronounced at PF (since it is phonetically weak, and it has no lexical
semantic content), in accordance with (12ii). (An incidental point to note in
passing is that here and elsewhere, I set aside the possibility that subjects
transit through spec‑TP on their way to spec‑SUBJP.)
The assumption that Subject Drop sentences are truncated and project
no peripheral constituents above SUBJP accounts for why Subject Drop is
not found in sentences like those in (11) above, since it is clear that in each
sentence in (11) there will be one or more projections above the SUBJP
projection containing the unpronounced subject. For example, if root
wh‑questions involve an inverted auxiliary moving from SUBJ to FIN, and
a focused wh‑operator moving through spec‑FINP into spec‑FOCP (as ar-
gued in §5.1.2), a sentence like (11a) Which pub did you go to? will have the
derivation below (simplified by not showing the internal structure of the
VP constituent, and not showing movement of the subject you from inside
VP to the edge of SUBJP, in order to reduce visual complexity):
(15) FORCEP

FORCE FOCP
ø
QP FOC′
which
pub FOC FINP
ø
QP FIN′
which
pub FIN SUBJP
did+ø
DP SUBJ′
you
SUBJ TP
did
T VP
did go to which pub
354 Abbreviated registers

The outcome is a structure in which the (bold-printed) subject you is not on

the edge of a root projection (since it is in spec‑SUBJP, and the root projec-
tion is FORCEP). Consequently, the subject cannot be null/unpronounced
via (12ii) – thereby accounting for the ungrammaticality of (11a) *Which
pub did <you> go to?
As we have seen, Rizzi argues that Truncation in abbreviated styles of
spoken English can be handled by treating ‘certain informal registers of
English … as allowing truncation’ (Rizzi 1998: 28). By contrast, more for-
mal registers of English do not allow ellipsis of subjects. Why should this
be? Rizzi’s answer is that more formal varieties don’t allow truncation of
the clause periphery, so that the counterpart of (14) in more formal styles is
the untruncated FORCEP structure below:

(16) FORCEP

FORCE FINP
ø
FIN SUBJP
ø
DP SUBJ′
it
SUBJ FP
ø
ADVP F′
clearly
F TP
ø
T PERFP
has done her a lot
of good
Since the pronoun it is not on the edge of a root projection in (16) (because
it is in spec‑SUBJP, and the root projection is FORCEP), it cannot receive
a null spellout. On this view, the difference between informal registers
of English which allow Subject Drop in root clauses and formal registers
which don’t is that informal registers allow truncated structures like (14)
above in which there are no peripheral projections above SUBJP, whereas
formal registers don’t and instead require full/untruncated FORCEP struc-
tures like (16).

7.1.3 Subject+Auxiliary Drop

Although our discussion hitherto has centred on Subject Drop, the assump-
tion that peripheral projections above SUBJP can be truncated enables us
to account for sentences such as the following, where both a subject and an
 7.1 Subject Drop 355

auxiliary following it can undergo sentence-initial ellipsis and be dropped

at the beginning of a sentence.

(17) a. <I’ve> been waiting for over an hour

b. <I’m> not feeling too good at present
c. <There’s> been a bit of a mix-up
d. <I’ll> see you in the morning (Mackenzie 1998: 288)
e. <I’d> love to see you ([Link])
f. <It’s> been a long time (Nariyama 2004: 247)

If we take such sentences to be truncated above SUBJP and the truncated

subject and auxiliary to be on the edge of the same SUBJP projection (which
is plausible, since the auxiliary cliticises onto the subject in 17), then a sen-
tence like (17a) will have the truncated syntactic structure shown below
(simplified by not showing the internal structure of TP, and not showing
have raising from T into SUBJ):

(18) SUBJP

DP SUBJ′
I
SUBJ TP
(ha)ve been waiting for over an hour
Under the Truncation account, both the subject I and the auxiliary have can
receive a null spellout at PF via (12ii), because they are phonetically weak,
because they are functors with no lexical semantic content, and because
they are on the edge of a root projection (by virtue of occupying the speci-
fier/head positions of the root projection SUBJP).
The SUBJP analysis can also be extended to sentences such as the fol-
lowing, where only the subject undergoes ellipsis, not the auxiliary:

(19) a. <It> isn’t raining much at present

b. s p ea k er a : How are you?
s p ea k er b : <I> could be better (Napoli 1982: 94)
c. s p ea k er a : How’s the veal today?
s p ea k er b : <It> has been better (Napoli 1982: 95)

An analysis of (19a) consistent with the SUBJP analysis of Subject Drop

outlined above is to suppose that the subject is on the edge of SUBJP, but
that the negative auxiliary isn’t is in T (where it is adjacent to the NEGP
constituent in which the clitic n’t originates). This would mean that (19a)
has the structure shown in highly simplified form below:
356 Abbreviated registers

(20) SUBJP

DP SUBJ′
it
SUBJ TP
ø
T NEGP
isn’t raining much at present

This would allow the subject it to be unpronounced in accordance with

(12ii), since it is positioned on the edge of the root projection SUBJP.
Although this module has focused on spoken English, it should be noted
that Subject Drop is also found in a wide variety of registers of written
English. For example, Subject Drop is found in diary entries like (21a) be-
low, encyclopedia entries (21b), biographies (21c), instructional materials
(e.g. recipe books 21d, stage directions 21e, product labels 21f), postcards
(21g), emails (21h), text/SMS messages (21i), social media (e.g. Facebook
21j, Twitter 21k), and telegrams (21l). In the examples below, the subject of
each underlined (auxiliary or lexical) verb is omitted/dropped; (21a–21g)
are from Haegeman (2019), and (21h–21l) are from the sources specified:

(21) a. Sat there for 3/4 hours (The Diary of Virginia Woolf 1940: 334)
b. Teeters hind part of body almost constantly, bobs head (entry for
‘Common Sandpiper’ in Collins Nature Guide: Birds of Britain and
Europe 1994: 112)
c. Exhibits paintings in a student art show at 2 the Dalmau Gallery,
Barcelona. Experiments with cubism. (Dali: The Salvador Dali
Museum Collection, Little, Brown & Co., Boston, 1991: 181; cited in
Matushansky 1995).
d. Serves four. If refrigerated, will keep a week to 10 days (E. Kennedy,
Things My Mother Told Me, Branden Press, Boston, 1975: 105)
e. Tosses the pencil. Starts on another ([Link]
f. Contains caffeine (label on a can of Coca Cola)
g. Wish you were here (holiday postcard)
h. Still have horrible cold (Scott 2010: 12)
i. Can’t believe it’s that time of week already (Tagg 2009: 175)
j. Hope you’re happy in your new country (Deneys 2020: 99)
k. Survived my first rugby match (Daeninck 2019: 83)
l. Have been detained (Barton 1998: 51)

The register of written English which has attracted most attention in this
respect is diary writing, and the dropping of constituents in diary entries
has come to be known as Diary Drop. This phenomenon has been exten-
sively researched, particularly by Liliane Haegeman and her colleagues
(see inter alia Haegeman 1987a, 1990a, 1990b, 1997, 2000a, 2002b, 2007a,
 7.1 Subject Drop 357

2013, 2017, 2019, 2021; Haegeman & Ihsane 1999, 2001; Haegeman &
Starke 2021). In Exercise 7.1 of the Workbook module, you can explore
potential differences between Subject Drop in spoken English, and Subject
Drop in the diary register of written English.
To summarise: Module 7.1 began by discussing the phenomenon of Sub-
ject Drop in spoken English, and went on to outline Rizzi’s Truncation analy
sis of it. We saw that under his analysis, truncating peripheral projections
above SUBJP in root clauses leaves the subject as a root specifier (hence
the highest constituent present in the syntax), and this allows material on
the edge of a root projection (its head and any specifier that it may have)
to be unpronounced in the PF component. As we have seen, the Truncation
analysis provides a principled syntactic account of why only the subjects of
root clauses can generally undergo this kind of ellipsis in spoken English,
and only when they are sentence-initial. The same Truncation analysis can
also account for how both subject and auxiliary can undergo sentence-
initial ellipsis and be dropped in sentences like (17).
The Truncation analysis outlined here provides a principled account of
Subject Drop in spoken English: nonetheless, it runs into a number of po-
tential problems. One such is the following: if FIN is the locus of finiteness/
mood features in a clause, how does a finite T auxiliary like has in (14)
acquire its indicative mood/finiteness feature if the clause is a truncated
structure containing no FINP projection? Is a truncated clause assigned
indicative mood by default? Likewise, if (as suggested in §4.3.2) the subject
of a finite clause is assigned nominative case by a FIN head immediately
above it, how does the subject it get case-marked in a structure like (14)
if the truncated structure contains no FINP projection? After all, if (pro)-
nominals must have case, the absence of a case assigner for the subject in
a truncated structure like (14) will potentially cause the derivation to crash.
A second potential problem comes from the absence of FORCEP in a
truncated structure such as (14). After all, since the function of FORCEP is
to mark the force of a clause, the question posed by the truncation analysis
is how a truncated SUBJP structure like (14) comes to be interpreted as de-
clarative in force, if it contains no FORCE projection. A plausible answer
is that an indicative clause structure like (14) is interpreted as declarative
in force by default, in accordance with the clause-typing conditions below
(from Module 1.1):

(22) Clause Typing Conditions

(i) A clause is interpreted as being of a given type (interrogative or
exclamative etc.) if it contains a peripheral clause-typing specifier of
the relevant type
(ii) An indicative clause is interpreted as declarative in type by default if
it contains no peripheral clause-typing specifier
358 Abbreviated registers

Since (14) is an indicative clause containing no peripheral clause-typing

specifier, it is interpreted as declarative by default via (22ii).
You should now be able to tackle Exercise 7.1.

7.2 Auxiliary Drop

In addition to Subject Drop, colloquial English also allows another type of

sentence-initial ellipsis which I will refer to as Aux(iliary) Drop, whereby
an inverted auxiliary can undergo ellipsis if it is the first word in its sen-
tence. In this module, we look at this phenomenon and explore a range of
alternative accounts.

7.2.1 A Truncation account of Auxiliary Drop

The phenomenon of Aux Drop in colloquial English can be illustrated by
root/main clause yes–no questions such as the following:

(23) a. <Does> anyone want a cup of tea? (Rühlemann 2012: 5)

b. <Are> you not winning the game? (Schirer 2008: 37)
c. <Have> you seen any good movies lately? (Zwicky & Zwicky 1981:
536)

Aux Drop sentences like (23) clearly can’t be truncated SUBJP structures,
if (as argued in Module 5.1) inverted auxiliaries are on the edge of FINP.
Still, we might be able to preserve the spirit of the Truncation analysis if
we supposed that sentences like (23) are truncated FINP structures which
contain a null yes–no question operator in spec‑FINP that attracts the
inverted auxiliary to adjoin to FIN. This would mean that a sentence
like (23a) involves the successive Head Movement operations arrowed in
the structure below (simplified by not showing the internal structure of
VP, nor the null heads that the inverted auxiliary adjoins to each time it
moves):

(24) FINP

ADVP FIN′
OpYNQ
FIN SUBJP
does
QP SUBJ′
anyone
SUBJ TP
does
T VP
does want a cup of tea
 7.2 Auxiliary Drop 359

The yes–no question operator in spec‑FINP c‑commands (and hence

licenses use of) the polarity item anyone. The structure (24) is interpreted
as a yes–no question via (22i) by virtue of containing a yes–no question
operator in the clause periphery (here, in spec‑FINP). Such a structure
would allow for does to be unpronounced at PF via (12ii) because it is a
weak functor/auxiliary with no lexical semantic content, and because it
it is on the edge of the root projection FINP. Moreover, an analysis like
(24) can account for why the verb want is in the infinitive form in (23a) –
namely because want originates as the (head V of the VP) complement of
the auxiliary do (which selects an infinitive complement); likewise win-
ning in (23b) is in the progressive participle form because it originates as
the complement of the progressive auxiliary are; and seen in (23c) is in
the perfect participle form because it originates as the complement of the
perfect auxiliary have.
A question arising from the analysis of Aux Drop outlined in (24) is
why the non-modal auxiliaries do/be/have can have a silent spellout (as
we see in 23 above), but modal auxiliaries generally can’t – as illus-
trated by the examples below (showing judgements reported in Fitzpat-
rick 2005b):

(25) a. *<Can> anyone pick up John at the airport?

b. *<Will> anyone play the piano at the party tomorrow?
c. *<Could> anyone have picked up John at the airport yesterday?
d. *<Would> everyone be happier if classes were cancelled?
e. *<Should> everyone leave if the neighbours complain?

The answer to this question given by Schirer (2008: 18) is that ‘Auxiliaries
which cannot be deleted have semantic content which cannot be recovered,
while the deletion-eligible auxiliaries do, be, and have do not.’ This account
appeals to the Recoverability Condition of Chomsky (1964b) and Katz &
Postal (1964).
This same intuition is given a parsing implementation by Pesetsky in
terms of the following principle:

(26) Principle of Unambitious Reverse Engineering (PURE)

When determining the identity of unpronounced material in the course
of reverse-engineering a speaker’s syntactic derivation, the language
system of the hearer considers only the minimally semantically
contentful possibilities compatible with the morphosyntactic
environment (Pesetsky 2020: 5)

(The term ‘reverse-engineering’ means ‘parsing the structure produced by

a speaker in order to try and work out the identity of the missing constit-
uent/s’. The term ‘minimally semantically contentful possibilities’ means
360 Abbreviated registers

‘constituents with as little semantic content as possible’.). So, when try-

ing to identify the missing auxiliary in (23a) Anyone want a cup of tea?
the hearer conjectures that the missing auxiliary is does, because this is
the auxiliary with the least semantic content which allows a complement
headed by an infinitival verb like want, and also allows a third person
singular subject like anyone. Under the PURE account, the reason why the
auxiliaries have/be/do can be dropped but modals cannot is that modals
have too much semantic content to reverse-engineer.
And yet, sentences such as the following suggest that some modals can
sometimes be dropped:

(27) a. <Would> you like to go fishing this afternoon?

b. <Would> you mind if I went home early today?
c. <Will> tomorrow at four be okay? (after Carter & McCarthy
1995: 147)

How come? A plausible answer is that modal drop is possible where the
overall content of the sentence gives sufficient clues to the nature of the
missing modal. So, for example, would like to is a frequent collocation,
so it’s easy enough to infer that the missing modal in (27a) is would.
Similarly, since would is used in counterfactual conditionals like I would
feel annoyed if he went home early, once again it can readily be inferred
that the missing modal in (27b) is would. In much the same way, the
presence of the future expression tomorrow at four in (27c) enables the
inference to be drawn that the missing auxiliary is the future modal will.
On this view, all unstressed auxiliaries can in principle be dropped, but in
cases where a modal is dropped, hearers may struggle to reverse-engineer
(i.e. identify) the missing modal if the sentence contains insufficient clues
as to its identity.
However, the Truncation account of Aux Drop outlined in (24) above faces
a potential theoretical problem. Under the analysis of root/main clause
yes–no questions in §5.3.1, the interrogative operator in yes–no questions
is focused, and its criterial position is spec‑FOCP. It therefore follows that
the truncated FINP structure in (24) above will crash at the LF interface,
because the operator is not in the criterial spec‑FOCP position required for
the structure to be interpreted as a yes–no question. One potential way of
circumventing this problem while still maintaining that Aux Drop involves
Truncation would be to suppose that Aux Drop sentences are truncated
at FOCP. This would mean that an Aux Drop sentence like (23a) <Does>
anyone want a cup of tea? has the truncated FOCP structure in (28) below
(with the yes–no question operator raising to spec‑FOCP, and the inverted
auxiliary does raising to FOC):
 7.2 Auxiliary Drop 361

(28) FOCP

ADVP FOC′
OpYNQ
FOC FINP
does
ADVP FIN′
OpYNQ
FIN SUBJP
does
QP SUBJ′
anyone
SUBJ TP
does
T VP
does want a cup of tea
In the LF component, the structure will be interpreted as a yes–no question
by virtue of containing a focused yes–no question operator in its criterial
position in spec‑FOCP. In the PF component, the operator and the inverted
auxiliary can be unpronounced via (12ii), by virtue of being on the edge of
the root projection (FOCP). The structure in (28) can thus be spelled out at
PF as the Aux Drop structure (23a) Anyone want a cup of tea?
However, a potential empirical problem arising from the truncated FOCP
analysis in (28) is that it requires us to suppose that the inverted auxiliary
does raises from FIN to FOC, in order to get to the edge of FOCP and thus
be eligible to be given a null spellout in the PF component. And yet, the
analysis of questions in §5.1.2 presented empirical evidence from sentences
like (29) below that inverted auxiliaries move only as far as FIN (and not
into FOC) in root/main clause questions:

(29) a. What at present do you need? ([Link])

b. How long, in the light of this promotion, will you stay here? (Ross
1967: 31)
c. What else, with no money to speak of, could he possibly have done?
(Kayne 1998: 155, fn. 66)
d. Under what circumstances, during the holidays, would you go into
the office? (Sobin 2003: 193)
e. Why, when asked, did all the students I interviewed say they would
take ‘study drugs’ again? ([Link])
f. How often, that book, has he praised it in public? (Kayne 1998: 155,
fn. 66)

On the assumption that the italicised interrogative XPs in (29) are in

spec‑FOCP, it follows that the bold-printed inverted auxiliaries cannot be
in FOC, since the two are separated by underlined XPs which (under the
362 Abbreviated registers

Functional Specifier Hypothesis) serve as the specifiers of some other per-

ipheral projection. A more plausible analysis is that the italicised interroga-
tive XPs in (29) are in spec‑FOCP, the underlined intervening XPs are in
spec‑MODP in (29a–29e) and in spec‑TOPP in (29f), and the bold-printed
inverted auxiliaries occupy the head FIN position of FINP. If – contrary
to what is suggested here – inverted auxiliaries moved to FOC, the bold-
printed auxiliaries in (29) would (wrongly) be predicted to precede the
underlined XP modifiers. See §5.1.2 for further evidence supporting the
conclusion that inverted auxiliaries move to FIN, not to FOC.
But if the inverted auxiliary doesn’t move to FOC in (28) and instead
moves only as high as FIN, it would not be eligible for ellipsis, since it
would not be on the edge of the root projection FOCP. The bottom line is
that whichever version of the Truncation analysis we adopt (the FINP analy
sis in 24 or the FOCP analysis in 28), we run into theoretical or empirical
problems (relating to the position of the operator, or of the inverted auxil-
iary) which undermine the analysis.

7.2.2 A Left Edge Ellipsis account

A further problem for the Truncation analysis of Auxiliary Drop outlined
in §7.2.1 is posed by sentences such as the following, which involve both
Aux Drop and Subject Drop:

(30) a. <Do you> mind if I smoke? (Mackenzie 1998: 288)

b. <Are you> going to be long? (Nariyama 2004: 255)
c. <Are there> any pears today? (Napoli 1982: 97)
d. <Is there> anybody else with a comment? (Rühlemann 2012: 5)
e. <Have you> ever been to Chicago? (Schmerling 1973: 580)

At first sight, such sentences might seem unproblematic for the Truncation
account. After all, we might simply assume that they are truncated FINP
structures, so that a sentence like (30a) has the syntactic derivation shown
in simplified form below:

(31) FINP

ADVP FIN′
OpYNQ
FIN SUBJP
do
DP SUBJ′
you
SUBJ TP
do
T VP
do mind if I smoke
 7.2 Auxiliary Drop 363

The null yes–no question operator will license the use of the polarity item
mind, and trigger Inversion of the auxiliary do. We could then suppose that
do (being on the edge of the root projection FINP) receives a null spellout
in the PF component via (12ii). However, this would derive You mind if I
smoke? And while this is perfectly grammatical, it leaves unanswered the
question of how the pronoun you also comes to be dropped in (30a) Mind if
I smoke? After all, you is on the edge of SUBJP, but SUBJP is not the root
projection in the structure, so the ellipsis operation in (12ii) can’t apply
here. Similar problems arise if we suppose that Mind if I smoke? is a trun-
cated FOCP structure with the operator and auxiliary on the edge of FOCP
(as in 28 above): giving a null spellout to the edge of FOCP will still leave
the subject intact on the edge of SUBJP.
But if we reject the assumption in (24, 28, 31) that root yes–no questions
which show Auxiliary (+Subject) Drop have a syntactic structure which is
truncated above FINP or FOCP, what alternative are we left with? An inter-
esting answer is provided by Schirer (2008). He suggests that Truncation
is a phonological rather than a syntactic phenomenon. More specifically,
he proposes that Aux+Subject Drop sentences (i.e. sentences in which both
auxiliary and subject have been dropped) have a full/untruncated syntac-
tic structure which projects all the way up to FORCEP, but that in the PF
component ‘The phonology … allows the deletion of material when it ap-
pears on the left edge of the sentence’ (Schirer 2008: 15). Let’s adopt the
following implementation of his proposal (adapted to fit the framework
used here):

(32) Left Edge Ellipsis account of Auxiliary (+Subject) Drop

(i) Reduced clauses have the same full syntactic structure as their
unreduced counterparts
(ii) The left edge of a continuous set of one or more projections at the
top of a tree in a root clause can undergo ellipsis in the phonology

To see how this account works, consider the sentences below, where (33a)
involves Aux Drop, and (33b) Aux+Subject Drop:

(33) a. <Do> you need any help?

b. <Do you> need any help?

Let’s suppose that both sentences have the full syntactic structure below
(simplified by not showing the internal structure or null constituents of TP):
364 Abbreviated registers

(34) FORCEP

FORCE FOCP
ø
ADVP FOC′
OpYNQ
FOC FINP
ø
ADVP FIN′
Op YNQ
FIN SUBJP
do
DP SUBJ′
you
SUBJ TP
do need any help
The structure in (34) will be handed over to the PF and LF components
at the end of the syntactic derivation. In the LF component, (34) is
interpreted as a yes–no question in accordance with (22i) by virtue of
containing a yes–no question operator on the edge of FOCP. In the PF
component, the Left Edge Ellipsis operation in (32ii) will allow a null
spellout to be given to the left edge of FORCEP, or of FORCEP+FOCP
(i.e. to the left edge of FORCEP as well as to the left edge of FOCP), or of
FORCEP+FOCP+FINP … and so on. This raises the possibility that all the
material on the left edge of the FORCEP+FOCP+FINP projections in (34)
can receive a null spellout at PF, so deriving (33a) You need any help?
An alternative possibility allowed for by (32ii) is that the left edge of the
FORCEP+FOCP+FINP+SUBJP projections can receive a null spellout, so
deriving (33b) Need any help? Thus, the Left Edge Ellipsis account in (32)
can handle both Aux Drop sentences like (33a) and Aux+Subject Drop
sentences like (33b) in a unitary fashion. Under this analysis, truncated
projections are present in the syntax, but given a null spellout in the
phonology. This contrasts with Rizzi’s treatment of truncated constitu-
ents, in which they are not present at any level of the derivation (neither
in the syntax, nor or in the phonology).

7.2.3 A prosodic account

Both the Truncation account outlined in §7.2.1 and the Left Edge Ellipsis
account sketched in §7.2.2 attempt to characterise the deletion domain (i.e.
the set of constituents which can undergo sentence-initial ellipsis) in purely
structural terms as the left edge of one or more peripheral projections.
However, there is evidence that the structural approach is potentially mis-
guided, and that prosodic factors (involving stress and intonation patterns)
play a pivotal role (as argued by Thrasher 1973; Napoli, 1982; Zwicky &
 7.2 Auxiliary Drop 365

Pullum 1983; Gerken 1991, 1994; Sigurðsson & Maling 2007, 2010; and
Weir 2008, 2009a, 2009b, 2012, 2016, 2017, 2018, 2022). To see why, con-
sider the following sentences:

(35) a. Has the professor arrived yet?

b. <Has> the professor arrived yet?
c. <Has the> professor arrived yet?
d. <Has the pro> ’fessor arrived yet? (Napoli 1982: 85)

Let’s suppose that the sentences in (35) all have the (full/untruncated)
syntactic structure below (simplified by not showing the internal structure
of TP):

(36) FORCEP

FORCE FOCP
ø
ADVP FOC′
OpYNQ
FOC FINP
ø
ADVP FIN′
OpYNQ
FIN SUBJP
has
DP SUBJ′

D NP SUBJ TP
the professor has arrived yet

One possibility allowed for by the Left Edge Ellipsis account in (32) is
that the left edge of the FORCEP+FOCP+FINP projections in (36) can
be given a silent spellout at PF, so that only constituents of SUBJP get
spelled out overtly at PF. This will derive (35b) <Has> the professor
arrived yet?
But now consider what happens in (35c) <Has the> professor arrived yet?
It might at first sight appear as if what is unpronounced here is the left edge
of the FORCEP+FOCP+FINP+SUBJP projections; but the left edge of SUBJP
includes the whole of the DP the professor, and yet only the first word
of that DP (= the) is silent, not the whole of the edge of SUBJP. The same
problem arises in (35d) <Has the pro> ’fessor arrived yet? where once again
what is deleted is not the whole DP the professor on the edge of SUBJP, but
rather the word the and the first syllable (pro) of the word professor. Napoli
argues on the basis of such considerations that the domain of the relevant
deletion operation cannot be identified in syntactic terms, because ‘A syn-
tactic deletion rule cannot delete part of a word’ (Napoli 1982: 99). Instead,
she concludes that in sentences like (35b–35d) ‘a phonological deletion rule
366 Abbreviated registers

is operative’ (Napoli 1982: 99). And she makes the following suggestions
about the nature of the rule:

The rule deletes unstressed (or lightly stressed) initial material. Dwight Bolinger
(personal communication) suggests that if we … divide the intonation contour
of a sentence into the prehead (the part preceding the first main accent), the
head (from the first main accent to the last main accent), and the post-head
(the falling part after the last main accent), we could say that the rule deletes
the prehead. (Napoli 1982: 99)

She further notes Bolinger’s suggestion that the rule can operate not only
at the clause level, but also at the phrase level (so that the PP of course
can be truncated to course), and at the word level (so that because can be
truncated to coz).
Weir (2012: 106) characterises the relevant phenomenon as ‘the result of
a general process of weak syllable drop at the left edge of prosodic phras-
es, in order to satisfy a constraint requiring prosodic phrases to start with
accented syllables’. He takes prosodic phrases to include clauses, maximal
projections, and words. He formalises his approach in terms of Optimality
Theory, but I will leave interested readers to read the technical details of his
work for themselves (see Weir 2012: 110–19). For present purposes, I will
characterise the phenomenon informally as follows:

(37) Prosodic account of Weak Syllable Drop

(i) Reduced clauses have the same full syntactic structure as their
unreduced counterparts
(ii) A continuous sequence of one or more weak syllables at the
beginning of a prosodic phrase (e.g. a sentence) can undergo ellipsis
at PF

In the case of a sentence like Has the professor arrived? the three syl-
lables has+the+pro are weak; has is deleted on its own in (35b), has+the
are deleted together in (35c), and has+the+pro are all three deleted in
(35d).
The overall conclusion to be drawn from our discussion in this module
and the last is that sentences involving Subject Drop and/or Auxiliary Drop
in informal styles of spoken English have a full (untruncated) syntactic
structure, but can be given an abbreviated spellout via an ellipsis operation
in the PF component which is structural in nature under the Left Edge El-
lipsis account (32), and phonological in nature under the Prosodic Account
(37) – though only the latter approach can account for the types of reduced
clause found in (35c, 35d).
You should now be able to tackle Exercise 7.2.
 7.3 Article Drop 367

7.3 Article Drop

Thus far in this chapter, I have focused on structures in spoken English in

which subjects and auxiliaries are dropped. However, in this module, I turn
to look at a very different phenomenon (characteristic of certain written
registers) which I shall refer to as Article Drop, since it involves omission
of the definite/indefinite articles the/a (where a should be understood to
include its variant an as well).
Article Drop is restricted to occurring in written English, and is found
in many such registers: for example, a corpus of posts on the social media
platform Twitter collected by Daeninck (2019) includes examples of Article
Drop such as the following:

(38) a. <The> motive was rage fueled by rejection

b. <The> entire cast was on fire
c. <The> suspect planted blood in <the> RAV4 …
d. If not used to these sorts of climbs maybe best to take her next
time or not try her on <the> tallest mountain in England for <the>
first try!

Although Article Drop is relatively infrequent in Daeninck’s corpus of

tweets (with only 6 per cent of articles in her corpus being dropped, by
my calculations), there are other written registers where it is much more
common. For example, Article Drop is pervasive (to the point of being the
norm) in a very different kind of abbreviated register widely referred to as
headlinese, found in the headlines of newspaper and magazine articles.
To illustrate this, I collected a sample of 390 front-page headlines in
British newspapers over a six-week period from the beginning of April to
mid-May in 2021. My sample showed a very high frequency of Article Drop
in obligatory contexts – that is, contexts where an article would be required
in unabbreviated registers (e.g. modifying an NP headed by a singular com-
mon noun which is countable). My sample contained 462 instances of Ar-
ticle Drop (290 of the-drop, 172 of a-drop); overall, articles were dropped
in 96 per cent of obligatory contexts (the same figure as reported for Dutch
newsflash headlines by Oosterhof & Rawoens 2017). There were frequently
multiple instances of (the same or different) articles being omitted within a
single sentence – as below:

(39) a. <The> Bank predicts <the> strongest year since <the> Second World
War as Britain surges out of lockdown (The Times, 7 May)
b. <A> Tory peer victimised <a> female worker at <a> firm (Daily
Telegraph, 14 April)
368 Abbreviated registers

c. <The> EU unveils <a> big push for <the> Pfizer vaccine (Financial
Times, 15 April)
d. <A> majority of Britons trust <the> Oxford jab, <a> poll shows (The
Times, 9 April)

Unlike Subject Drop or Auxiliary Drop, Article Drop is not restricted to

sentence-initial position, since (as we see from 38, 39 above) it can equally
affect nominals which are subjects, direct objects or prepositional objects.
Moreover, Article Drop is not restricted to root clauses, since it can also
affect (italicised) articles in non-root clauses like those bracketed below:

(40) a. <A> Court rules [Boris owes <a> £535 debt] (Daily Mirror, 13 May 21)
b. <The> rota will mean [<the> Queen is not alone] (Daily Mirror, 19 April)
c. Downing Street sources claim [<the> former chief adviser released
<the> PM’s messages out of spite] (Daily Telegraph, 23 April)
d. <The> Queen says [<the> Duke’s death ‘has left a huge void’ in her
life] (Daily Telegraph, 12 April)
e. Stocks retreat [after <a> US inflation surge sparks fears [<the>
economy is overheating]] (Financial Times, 13 May)
f. Starmer hits out [as <the> Tories reject <the> Labour call for <a>
Commons inquiry into lobbying] (Metro, 16 April)

Articles are omitted because they are words with little informational con-
tent or impact, and headlines aim to be as concise, informative and impact-
ful as possible.
Indeed, prescriptive manuals and style sheets that set out editing rules
for composing headlines explicitly instruct subeditors to omit articles.
For example, Williams (2013) outlines eight grammar rules for newspaper
headlines, one of which is: ‘Leave out articles (a, an, the).’ In the same vein,
Ghosh (2022) prescribes ‘Avoid the use of articles like a, an, the.’ Likewise,
the St Petersburg College guide on ‘How to write a news article: Headlines’
([Link]) specifies ‘Do not use articles – a, an, the.’ How-
ever, in the remainder of this module, I will argue that there are structural
conditions on Article Drop which undermine any such simplistic prescrip-
tive rule-based account. I will explore a number of possible syntactic ac-
counts, starting with a Truncation account.

7.3.1 A Truncation analysis of Article Drop

A suggestion which might at first sight seem plausible is that Article Drop
involves use of truncated nominals which contain no superordinate nom-
inal functional projections (like DP, ARTP or QP), but rather are truncated
NPs (i.e. NPs which don’t project any functional nominal superstructure
above NP). This would mean (for instance) that the complement of the
 7.3 Article Drop 369

verb reject in (40f) above is not the DP … ARTP structure in (41a) below
(containing ellipsed articles), but rather the NP … NP structure in (41b) (not
containing any articles):

(41) a. [DP <the> [NP Labour call for [ARTP <a> [NP Commons inquiry into
lobbying]]]]
b. [NP Labour call for [NP Commons inquiry into lobbying]]

However, evidence against positing that all nominals in headlines are trun-
cated NPs with no functional superstructure of any kind comes from the
observation that many NPs in headlines (such as those underlined below)
are modified by quantifiers and numerals like those bold-printed below:

(42) a. Civil service staff <are> ordered to reveal all second jobs (The Times,
15 April)
b. Minnesota <is> hit by more unrest (Financial Times, 14 April)
c. Travel abroad <is> to become legal, but tests and quarantine will
remain for most destinations (Daily Telegraph, 1 May)
d. Free kits <are> to be available for every adult in England (Daily
Mirror, 5 April)
e. No end <is> in sight as Johnson says normal is some way off (Daily
Telegraph, 6 April)
f. Only 14 breaches of lobbying rules <are> found in two years (The i,
16 April)

On the assumption that the highlighted nominals in (42) have the status of
QP or NUMP (i.e. quantifier or numeral phrases), it is clearly untenable to
maintain that all layers of functional superstructure above NP are truncated
in headlinese.
But could we instead maintain that nominals in headlinese show trunca-
tion of the highest (DP) layer of structure? This, too, seems to be too strong
a claim, since possessive/genitive nominals like those underlined below are
frequently found:

(43) a. Northern Ireland’s First Minister <is> toppled in <a> backlash over
Brexit (The i, 29 April)
b. Bozo’s phone number <was> on the internet for 15 years (Daily Star,
1 May)
c. Johnson’s Mustique holiday <is> added to inquiries into potential
rule breaches (Financial Times, 11 May)
d. EY’s Wirecard audits <are> slammed in <a> report (Financial Times,
19 April)
e. Starmer’s leadership <is> in crisis over <the> reshuffle move to
demote Rayner (The Guardian, 10 May)
370 Abbreviated registers

Moreover, (underlined) nominals with an (italicised) pronominal possessive

modifier are also frequent in newspaper headlines:

(44) a. <The> young <are> turning their backs on learning to drive (Daily
Telegraph, 26 April)
b. <The> Duke planned his funeral with military precision (Daily
Telegraph, 16 April)
c. Cameron <is> told: your reputation is in tatters after Greensill
lobbying (The Guardian, 14 May)
d. Harry <is> set to miss her 95th birthday (Daily Mirror, 19 April)
e. <There are> 10 days to save my life (Sunday People, 2 May 21)

Under the analysis in Radford (2020: 201) nominal and pronominal posses-
sives like your car/the queen’s car are DPs with the skeletal structure below
(with the possessive affix being spelled out as -r on your, and as ’s on the
queen’s):

(45) DP

D′

DP D NP
you/the queen AfPOSS car

If we adopt an analysis along similar lines, the frequent use of nominal/

pronominal possessive structures like (43/44) in headlines undermines the
claim that nominals in headlinese involve DP truncation.
Instead, what seem to be systematically truncated are the articles the/a.
Interestingly, Article Drop even occurs inside possessive DPs in structures
like the following (where the bold-printed article in the specifier position of
the DP in square brackets is truncated):

(46) a. <The> Queen says [<the> Duke’s death] ‘has left a huge void’ in her
life (Daily Telegraph, 12 April)
b. [<The> Queen’s final preparations] aim to steer <the> Royal family
into calmer waters (Daily Telegraph, 16 April)
c. Downing Street sources claim <the> former chief adviser released
[<the> PM’s messages] out of spite (Daily Telegraph, 23 April)
d. [<The> BBC’s Philip tributes] set <a> record for complaints (The
Guardian, 13 April)
e. [<The> Alawi sect’s faith in Assad] <is> knocked by Syrian crises
(Financial Times, 14 April)
 7.3 Article Drop 371

In all thirteen such structures in my sample of headlines, the article was

always deleted (i.e. my sample contained no DPs like the Duke’s death).
The bracketed possessive DP constituents in (46) have clearly not been
truncated, and retain the affix ’s which (on one view) would be the head
D of the DP. If we were to adopt a Truncation approach, we would have to
say that the specifier of the overall DP is truncated above NP, so that the
bracketed DP in (46a) above has the structure below:

(47) DP

NP D′
Duke
D NP
’s death

This would mean that the bold-printed specifier in (47) would not be the
DP constituent required in non-abbreviated registers, but rather a trun-
cated NP.
However, there are several reasons to be sceptical about the Truncation
analysis of the specifier Duke in (47). For one thing, Truncation typically
affects constituents on the left edge of a root projection; and yet (as we
have seen) articles are omitted in a wide range of positions within head-
lines. Secondly, it is potentially inconsistent to posit that a nominal is trun-
cated to NP when (as in 47) it is contained inside an untruncated nominal
which projects all the way up to DP. Thirdly, the Truncation account fails to
account for why the NP Duke receives the same interpretation as a definite
DP like ‘the Duke’, and not an indefinite interpretation like ‘a Duke’. More-
over, if we take Duke to be an unaccusative argument of the noun death,
(47) also violates the DP Hypothesis argued for by Longobardi (1994, 1996,
2001) whereby definite arguments are required to be DPs. In addition, if
articles are the locus of the number, person and case properties of nominals
(with these features being transmitted to the NP below them via Concord),
it could be argued that a truncated NP cannot receive number or case,
so causing the derivation to crash. Moreover, absence of the article (and
its φ-features) would then pose problems in accounting for subject-verb
agreement in a (constructed) headline like Sheep runs amok.

7.3.2 Null article and article ellipsis accounts

One way of overcoming the objections to the Truncation approach outlined
in §7.3.1 would be to suppose that the abbreviated nominal Duke in (46a)
is an untruncated DP headed by a null article, so that the DP Duke’s death
has the structure below:
372 Abbreviated registers

(48) DP

DP D′

D NP D NP
ø Duke ’s death

Such a structure is preferable to the NP analysis (47), firstly because it

makes for structural continuity with unabbreviated registers (by treating
the specifier Duke as a DP), and secondly because it conforms to the DP
Hypothesis (in that the argument of the unaccusative noun death is the
DP ø Duke). Under the analysis in (48), Article Drop in headlines is not the
result of structural Truncation (i.e. non-projection of DP/ARTP), but rather
the result of this register having a null article which is used in place of the
overt articles the/a. (For an analysis positing null articles in headlines in
German, see Lemke et al. 2017, and Reich 2017.)
However, the null article hypothesis raises the question of whether it
is plausible to suppose that a register like headlinese could have devel-
oped a functor of its own (a null article) which is not found in unab-
breviated (or spoken) registers of the same language. Moreover, it poses
the learnability question of how such a null article could be acquired.
Certainly not via the normal process of L1 acquisition, since infants
acquiring their L1 can’t read newspaper headlines. Nor via the normal
process of L2 acquisition, since while transfer effects might lead an L2
learner whose L1 lacks articles to omit them, there are no such transfer
effects at work in headlinese: on the contrary, headline writers omit art-
icles which are obligatory in non-abbreviated registers of their native
language.
A more plausible analysis would arguably be to suppose that the abbre-
viated nominal Duke in (46a) has the same syntactic structure (49) below as
its untruncated counterpart the Duke’s death:

(49) DP2

DP1 D′

D NP D NP
the Duke ’s death

That is to say, the definite article the would be present in the syntax, but
(in the headlinese register) would undergo a type of ellipsis in the spellout
component which can be characterised informally as follows:

(50) Article Drop

An article which is the first word in the functional projection that it
heads can be dropped
 7.3 Article Drop 373

Saying here that an article can be ‘dropped’ means that an article present in
the syntax can be given a null spellout/null printout. Since the article the in
(49) is the first word of the DP1 projection that it heads, it can be dropped,
as in (46a) – and a parallel analysis can be proposed for (46b–46e).
The ellipsis account of Article Drop outlined in (50) above predicts that
articles can only be dropped if they are the first word in a DP/ARTP con-
taining them. This is true of every single one of the dropped articles in my
own sample of headlines. But what about the following headlines (the first
being internet-sourced, and the second being a spoof headline constructed
by me), where the dropped article is not initial within the italicised nomi-
nals containing it?

(51) a. Boris Johnson’s 2020: A tumultuous year that upturned all <the>
PM’s hopes (The Independent, 29 December 2020)
b. Only <the> PM’s allies stand by him

If we suppose that the italicised nominals are QPs headed by the quanti-
fiers all/only, the italicised strings in (51a/51b) will have the structure in
(52a/52b):

(52) QP

DP2

DP1 D′

QP D NP D NP
a. all the PM ’s hopes
b. only the PM ’s allies
This means that although the article the is not initial within (i.e. is not the
first word of) the QP (since it follows the Q all/only), it is initial within the
DP1 projection that it heads and thus can undergo the Article Drop oper-
ation in (50).
The condition that only an initial article can be dropped also accounts
for my intuition that the (bold-printed) articles cannot be deleted in (con-
structed) headlines such as the following (an intuition shared by Andrew
Weir, pc):

(53) a. MEPs earn [twice the salary of MPs], report reveals

b. [What the hell] is BoJo up to?
c. [What a nasty piece of work] PM’s aide turns out to be!
d. Cummings creates [quite a stir]

If the bracketed nominals are DPs/ARTPs headed by the/a, it follows from

the account of Article Drop in (50) that the bold-printed articles cannot
374 Abbreviated registers

be dropped in (53) because they are not initial within the (bracketed) DP/
ARTP that they head, being preceded by the italicised items twice/what/
quite (which, on one view, are specifiers of the articles).
The account of Article Drop outlined above (whereby it involves not
pronouncing/printing articles which are present in the syntax) raises a
number of questions in its wake, however. One (raised by an anonymous
reviewer) is why an article can only be omitted when initial within its
projection, and not when medial as in structures like (53). A plausible
answer is to suppose that Article Drop is a subtype of Left Edge Ellipsis
which involves giving a null spellout to the edge of a DP or ARTP (where
the edge comprises the head and any specifier that it has). Where the left
edge of the DP/ARTP comprises an article on its own, the Recoverability
Condition of Chomsky (1964b) and Katz & Postal (1964) allows deletion
of the article, because it is a functor with minimal semantic content.
But where (as in the nominals bracketed in 53 above), the article has a
specifier, the Recoverability Condition precludes deletion of the edge of
the DP/ARTP because the specifier has substantive lexical semantic con-
tent. In more concrete terms, this means that in the DP twice the salary
of MPs in (53a), the edge of the DP (= the string twice the) cannot be
deleted because twice has semantic content which cannot be recovered
if it is deleted.
Another question arising from the analysis in (50) is why, as observed by
Mårdh (1980: 153), an article can’t be deleted in a (constructed) headline
such as:

(54) Motorway speed limit to be reduced to fifty miles an hour in bad

weather

A plausible answer is that the Recoverability Condition makes the article

an undroppable in (54) because it functions as a distributive quantifier,
and recoverability considerations prevent quantifiers from being dropped
(because they are too contentful).
An interesting restriction on Article Drop observed in much work on
headlines is that an article cannot be dropped from a direct object if the
subject of the sentence contains an article. Stowell (1999) captured this
generalisation in terms of a constraint which can be paraphrased informally
as follows:

(55) C‑command Condition on Article Drop

No nominal allows Article Drop if it is c‑commanded by a nominal with
an overt article
 7.4 Be Drop 375

How this constraint works can be illustrated in terms of the following set of
hypothetical headlines, where ø denotes a dropped article:

(56) a. A man bites a dog

b. ø man bites a dog
c. ø man bites ø dog
d. *A man bites ø dog

In none of these headlines does the subject a/ø man violate the constraint,
since the subject has no nominal constituent c‑commanding it. Nor does the
object a dog violate the constraint in (56a, 56b), because it has not under-
gone Article Drop. And although the object ø dog in (56c) has undergone
Article Drop, it is c‑commanded by a subject ø man which has also
undergone Article Drop, so there is again no violation of the C‑command
Condition in (55). By contrast, the article-dropped object ø dog in (56d)
does indeed violate the constraint, since it is c‑commanded by the article-
containing subject a man.
To summarise: in Module 7.3 we have looked at Article Drop, which is
sporadic in some registers (e.g. social media like Twitter), but systematic
in others (e.g. newspaper headlines). I argued that this involves a definite
or indefinite article which is present in the syntax not being pronounced/
printed if it is the first word of the projection that it heads.
You should now be able to tackle Exercise 7.3.

7.4 B e Drop

In this module, I turn to look at a further phenomenon characteristic of

the headlinese register, which I will term e Drop: by this, I mean omitting
finite (indicative) forms of b e in obligatory contexts (i.e. in contexts where
adults would require the use of such a form of b e in untruncated registers).
Prescriptive guidelines on writing headlines often explicitly instruct sub-
editors to omit (finite) forms of b e . For example, Williams (2013) specifies
‘Leave out to be’, advising subeditors to say ‘Residents unhappy about new
road’ rather than ‘Residents are unhappy … .’ Likewise, Ghosh (2022) recom-
mends ‘Avoid the use of is, are, was, to be, etc.’ And in the same spirit, the
Headline Writing Guidelines on [Link] (2020) instruct head-
line writers to ‘Omit forms of the verb to be (is, are, was, were).’ However,
in this module, I will argue that any such prescriptive rule-based account is
oversimplistic (in that it overlooks structural conditions on e Drop), and I
present (and evaluate) alternative syntactic accounts.
376 Abbreviated registers

7.4.1 Characteristics of Be Drop

My sample of newspaper headlines contains 192 cases of e Drop, 79
involving passive b e as in (57a) below, 64 involving copular b e as in (57b),
39 involving the future b e t o structure as in (57c), and 10 involving pro-
gressive b e as in (57d):

(57) a. Minnesota <is> hit by more unrest (Financial Times 14 April)

b. Labour <is> in turmoil after Tories inflict huge defeats (The
Guardian, 8 May)
c. Europe <is> to welcome Brits (The Sun, 22 April)
d. Glaciers <are> melting twice as fast as 20 years ago (The Guardian,
29 April)

e Drop is not restricted to root clauses, but is also found in subordinate

clauses like those bracketed below (where the relevant instance of Be Drop
is bold-printed):

(58) a. <A> corrupt police officer <is> caught [when <a> gangster group
<is> hacked] (The Times, 13 May)
b. Angela Rayner <is> ‘fired’ [as Labour <is> gripped by post-poll
rancour] (The Observer, 9 May)
c. High streets feel the rush [as lockdown curbs <are> eased] (The
Guardian, 13 April)
d. Holiday sun <is> on <the> horizon [as restrictions <are> set to ease
in June] (The Times, 9 April)

e Drop is systematic, in that finite forms of b e are dropped in 97 per cent

of obligatory contexts in my sample of headlines. Exceptional cases where
(bold-printed) b e is not dropped fall into the types illustrated below:

(59) a. Bozo is a buffoon says <a> Tory (Daily Star, 4 April)

b. Cameron <is> told: your reputation is in tatters after Greensill
lobbying (The Guardian, 14 May)
c. After <a> week of sun, it’s <a> bank hol chiller … just as hot tubs
boom (Daily Star, 3 April)
d. Cummings <is> warned he isn’t in <the> clear over <the> Chatty Rat
leak (Daily Telegraph, 26 April)

In cases like (59a, 59b), the italicised clause is a direct quotation (albeit
not enclosed in inverted commas), and if the speech is quoted verbatim,
it will retain the copula is from the original. (59c, 59d) could fall under a
constraint (discussed below) that personal pronoun subjects don’t allow e
Drop, or alternatively could fall under a different condition, to the effect
that be can’t be dropped on its own when part of a clitic cluster.
 7.4 Be Drop 377

Since indicative forms of b e function as auxiliaries, and since we saw in

Module 7.2 that the other two non-modal auxiliaries (h av e / d o ) can under-
go left edge ellipsis, we might expect to find that h av e and d o can also
undergo ellipsis in headlines. However, my sample contains no examples
of ave / o Drop; on the contrary, finite forms of h av e and d o are overtly
spelled out, as in the examples below:

(60) a. Seven have died from blood clots after 18 m AstraZeneca jabs,
watchdog says (Financial Times Weekend, 3 April)
b. Whitehall dismisses claim former aide has been cleared (The Times,
26 April)
c. Queen says Duke’s death ‘has left a huge void’ in her life (Daily
Telegraph, 12 April)
d. Jez: Life of crime does pay (Daily Star, 11 May)

Hence, I have employed the narrower term e Drop here, rather than the
more general term Aux Drop.
An interesting constraint which appears to hold in my sample of head-
lines is that in all cases of e Drop, the subject of b e is always a nominal,
never a personal pronoun. This is puzzling, in view of the fact that a clause
(like that bracketed in 61a below) in which b e undergoes Gapping allows
a (bold-printed) personal pronoun subject – as does a clause like (61b) in
African American Vernacular English, where a null copula/ø is used in a
context where standard English would use the contracted copula’s:

(61) a. If you were me and [I <were> you] – would you believe it, even if it
were true? (J. L. Chalker, The Return of Nathan Brazil, Hachette UK,
2013)
b. He ø fast in everything he do (Labov 1969: 717)

In my sample of headlines, (bold-printed) nominative personal pronouns only

occur as subjects of (italicised) overt finite auxiliaries or finite verbs, as in:

(62) a. <A> Senior Tory says Johnson should quit if he broke donation rules
(The Guardian, 3 May)
b. Harry: I left <the> UK to break <the> cycle of pain (The Times, 14 May)
c. 14 years on, <the> McCanns tell <their> missing daughter: We are
waiting for you (Daily Mirror, 13 May)
d. On her 18th birthday … Madeleine parents say they’ll never give up
(Daily Express, 13 May)
e. Cummings <is> warned he isn’t in <the> clear over <the> Chatty Rat
leak (Daily Telegraph, 26 April)
f. Her majesty <is> to spend more time at Windsor Castle where she is
‘most comfortable’ (Daily Express, 14 April)
378 Abbreviated registers

So how can we account for e Drop sentences not allowing pronominal

subjects?
One possible answer is a pragmatic one: b e is dropped in order to avoid
the use of items with minimal semantic content and impact; and (as sug-
gested by Mike Jones, pc) subject pronouns are not used because subjects
in headlines announce new topics, and pronouns are not suited to this use.
However, the assumption that b e is dropped purely because it has minimal
content/impact fails to account for an important aspect of its distribution –
namely that e Drop is incompatible with Wh‑Movement (more generally,
with constituents undergoing A‑bar Movement into peripheral positions)
– as we see from the observation that my data don’t contain any e Drop
headlines such as the following (and my intuition is that these are not well-
formed as headlines):

(63) a. *What on earth <is> Bozo up to?

b. *What chaos Bozo <is> causing!

The absence of headlines like (63) suggests an alternative syntactic approach

under which e Drop headlines are truncated clauses with a truncated
periphery.
In the next section, I explore possible ways of implementing such an
approach.

7.4.2 Truncation accounts of B e Drop

One potential implementation of the Truncation analysis would be to sup-
pose that e Drop clauses have a truncated structure which projects only
up to SUBJP. On one variant of this idea, (57a) Minnesota hit by more
unrest would have a truncated structure along the lines below (simplified
by not showing the internal structure of the VOICEP in which the passive
auxiliary is originates before raising through T into SUBJ):

(64) SUBJP

DP SUBJ′
Minnesota
SUBJ T′
<is>
T VOICEP
is hit by more unrest
We can then suppose that (because of its minimal semantic content) b e
can receive a <null spellout> in headlinese when it is the head of a root
projection. Still, this analysis would have to be modified in some way
in order to account for the observation made earlier that e Drop is not
restricted to root clauses, but rather can also apply to subordinate clauses
 7.4 Be Drop 379

like those bracketed in (58) above. Perhaps, then, we could suppose that e
Drop applies in truncated clauses (i.e. clauses which contain no peripheral
FORCEP … FINP projections).
Still, such an approach raises the question of which clauses can (and can’t)
be truncated (and why), and what the mechanism is by which be comes to
receive a null spellout. Under the analysis in (64), this can’t be Left Edge
Ellipsis, since this gives a null spellout to all constituents on the left edge of
the highest projection in a clause – and this would lead us to expect that be
can only have a null spellout if its subject has a null spellout as well. But
this is decidedly not the case: all e Drop sentences in my data have overt
subjects (and indeed in 64, the subject Minnesota receives an overt spellout).
One potential way of addressing this issue might be to suppose that b e
raises to the head of SUBJP, whereas the subject (in the syntax, at least)
raises only as far as spec‑TP (an idea familiar from §6.3.1). This would
mean that in place of (64) above, we have the revised truncated structure
in (65) below:

(65) SUBJP

SUBJ TP
is
DP T′
Minnesota
T VOICEP
is is hit by more unrest

In the PF component, is will receive a null spellout in SUBJ via the condi-
tion in (12ii) above that the left edge of the highest projection in a truncated
clause can be unpronounced in the PF component. The overall structure
will thus be spelled out as Minnesota hit by more unrest. The structure in
(65) will be interpreted as declarative in force by the default Clause Typing
Condition (22ii).
However, two important questions raise doubts about the plausibility of
the derivation in (65). One is why the subject does not raise to the criter-
ial position for subjects (spec‑SUBJP). One possibility is that SUBJ in (65)
carries a defective set of agreement features (e.g. perhaps number but not
person), and this is enough for SUBJ to attract b e , but not enough for it to
trigger raising of the subject to spec‑SUBJP in the syntax (if subject raising
is contingent on the presence of person). Still, this raises the further ques-
tion of what interpretation SUBJP could receive in the LF component if it
does not house the subject: one answer could be that the subject raises to
spec‑SUBJP in the semantic/LF component.
A second question raised by the derivation in (65) – and also by that in
(64) – is why the subject in e Drop sentences can’t be nominative. A plausible
380 Abbreviated registers

answer is that if FIN is the locus of nominative case (in the sense that a finite
FIN head assigns nominative case to subjects, as claimed in §4.3.2), truncating
clauses at SUBJP would account for the absence of nominative pronouns like
she as subjects in e Drop headlines. If be lacks one or more agreement fea-
tures, it follows that unless it undergoes ellipsis, the derivation will crash at PF
(because English has no spellout for a present tense form of be with defective
agreement features). On this view, ellipsis of be serves to rescue a structure
which would otherwise crash, since English has no overt spellout for a tensed
verb with a defective set of agreement features. This would be in line with
research dating back to Ross (1969) and Chomsky (1972) arguing that ellipsis
operations can rescue derivations which would otherwise crash.
Since the subject is not within the domain of any case assigner (irrespec-
tive of whether we adopt the implementation of the SUBJP analysis in 65,
or the variant of it in 64), we can suppose that the subject receives default
accusative case via a Default Case Assignment Rule which was given the
following informal characterisation in §4.1.2:

(66) Default Case Assignment

A (pro)nominal which does not fall within the domain of (i.e. which
is not c‑commanded by) a local (i.e. nearby) case assigner receives
accusative case by default

We could then suppose that in a (constructed) headline like (67) below, the
italicised subject receives accusative case by default, via the case assign-
ment condition (66) above:

(67) Us poor Brits <are> victimised by Eurocrats … again!

However, what remains to be accounted for under this story (as pointed out by
Andrew Weir, pc) is why we don’t find ?*Us victimised again! (albeit this is not
as bad as *We victimised again). One possibility would be to invoke the prag-
matic analysis offered by Mike Jones (pc) and mentioned earlier in §7.4.1, and
suppose that subject personal pronouns are not used because subjects in head-
lines announce new topics, and personal pronouns are not suited to this use.
The assumption in (65) that e Drop headlines are truncated clauses
with defective agreement and defective subjects which raise only as far as
spec‑TP invites comparison with a phenomenon in Belfast English which
Henry (1995) refers to as ‘singular concord’. This involves s-forms of aux-
iliaries/verbs being used to spell out items which are specified for present
tense but not for agreement. Consequently, singular concord clauses allow
plural subjects as in:

(68) a. The eggs is cracked (Henry 1995: 16)

b. These cars goes very fast (Henry 1995: 16)
 7.4 Be Drop 381

However, singular concord is incompatible with nominative subjects:

(69) a. *They is cracked (Henry 1995: 16)

b. *They goes very fast (Henry 1995: 16)

Furthermore, it is also incompatible with Auxiliary Inversion:

(70) *Is the eggs cracked? (Henry 1995: 16)

Alison Henry argues that in singular concord clauses, the subject ‘moves
only as far as SPEC/Tense’ (Henry 1995: 17). Auxiliary Inversion is blocked
in sentences like (70) because locality constraints require is to move through
SUBJ into FIN, but is cannot move to SUBJ because it has no agreement
features. In addition, subjects in singular concord clauses carry default
(accusative) case:

(71) Us and them/*We and they is always arguing (Henry 1995: 18)

Thus, the potential parallels between e Drop in headlines and singular

concord in Belfast English are suggestive.
Returning now to headlines, it should be noted that something which is
potentially problematic under the Left Edge Ellipsis account outlined in (65)
above is why only b e can be dropped in headlines, not other auxiliaries.
After all, structures involving Left Edge Ellipsis allow ellipsis not only of
be , but also of a range of other auxiliaries – as illustrated by the examples
of Aux Drop below:

(72) a. <Are> you not winning the game? (Schirer 2008: 37)
b. <Does> anyone want a cup of tea? (Rühlemann 2012: 5)
c. <Have> you seen any good movies lately? (Zwicky & Zwicky 1981:
536)
d. <Will> tomorrow at four be okay? (after Carter & McCarthy 1995:
147)
e. <Would> you like to go fishing this afternoon?

This asymmetry between the range of auxiliaries which can be dropped

in Aux Drop structures like (72) on the one hand (where b e , h av e , d o and
modals like wi and wo u d can be dropped) and headlines like (57) on
the other (where only b e can be dropped) poses a substantial challenge to
the assumption that headlines undergo Left Edge Ellipsis. In terms of the
analysis in (65), this amounts to asking why only b e can raise to SUBJ
(and thereby get into a position on the edge of a root projection where it
can undergo Left Edge Ellipsis), and not other auxiliaries. Perhaps this is
a reflex of the property that b e is more mobile than other verbs (being the
only verb that can raise out of VP in most varieties of English), and has a
wider range of agreeing forms (am/are/is/was/were) than other verbs – and
382 Abbreviated registers

indeed some earlier work suggested that verbs with a rich set of overt
agreement forms raise to a high position within clauses (e.g. Vikner 1995;
Rohrbacher 1999). Still, this seems an implausible rationale if we are posit-
ing that SUBJ carries a defective set of agreement features in (65).
A further asymmetry is that whereas Aux Drop occurs only in main/root
clauses like (72), e Drop is found in root and non-root clauses like – for
example, in subordinate clauses like the when/as-clauses in (58) above. An
additional potential problem that arises with the analysis in (65) is that
it is generally the case that an auxiliary only raises to SUBJ if its subject
raises to spec‑SUBJP, and this condition is violated in (65). Overall, then, a
range of considerations argue against treating e Drop clauses as truncat-
ed SUBJP structures with a SUBJ head with defective agreement features
(perhaps carrying number but not person) which attracts b e (but not its
subject) to raise to the edge of SUBJP, as in (65) above.
Accordingly, let’s explore an alternative variant of the Truncation ac-
count of e Drop which supposes that it is tense (not agreement) which is
defective in e Drop clauses (let’s call this the Tense Deficit Hypothesis).
This seems plausible, since Stowell (1999) reported that it is a characteristic
of headlines that they often use present tense to report past events, as in
the following example:

(73) Minister of Defence White dies at 43 (de Lange 2008: 70)

It would seem that present tense serves as a default tense value in such cases.
One possible implementation of the Tense Deficit Hypothesis is to sup-
pose that e Drop clauses are truncated SUBJP+TP+VP structures in which
T contains an affix which is unspecified for tense [u-Tns] in the syntax, and
is assigned a discourse-determined tense value at LF, which will be [Pres-
Tns] in the case of (57a) Minnesota hit by more unrest (though this would
require us to abandon the Legibility Condition of §3.1.1, requiring all fea-
tures to be valued at the end of the syntactic derivation). On this view, (57a)
would be a truncated clause with the syntactic structure below, if b e raises
to adjoin to the affix in T (with the structure being simplified by not show-
ing the internal structure of VP):

(74) SUBJP

DP SUBJ′
Minnesota
SUBJ TP
ø
T VOICEP
BE+Af
VOICE VP
BE hit by more unrest
 7.4 Be Drop 383

The resulting be +Af head cannot be spelled out as an overt form of b e

(since there is no overt spellout for forms of b e with an unvalued tense
feature). Instead let’s suppose that it receives a null spellout at PF, because
be has a null counterpart ø which is used as a default spellout in contexts
where no overt form of b e is available – including in small clause structures
like that bracketed below

(75) I consider [this analysis ø potentially flawed]

We could then suppose that the b e +Af head in (74) receives the same default
null spellout ø as is found in small clauses like that bracketed in (75) – and
perhaps also in root clauses like that italicised in ‘What? John ø in prison?
No way!’

7.4.3 B e Drop and Article Drop

I now turn to touch on another characteristic of e Drop headlines which
appears to be empirically robust – namely that headlines which show e
Drop typically show Article Drop too, as illustrated by the examples below:

(76) a. <A> wife <is> fighting for life after <a> knife horror at <a> 2m
country estate (Daily Express, 9 April)
b. Starmer <is> ejected from <a> pub by <an> angry landlord (Daily
Telegraph, 20 April)
c. Cummings <is> to blame <the> PM for <the> Covid death toll
(Sunday Times, 25 April)
d. <The> under-40s <are> to be offered <an> alternative to <the>
Oxford jab (The Independent, 7 May)
e. <An> embryo <is> made from <a> monkey and <a> human (The
Times, 16 April)
f. <The> UK <is> to use <the> presidency of <the> G7 to take on China
(Daily Telegraph, 3 April)
g. <The> PM <is> <the> subject of <a> £535 county court order (Daily
Telegraph, 13 May)

In my sample, 100 per cent of all subjects in e Drop headlines showed

Article Drop in obligatory contexts for articles, and 91 per cent of all
non-subject constituents. Exceptions to this generalisation (i.e. e Drop
clauses containing highlighted non-subject nominals with overt articles)
fall into the types illustrated below:

(77) a. India <is> braced for 500,000 infections a day (Sunday Times, 25 April)
b. Boris <is> painted into a corner (Daily Mail, 29 April)
c. Boris <is> on the ropes (Daily Mail, 27 April)
d. <A> Covid pill to take at home <is> on the way (Daily Express,
21 April)
384 Abbreviated registers

e. Britain <is> on the offensive in <a> race with China for crucial
minerals (The Times, 5 April)
f. Cars and whisky <are> to the fore as Johnson chases <a>
‘challenging’ India trade deal (Financial Times, 16 April)

More specifically, they fall into two categories. In (77a), the article a has a dis-
tributive sense (paraphrasable as ‘each’), and it has often been noted (in work
dating back to Mårdh 1980) that articles which function as generic or distribu-
tive quantifiers can’t be dropped (arguably for the same recoverability reasons
that quantifiers like each/every/any can’t be dropped). The remaining articles
bold-printed in (77b–77f) are intrinsic parts of idioms/set phrases (paint into
a corner, on the ropes, on the way, on the offensive, to the fore), and these are
widely reported to be resistant to Article Drop. Still, it should be noted that my
sample also contains some potential fixed phrases with dropped articles like
‘on <the> brink of ’, ‘in <the> wake of ’, ‘in <the> pipeline’, and ‘on <the> way’;
thus, it would appear that such phrases allow (but do not require) Article Drop.
The question raised by the discussion above is what is the nature of
the correlation between e Drop and Article Drop. An intriguing question
(raised by Giuliano Armenante, pc) is whether there could be some relation
between Article Drop and e Drop in headlines on the one hand, and the
cross-linguistic observation made by Bošković (2012) that languages with-
out articles typically also lack TPs on the other. For Bošković, nominals in
languages without articles project only up to NP (not up to DP), and occur
in clauses which lack TP and which also have subjects assigned default
case. If personal pronouns like him/her/it/them are pronominal DPs, such
an analysis would account for the absence of personal pronoun subjects in
e Drop sentences (if the latter have NP subjects).
Still, a potential challenge to the claim that the subjects of e Drop sen-
tences do not allow DP subjects comes from the (spoof) headline in (67)
above, repeated as (78) below:

(78) Us poor Brits <are> victimised by Eurocrats … again!

The problem which (78) poses is that it is a e Drop sentence which has the
seeming DP us poor Brits as its subject. However, it should be noted that this
type of subject was not attested in my sample of e Drop headlines (though
of course this could have been an artefact of the relatively small sample size).
Another potential challenge to the claim that e Drop sentences don’t al-
low DP subjects is posed by sentences like (43) above, repeated as (79) below:

(79) a. Northern Ireland’s First Minister <is> toppled in <a> backlash over
Brexit (The i, 29 April)
b. Bozo’s phone number <was> on the internet for 15 years (Daily Star,
1 May)
 7.4 Be Drop 385

c. Johnson’s Mustique holiday <is> added to inquiries into potential

rule breaches (Financial Times, 11 May)
d. EY’s Wirecard audits <are> slammed in <a> report (Financial Times,
19 April)
e. Starmer’s leadership <is> in crisis over <the> reshuffle move to
demote Rayner (The Guardian, 10 May)

Here, the underlined subjects of the e Drop sentences in question would

appear to be DPs – at least if we adopt the DP analysis of possessive nom-
inals in (45) above. Still, one way round this might be to follow Kayne
(1994: 26) in supposing that possessors are not in spec‑DP, but rather in a
lower projection headed by the possessive affix ’s (let’s call this POSSP).
On this view, we could suppose that the underlined constituents in (79) are
defective (DP-less) nominals which project only up to POSSP, not up to DP.
In more concrete terms, this would mean that the nominal Starmer’s lead-
ership in (79e) has the structure below:

(80) POSSP

NP POSS′
Starmer
POSS NP
’s leadership

We could then maintain the view that nominals in e Drop sentences lack
the DP projection found in non-abbreviated registers.
Bošković (2012: 217) suggests that TP-less clauses contain a subject pro-
jection (which I will here take to be SUBJP), but no tense projection. If so,
and if e Drop clauses are truncated at SUBJP (as suggested in §7.4.2), a
headline such as (76a) above could have a derivation along the simplified
lines shown below:

(81) SUBJP

NP SUBJ′
wife
SUBJ PROGP
ø
PROG VP
<BE> wife fighting for life after knife
horror at 2m country estate

On this view, the clause subject (wife) is an NP which raises from spec‑VP
to spec‑SUBJP, and receives default accusative case via (66). The progres-
sive auxiliary be cannot be assigned tense (because there is no TP projec-
tion), and will remain in PROG if SUBJ can only attract a tensed T‑head
386 Abbreviated registers

to adjoin to it. If SUBJ contains an agreement affix which is lowered onto

be at PF, b e will end up marked for agreement but not tense. But since PF
has no overt spellout for any (auxiliary or non-auxiliary) verb inflected for
agreement but not tense, b e will be given a null spellout by default, and
the overall structure in (81) is spelled out as Wife fighting for life after knife
horror at 2m country estate.
An interesting technical issue arising from the ‘no TP, no DP’ derivation
in (81) concerns how the subject NP wife raises to spec‑SUBJP. Under the ac-
count of movement of subjects to spec‑TP within the CP+TP+VP framework
in §3.1.1, a T probe is said to attract the closest XP goal carrying a person
feature to become its specifier (provided that the goal is active). If we trans-
pose this account into the SUBJP analysis of subjects assumed here, this
would mean that SUBJ attracts the closest XP carrying person to become its
specifier. But if D is the locus of the person features of nominals, then SUBJ
will be unable to attract the NP wife in (81) to become its specifier, because
this NP is not projected further into DP, and so lacks person. One way round
this would be to suppose that SUBJ attracts the closest XP that it agrees with
in respect of one of more φ-features to become its specifier. We could then
say that SUBJ in (81) carries a number feature which allows it to attract the
NP wife to move to spec‑SUBJP because the NP wife is specified as singular
in number, and is active by virtue of its unvalued case feature.
Indeed, we might even go further and suppose that SUBJ in e Drop
headlines is a defective head which enters the derivation carrying an un-
valued number feature, and attracts the closest XP carrying number to raise
to spec‑SUBJP: in the case of (81), this XP will be the NP wife. If SUBJ in
e Drop structures lacks a person feature, and if SUBJ can only attract an
XP with person if SUBJ carries a person feature, it follows that the defective
SUBJ in e Drop structures (by virtue of carrying number but not person)
cannot attract a DP carrying person to become its subject. This would pro-
vide a principled account of the observation reported above that e Drop
sentences don’t have DP subjects (like the queen or she).
Under the approach outlined above, there are potential parallels between
languages which lack DP and TP projections on the one hand, and head-
lines which exhibit e Drop and Article Drop on the other. Still, a question
arising from this approach concerns the exact nature of the inter-relation
between DP and TP. Bošković (2012: 216) argues that TP is the clausal
counterpart of the nominal constituent DP, so that ‘the lack of DP in a lan-
guage would imply the lack of its clausal counterpart, namely TP’ (2012:
216) and hence ‘article-less languages lack TP’ (2012: 221). But why is there
a correlation between articles and tense?
One possible answer (suggested by Giuliano Armenante, pc) can be of-
fered in relation to work by Schwarz (2012) on the temporal properties of
 7.4 Be Drop 387

DPs. I will present the key ideas in a highly simplified form here, for the
sake of readers unfamiliar with work on formal semantics. In this con-
nection, consider the interpretation of the articles a/the in the following
examples:

(82) a. A/the hostage is bleeding after being pistol-whipped

b. A/the hostage is safe after being released

In (82a), the phrase a/the hostage means ‘a/the person who is a hostage’,
whereas in (82b) it means ‘a/the person who was a hostage’ (but no longer
is after being released). One way of accounting for this is to suppose that
articles carry a tense which can either be bound by (i.e. interpreted as the
same as) that of the finite verb/auxiliary in the containing clause (as in
82a), or can be contextually determined (as in 82b). This opens up the pos-
sibility that it is a tense deficit in the article which is at the root of Article
Drop. Giuliano makes the following suggestion about tense-dropped (i.e.
e Drop) headlines:

Tense-dropped headlines are not situated in a salient context (since a temporal

reference is missing). This might explain why the article is dropped (It can’t
be bound, since its binder is dropped, and it can’t be contextually evaluated
either.) (Giuliano Armenante, pc)

The phrase ‘It can’t be bound, since its binder is dropped’ can be para-
phrased informally as ‘It can’t be interpreted as having the same tense as
be , since be is tenseless.’ One way of implementing this idea within the
framework used here is to suppose that overt articles carry a tense feature
which needs to be valued, and that in e Drop clauses, this feature can’t be
valued by the tense feature on b e (since this is unspecified), and can’t be
contextually valued either (since the absence of tense on b e means there is
insufficient contextual information). This entails that the use of overt arti-
cles in e Drop clauses will cause the derivation to crash; and this results
in the article receiving a null spellout by default (or forces the use of a null,
tenseless article).
Still, while this analysis accounts for why Tense Drop (i.e. e Drop)
sentences like (76) above show Article Drop, it begs the question of why
Article Drop is also found in clauses like those below which contain a
(bold-printed) finite auxiliary or non-auxiliary verb which is either present
tense (as in 83), or past tense (as in 84):

(83) a. Jez: <A> life of crime does pay (Daily Star, 11 May)
b. <A> mum tells of <an> encounter with <a> madman (Sunday People,
25 April)
c. <A> cyber attack shuts down <a> US oil pipeline (The i, 10 May)
388 Abbreviated registers

d. <The> PM confirms <a> huge easing of lockdown (Metro, 11 May)

e. <The> navy heads home but <the> dispute rages on (The
Independent, 7 May)

(84) a. <The> Tory party paid <the> bill for <a> 58,000 flat makeover (Daily
Mail, 27 April)
b. <A> Tory peer victimised <a> female worker at <a> firm (Daily
Telegraph, 14 April)
c. <A> terror victim helped <an> attacker catch <a> train (Daily
Telegraph, 15 April)
d. <The> Duke planned his funeral with military precision (Daily
Telegraph, 16 April)
e. <A> top Cameron mandarin took <a> job from Greensill (The Times,
14 April)

One possible answer is to suppose that Article Drop results from failing to
mark nominal tense, and e Drop from failing to mark clausal tense (on
be ). It would then follow that Article Drop can occur irrespective of whether
clausal tense is dropped (because there is no overt article on the nominal
needing to be assigned a tense value), but that Tense Drop/ e Drop requires
Article Drop, because the tense feature on overt articles can’t be valued in
the absence of clausal tense (for the reasons set out in the indented passage
in the preceding paragraph).
An alternative account of the correlation is developed by Reich (2017).
He argues that headlines are used to comment on some event e which is the
topic of the headline. He maintains that headlines contain a null e-topic
in spec‑TOPP, and that this event topic binds (and thereby licenses) a null
copula/NC, and can also bind (and license) a null article/NA on one or more
nominals. Although I will not attempt to sketch his analysis here (since it
requires familiarity with work in formal semantics), I note that it follows
from his analysis that the use of a null article or null copula in headlines
is subject to a condition to the effect that neither a null article nor a null
counterpart of b e can be c‑commanded by a nominal containing an overt
article, or by an overt copula. This condition predict the pattern of (un)-
grammaticality below (NA = null article; NC = null copula):

(85) a. A policeman is awarded the Victoria Cross

b. NA policeman NC awarded NA Victoria Cross
c. NA policeman NC awarded the Victoria Cross
d. NA policeman is awarded the Victoria Cross
e. *NA policeman is awarded NA Victoria Cross
f. *A policeman NC awarded the Victoria Cross
g. *A policeman NC awarded NA Victoria Cross
 7.4 Be Drop 389

Examples (85b–85d) satisfy the C‑command Condition because no NA or

NC is c‑commanded by an overt copula or by a nominal headed by an
overt article. By contrast, in (85e) the second NA violates the C‑command
Condition because it is c‑commanded by the overt passive auxiliary is.
In (85f), the use of NC violates the C‑command Condition because it is
c‑commanded by a nominal headed by an overt article (a policeman). And
in (85g) both NA and NC are c‑commanded by a nominal headed by an
overt article, in violation of the c‑command requirement.
However, I would note that my own judgement is that (85e) is fine, and
indeed my sample of headlines includes the following:

(86) Navalny is moved to NA prison hospital (Financial Times, 20 April)

where the null article is c‑commanded by the overt passive auxiliary is.
This suggests that the C‑command Condition needs to be modified, perhaps
as below:

(87) Generalised C‑command Condition on Article/B e Drop

Neither a null article nor a null counterpart of b e can be c‑commanded
by a nominal with an overt article

This yields the same grammaticality judgements for all the examples in
(85), except (85e), which is predicted to be fine because neither null article
is c‑commanded by a nominal with an overt article.
Still, it seems odd that a seemingly arbitrary syntactic condition like
(87) should govern the use of a null article/null b e in headlines. We might
therefore wonder what the conceptual basis of this condition is. Since the
c‑commanding nominal which bars the use of a null article/null b e lower
down in the clause is invariably the subject (hence the first constituent
in the reduced clausal structure of headlines), (87) amounts (in layman’s
terms) to positing that if you start a headline with an overt article on the
subject, you are thereby signalling that you are using a non-abbreviated
register, and hence considerations of cohesion (i.e. continuity of style/
register) demand that you don’t use null articles or null b e (which are char-
acteristic of abbreviated registers) lower down in the clause. Note, however,
that the converse doesn’t appear to be the case: for example, in headines
like (85c, 85d) you can start the sentence in an abbreviated register (by
using a subject with a null article) and then switch to a non-abbreviated
register and use an overt copula/overt article later/lower in the structure.
This would suggest that there is a constraint on code-switching (more pre-
cisely, register-switching) to the effect that switching from a non-standard
(abbreviated) register to a standard (unabbreviated) register is permitted,
but the converse (i.e. switching from a standard unabbreviated register to a
non-standard abbreviated one) is avoided.
390 Abbreviated registers

To summarise: in Module 7.4 I have examined the phenomenon of e

Drop in headlines. I presented the two alternative accounts of e Drop
below:

(88) Agreement Deficit analysis of B e Drop

e Drop clauses are truncated SUBJP constituents, headed by a SUBJ

with defective agreement properties (e.g. with number but not person).
SUBJ attracts be to raise to adjoin to it, but (being defective) cannot
attract the subject to move to spec‑SUBJP; and since the clause lacks
FIN, its subject cannot be assigned nominative case. Instead, the subject
raises only as far as spec‑TP, and is assigned accusative case by default.
e receives a null spellout by default at PF, because there is no overt

spellout for defective forms of b e .

(89) Tense Deficit analysis of B e Drop

e Drop clauses are truncated SUBJP constituents with a tense deficit,

which means that be receives a null spellout by default, and its subject
receives default case. On one view, e Drop clauses contain a TP whose
head T contains an affix with an unvalued tense feature; on another
view, they lack both TP and DP projections, and have NP subjects.

I also noted that e Drop and Article Drop are correlated, and looked at
accounts of this correlation, concluding that it can be described in terms of
the Generalised C‑command Condition in (87), which may ultimately be redu-
cible to a constraint on register-switching, permitting switching from a non-
standard (abbreviated) register to a standard (unabbreviated) register, but not
from a standard (unabbreviated) register to a non-standard (abbreviated) one.
You should now be able to tackle Exercise 7.4.

7.5 Object Drop

An interesting characteristic of certain abbreviated registers of written English

is that they exhibit a phenomenon often referred to as Object Drop, because
it involves an object being dropped. This is widely attested in instructional
registers such as recipe books, maintenance or repair manuals, stage dir-
ections, product labels, and so on, but is also found in a range of informal
written registers, such as diaries, SMS text messages, and social media posts.
Typical examples I have gathered from the labels on a number of household
products are shown below (with the dropped object shown as ø):

(90) a. Don’t stir ø. Don’t thin ø. Brush ø on using light, even strokes in the
direction of the grain, taking care to avoid overlap (Instructions on a
tin of woodstain)
 7.5 Object Drop 391

b. Heat ø on full power. Mix ø halfway through cooking ø and leave

ø to stand for 1 minute. Stir ø before serving ø (Instructions on a
packet of frozen noodles)
c. Shake ø gently before use. Store ø in the original container. Use ø
within two months of first use. Do not store ø above 25˚C. Do not
freeze ø. (Instructions on a nasal spray)
d. Pour ø into a low pressure sprayer or watering can or bucket and
apply ø over the area to be cleaned. Allow the product 20 minutes to
lift the dirt/moss/algae from the surface and then scrub ø with a firm
brush. Finally wash ø off, preferably with a jet wash (Instructions on
a container of patio cleaner)
e. Mix the sachet contents with the water in a small saucepan. Bring ø
to the boil gently, stirring ø continuously. Simmer ø for 1–2 minutes,
until the sauce is smooth and thickened (Instructions on a packet of
sauce mix)

There are several reasons for positing that such sentences contain null
objects (in each of the positions marked ø) and that these objects are active
in the syntax. Thus, the null object can serve (for example) as the controller
of PRO in a sentence like (91a) below, or as the binder for a reflexive ana-
phor like itself in (91b), or as a constituent modified by secondary predi-
cates as in (91c):

(91) a. Knead the dough and leave ø in a warm place [PRO to rise]
b. Push the dough out and fold ø back on itself each fourth stroke (Neal
1985: 47)
c. Serve ø hot with lots of butter (Neal 1985: 35)

Massam & Roberge (1989: 135) note of the empty object/EO that ‘EO
receives a specific interpretation, rather than, for example, an arbitrary
reference. The empty category need not have a particular linguistic ante-
cedent; instead, its reference appears to be contextually defined.’
Massam & Roberge note (1989: 135–36) that the null/empty object is subject
to a constraint that it must appear in a position where it is theta-marked and
case-marked (more specifically, assigned accusative case) by the same transi-
tive verb. This means that it can’t occur as the subject of an ECM clause like
that bracketed in (92a) below, or as the object of a preposition like with in (92b):

(92) a. Put cake in oven. *Expect [ø to be done half an hour later]

b. Mix the lemon juice and chopped parsley. *Then sprinkle scallops with ø

A further constraint to note is that the null objects are always interpreted
as third person referential/non-expletive pronouns – as is the case with the
null objects in (90) above. Thus, the empty object is a specific, thematic,
accusative, third person referential pronoun.
392 Abbreviated registers

Key questions arising from the various observations above are: What
is the nature of the null objects in sentences like (90) above, and how can
we account for the constraints on their use? In this module, I compare and
contrast three alternative accounts of null objects in instructional registers.

7.5.1 A Topic Drop analysis of null objects

An analysis which has been widely advocated in the relevant literature
(dating back to Haegeman 1987a, 1987b) is one which claims that null
objects in instructional registers are the result of Topic Drop. A number of
different implementations of this analysis have been proposed in the liter-
ature, but for expository purposes, I’ll present only one here, based on the
following assumptions (with 93ii taken from the Truncation account of null
subjects in 12ii above):

(93) Topic Drop analysis of null objects

(i) Null objects arise when a pronoun is topicalised and thereby moves
to the edge of a TOPP projection at the root of a truncated clause
(ii) The edge of the highest projection in a sentence structure can be
unpronounced in the PF component, provided that it contains only
phonetically weak material with recoverable semantic content

Let’s look at how such an analysis might handle the use of null objects in a
recipe instruction such as (90a) Don’t stir. Let’s suppose that imperatives have a
null counterpart of you as their subject (below denoted as you) in spec‑SUBJP,
and also contain a null imperator operator (OpIMP) in spec‑FORCEP (cf. Han
1998, 2000a, 2000b, 2001; Haegeman & Greco 2017: 9, fn. 4). In addition, let’s
posit that the missing object in (90a) Don’t stir is the pronominal DP it (refer-
ring to a tin of woodstain) and that (via Topicalisation) it moves to become the
specifier of a TOPP projection at the root of the clause. Given these assumptions
(and others), the sentence will have the syntactic derivation below (simplified
in numerous ways, including by not showing the internal structure of SUBJP):

(94) TOPP

DP TO P′
it
TOP FORCEP
ø
ADVP FORCE′
OpIMP
FORCE FINP
ø
FIN SUBJP
don’t you don’t stir —
 7.5 Object Drop 393

When the structure in (94) is transferred to the PF component, the topic it

is unpronounced in accordance with the Topic Drop spellout rule in (93ii)
above, so deriving (90a) Don’t stir!
Some of the assumptions made in (94) are worth expanding on. For
one thing, I have assumed that although FORCEP is normally the high-
est projection in an embedded clause (in order to satisfy the selectional
requirement for a predicate which selects e.g. an exclamative clause as
its complement to be immediately adjacent to that complement), no such
requirement holds in root clauses – hence a TOPP can be positioned on
top of FORCEP in an exclamative root clause such as (95) below (where
the italicised exclamative phrase is in spec‑FORCEP under the analysis in
§5.4.1):

(95) [TOPP Cristiano Ronaldo, [FORCEP what a fantastic career he has had]]

A second assumption made in (94) is that don’t undergoes Auxiliary

Inversion in negative imperatives, and thereby ends up in FIN (preceding
the subject in SUBJP) – an assumption which gains empirical support from
the relative position of auxiliary and subject highlighted below:

(96) Don’t you dare contradict me!

A third assumption made in (94) is that the sentence contains a null impera-
tive operator whose criterial position is the edge of FORCEP. This operator
may well originate in a lower position (perhaps as the specifier of a FIN head
which is imperative in mood, where it triggers Auxiliary Inversion before
moving to spec‑FORCEP), though this is not shown in (94). Interestingly, no
Inversion takes place in positive imperatives, as we see from the fact that
we find You be careful! rather than *Be you careful! There is much more to
be said about imperatives (see the references in the Bibliographical notes),
but my concern here is with Topic Drop.
While the Topic Drop analysis in (94) offers the apparent theoretical ad-
vantage that it makes use of the same root ellipsis mechanism as was em-
ployed in the analysis of Subject Drop in (12ii), it turns out to be potentially
problematic in several respects. For one thing, it is a weak pronoun – as we
see from the fact that it can’t be modified by the focus particle even as in
(97a) below, it can’t occur in focus position in a pseudo-cleft like (97b), and
it can’t be a topic in a sentence like (97c):

(97) a. Even this/*Even it didn’t upset her

b. What upset him was this/*it
c. This/*It, I can’t stand

Moreover, the Topic Drop analysis in (94) fails to account for why the null
object in instructional registers (in sentences like 90 above) is always third
394 Abbreviated registers

person. After all, fronted overt topics can be first or second person, as illus-
trated by the fronted topics italicised in (98) below:

(98) Me too/Us journalists/You guys, he would never talk to

The fact that overt fronted topics are not subject to person restrictions
whereas null objects in instructional registers are restricted to being third
person raises questions about the plausibility the Topic Drop analysis of
null objects.
A further potential problem is that (as the examples below illustrate),
fronting of an italicised topic in an imperative yields a degraded outcome:

(99) a. *The weapons leave — behind! (Jensen 2007)

b. *This book, leave — on the table! (Bianchi & Frascarelli 2010)

The low acceptability of sentences like (99) can be be attributed to an inter-

vention effect, if imperatives contain a null imperative operator in their
periphery, and if (as in 94 above) the fronted topic illicitly moves across the
peripheral imperative operator in violation of the Intervention Condition
which bars likes from crossing likes. On one view, an intervention violation
would arise because a topic undergoing A‑bar Movement into a peripheral
position (spec‑TOPP) crosses an imperative operator which also undergoes
A‑bar Movement into a peripheral position (e.g. if it moves from spec‑FINP
to spec‑FORCEP). But if a null imperative operator is a barrier to fronting
a topic, the Topic Drop analysis in (94) would wrongly predict that null
objects would not occur: and yet null objects do indeed occur in impera-
tives, as we see from (90) above.
A final problem to note is that the Topic Drop analysis in (94) runs into
potential problems in relation to contrasts such as the following (from Weir
2017):

(100) a. Wait for ten minutes [before removing ø from oven]

b. *The chicken, wait for ten minutes [before removing — from the oven]

(101) a. WARNING: [Placing ø into microwave] may cause damage

b. *This dish, [placing — into a microwave] may cause damage

As these examples illustrate, a null object can occur inside a bracketed

clause which is an adjunct as in (100a), or a sentential subject (i.e. a clause
used as the subject of another clause) as in (101a). However, no topic can
be extracted out of an adjunct clause or a sentential subject, as we see from
the ungrammaticality of (100b) and (101b), because any such movement
would lead to violation of the Adjunct Island Constraint or Subject Island
Constraint of Ross (1967). Thus, the Topic Drop analysis wrongly predicts
that (100a, 101a) should not occur.
 7.5 Object Drop 395

Given the objections outlined above to the Topic Drop analysis in (93),
let’s turn to look at an alternative analysis proposed by Weir (2017), under
which Object Drop is taken to involve Article Drop.

7.5.2 An Article Drop analysis of null objects

The core assumptions underlying Weir’s (2017) Article Drop analysis of null
objects can be summarised informally as follows:

(102) Article Drop analysis of null objects

Null objects are DPs headed by a null article with an NP complement
which can optionally be unpronounced at PF (thereby giving rise to a
null object)

On this view, Article Drop involves a DP headed by a null D which has an

overt NP as its complement, whereas Object Drop involves a DP headed by
a null D which has a null NP as its complement. This analysis provides an
interesting perspective on the relation between pairs of sentences like those
below:

(103) a. Wait for ten minutes [before removing <the> chicken from oven]
b. Wait for ten minutes [before removing ø from oven] (=100a)

On one implementation of Weir’s analysis, the article-dropped object in

(103a) will have the structure in (104a) below, whereas the null object ø in
(103b) will be a DP with the structure in (104b) below:

(104) a. DP b. DP

D NP D NP
<the> chicken <the> <chicken>
On this view, null objects have essentially the same DP structure as arti-
cle-dropped objects, since both are DPs headed by a null determiner with
an NP complement, and they differ only in respect of whether the relevant
NP is overt as in (104a), or null as in (104b).
Potential evidence in support of the subsuming Object Drop within Arti-
cle Drop comes from the observation that the two frequently occur together
within the same sentences – as illustrated by the instructions below (taken
from the labels on various household products), where I have italicised
missing articles like <the> and bold-printed missing objects like <them/it>:

(105) a. Best before: see <the> side of <the> pack. Keep <them> frozen. If
thawed, use <them> within 24 hours, and do not refreeze <them>.
For <the> best results, cook <them> from frozen. If thawed, reduce
<the> cooking time accordingly (Cooking instructions on a packet of
frozen pancakes)
396 Abbreviated registers

b. For <the> best results, cook <them> from frozen. If thawed, reduce
<the> cooking time accordingly. Remove all <the> packaging before
cooking <them>. To microwave <them>: Place <the> contents
in a microwaveable container with 3 tbsp oil, cover <them> and
vent <them>. Microwave <them> on full power for 20 minutes
(Instructions for frozen noodles)
c. To open <it>, squeeze <the> cap and turn <it>. Pour 20 ml of
undiluted mouthwash, rinse <it> around <the> teeth and gums
for 30 seconds, then spit <it> out. Do not swallow <it>. To close
<it>, turn <the> cap until it clicks. (Instructions on a bottle of
mouthwash)
d. Squeeze <the> pouch gently to separate <the> rice grains. Tear
or cut <the> top of <the> pouch by approx. 2 cm. Place <the>
pouch upright in the microwave. Microwave <it> on full power for
2 minutes. Remove <the> pouch from <the> microwave, and tear
<it> open (Instructions on a packet of cooked rice)

The co-occurrence of Article Drop and Object Drop within the same register
is precisely what would be expected under Weir’s analysis in which Object
Drop is treated as involving Article Drop.
The Article Drop analysis of Object Drop in (102, 104) accounts for sev-
eral properties of Object Drop. Firstly, both Article Drop and Object Drop
are found only in written registers, not in the spoken language – hence
neither results in a grammatical outcome in a conversation like that below:

(106) speaker a : What shall I do with the baking tray, honey?

speaker b : Put the tray/it/*tray/*ø over there, sweetie

Secondly, null objects denote specific entities (as we see from sentences
like 90 above), in the same way as dropped articles typically have a specific
interpretation. Thirdly, null objects are always third person, as noted earl-
ier. This follows from the Article Drop analysis, since articles are inherently
third person, as we see from the observation that an article-modified nom-
inal requires third person agreement on a verb like that underlined below,
and binds a third person anaphor like himself/herself:

(107) The/A true scholar always applies himself/herself assiduously to research

A fourth property of null objects which follows from the Article Drop analy
sis is that null objects are never expletive: under the Article Drop analysis,
this follows from the absence of expletive NPs in English.
A fifth property accounted for under the analysis in (102, 104) is that null
objects can freely occur inside islands (as illustrated in 100a, 101a above),
since the same is true of null articles like those italicised below:
 7.5 Object Drop 397

(108) a. Wait for ten minutes [before removing <the> chicken from <the>
oven]
b. WARNING: [Placing <a> metal container into <a> microwave] may
cause damage

Overall, then, the null article analysis of Object Drop accounts for a wide
range of properties of null objects.
However, although the idea of subsuming Object Drop under Article
Drop is tempting, there are asymmetries between the two which potentially
undermine the Article Drop analysis of null objects. One such asymmetry
relates to the registers in which the two occur. For example, the sample of
390 newspaper headlines that I collected (discussed in §7.3.1) contained
462 instances of Article Drop, but not a single instance of Object Drop – a
discrepancy hard to account for under an analysis like (102, 104) which
treats Object Drop as a form of Article Drop. Moreover, in the (admittedly
small) sample of product labels which I looked at, I found that some (like
those in 105 above) involve both Object Drop and Article Drop, whereas
others (like those in 90 above) show Object Drop but not Article Drop – sug-
gesting that the two are different phenomena, and thus potentially under-
mining the symmetrical analysis in (104).
A further asymmetry between Object Drop and Article Drop concerns
the syntactic environments in which they apply. As we saw in §7.3.1 (and
illustrated below), Article Drop can apply to nominals in a wide range of
positions:

(109) a. <The> Bank predicts <the> strongest year since <the> Second World
War as Britain surges out of lockdown (= 39a)
b. <The> EU unveils <a> big push for <the> Pfizer vaccine (= 39c)
c. <The> rota will mean [<the> Queen is not alone] (= 40b)
d. Downing Street sources claim [<the> former chief adviser released
<the> PM’s messages out of spite] (= 40c)
e. Starmer hits out [as <the> Tories reject <the> Labour call for <a>
Commons inquiry into lobbying] (= 40f)

These examples show that Article Drop can apply to subjects (whether of
main, complement, or adverbial clauses), to direct objects, to prepositional
objects, to possessor/genitive nominals and so on. By contrast, only the-
matic objects of verbs can undergo Object Drop, not (as we saw in 92 above)
objects of prepositions, or subjects of ECM clauses. The Article Drop analy
sis leaves this asymmetry unaccounted for (as Weir 2017 acknowledges).
Given the problems posed by the Article Drop analysis, in the next sec-
tion, I will propose an alternative analysis, which (as far as I am aware)
has not previously been proposed for Object Drop in instructional registers.
398 Abbreviated registers

7.5.3 Null objects as instances of a null pronoun/pro

This section explores the possibility that dropped objects in instructional
registers are instances of an inherently null, specific, thematic, accusative,
third person referential pronoun (i.e. pronominal DP) pro. What this means
in more concrete terms in that a sentence like (90a) Don’t stir will have an
analysis along the following lines (where the null object pro is assigned
accusative case and theta-marked by the verb stir, and the null imperative
subject is shown as you):

(110) FORCEP

ADVP FORCE′
OpIMP
FORCE FINP
ø
FIN SUBJP
don’t
DP SUBJ′
you
SUBJ TP
don’t
T NEGP
don’t
NEG VP
n’t
DP V′
you
V DP
stir pro
(I set aside here various questions of detail, including whether the imperative
operator moves from some lower position – for example, from spec‑FINP,
thereby triggering FIN to attract the auxiliary don’t to adjoin to it. I also set
aside the question of whether the subject transits through spec‑TP on its
way to spec‑SUBJP.) Analyses of null objects in standard language varieties
in terms of pro have been proposed for a variety of languages, including
Chamorro (Chung 1984), French (Zribi-Hertz 1984), Italian (Rizzi 1986),
Imbabura Quechua, Korean and Thai (Cole 1987), Hungarian (Farkas 1987),
Norwegian (Åfarli & Creider 1987), Brazilian Portuguese (Farrell 1990), and
Catalan (Gavarrò 1991).
Rizzi (1986) argues that null constituents need to be both licensed (via a
local c‑command relation with an appropriate kind of head) and identified.
More specifically, Rizzi remarks that for any null constituent like pro, lin-
guistic theory should:
 7.5 Object Drop 399

specify (a) the conditions that formally license the null element (the conditions
that allow it to occur in a given environment) and (b) the way in which the
content of the null element (minimally, its phi-features) is determined or
‘recovered’ from the phonetically realized environment.
(Rizzi 1986: 518)

Accordingly, a question posed by the analysis in (110) is how the pro object
of stir is licensed, and how it is identified.
As far as licensing is concerned, a potential answer is that the object
pro found in instructional registers in English is licensed by being the-
ta-marked and case-marked (more specifically, assigned accusative case)
by a transitive verb which locally c‑commands it (the licensing verb for
the object pro being stir in 110). This will account for pro not being able to
function as an ECM subject in (92a) *‘Expect [pro to be done half an hour
later]’ because pro is assigned accusative case by the transitive verb expect
but theta-marked by the passive participle done. It will also rule out use of
pro as a prepositional object in a structure like (92b) ‘Then sprinkle scallops
with pro’, because pro is neither case-marked nor theta-marked by a tran-
sitive verb. Furthermore, the requirement for pro to be theta-marked would
account for why it can’t be used as an expletive.
As for identification, this relates to the problem of determining what
(person/number) φ-features pro carries. Farrell (1990) argues (in relation to
Brazilian Portuguese) that pro can either be identified via agreement with
an appropriate head (e.g. an auxiliary/verb richly inflected for person/num-
ber agreement can identify a pro subject), or by virtue of carrying intrinsic
φ-features of its own. Following this approach, let us suppose that the pro
found in Object Drop structures in instructional registers is an inherently
third person referential pronoun which can be either singular or plural – for
example, it is singular in (111a) below and hence substitutable by it, and
plural in (111b) and hence substitutable by them:

(111) a. Place pro in the fridge on its own to cool

b. Place pro in the fridge on their own to cool

Parallels can then be drawn with a different kind of pro found in English,
namely the null pro subject found in imperative structures such as the
following:

(112) a. Please pro keep yourself well wrapped up in winter!

b. Please pro keep yourselves well wrapped up in winter!

On the assumptions made here, imperative pro subjects are inherently

specified as second person singular/plural (singular in 112a, plural in
400 Abbreviated registers

112b): hence I showed the null second person imperative subject as you in
110 above, to avoid confusion with the dropped object pro, which is inher-
ently third person.
The key assumptions embodied in the pro analysis of null objects in the
recipe register can be summarised as follows:

(113) Pro analysis of null objects

Null objects in instructional registers are instances of a null, specific,
referential, thematic, third person, singular or plural, accusative
pronominal DP pro

The pro analysis can account for significant differences between Article
Drop and Object Drop. Since the two are distinct phenomena, it comes as
no surprise that headlinese should allow Article Drop but not Object Drop;
and conversely, it comes as no surprise that in some product labels like (90)
above we should find Object Drop but not Article Drop. Similarly, it comes
as no surprise that Article Drop can apply to prepositional objects (since
this applies freely to an initial article in any nominal), but Object Drop
cannot (since it only applies to pronominals which are theta-marked and
case-marked by a transitive verb).
A further potential consideration in favour of the pro analysis of null ob-
jects is that it provides a principled account of why pro (being a pronoun)
resembles an overt pronoun like it in being able to be used cataphorically
(i.e. referring forward to a bracketed nominal following it) in a sentence
such as the following:

(114) Remove it/ø from heat immediately if [<the> omelette] starts to burn

Moreover, like the overt pronoun them, the null object pronoun ø can have
split antecedents – as in the example below, where ø (like them) refers back
to the two separate underlined constituents:

(115) Sprinkle the flour over the cold, cooked rice in the mixing bowl and fold
ø/them together gently (Ito 1993: 77, sourced from Neal 1985: 59)

Overall, then, the pro analysis accounts for a wide range of properties of
null objects.
Still, it might be objected that the pro analysis is flawed from a theoret-
ical perspective, in that it is inconsistent with the view that English has the
typological characteristic that it is a ‘no pro’ language (i.e. one which lacks
pro constituents). However, this typological claim is called into question
by two observations. Firstly, imperatives allow second-person pro subjects
(serving as the antecedent of the italicised expressions) in sentences such
as (112) above, or (116) below:
 7.5 Object Drop 401

(116) a. Don’t pro sell yourself/yourselves short!

b. pro do it on your own!

Moreover, Collins (2021) argues at length that short passives (i.e. passives
which have no overt agent by‑phrase) contain a syntactically projected
pro agent (paraphrasable somewhat awkwardly as ‘by one’ in 117a below)
which can bind an anaphor like that italicised in (117a), or license a sec-
ondary depictive predicate like that italicised in (117b) (where we could
alternatively have an overt agent like by campers):

(117) a. Such privileges should be kept to oneself (Roberts 1987: 162)

b. At the commune, breakfast is usually eaten nude (Collins 2005: 101)

In relation to (117b), it is interesting to note that the object object pro found
in the recipe register can also license secondary predicates, as illustrated
earlier in relation to (91c) ‘Serve pro hot with lots of butter.’ Thus, the
claim that positing a null pro object in recipe registers creates a typological
anomaly is undermined by the observation that there are other types of pro
constituent found in English.
To summarise: Module 7.5 has been concerned with the null objects that
occur in a range of registers of written English, particularly instructional
registers. I began (in §7.5.1) by outlining a Topic Drop analysis under which
Object Drop clauses are truncated TOPP structures which involve a topic
being fronted and moved to spec‑TOPP in the syntax, and then being given
a null spellout in the phonology; I noted, however, that such an analysis
provides no principled account of why null objects can occur inside syn-
tactic islands. I then went on (in §7.5.2) to outline an alternative analysis
under which Object Drop is treated as a particular instance of Article Drop,
so that both article-dropped nominals and dropped objects are DPs head-
ed by a null determiner, differing only in whether the NP complement of
the determiner is overt or null. I noted, however, that there are significant
distributional differences between Article Drop and Object Drop which call
into question whether Object Drop can be subsumed under Article Drop:
for example, Article Drop is pervasive in headlinese, but Object Drop is not;
and Article Drop can freely apply to prepositional objects but Object Drop
cannot. Finally (in §7.5.3), I outlined a third analysis, under which dropped
objects in instructional registers are instances of an inherently null, specif-
ic, thematic, accusative, third person referential pronoun (i.e. pronominal
DP) pro. I noted that the pro analysis accounts for significant differences
between null objects and article-dropped nominals in respect of the regis-
ters and syntactic environments they occur in.
You should now be able to tackle Exercise 7.5.
402 Abbreviated registers

7.6 Summary

This chapter has been concerned with structures which are abbreviated, in
the sense that they involve dropping various kinds of constituent. I began
in Module 7.1 by outlining Rizzi’s Truncation analysis of Subject Drop in
colloquial spoken English sentences like <I> must have left my phone in
the car, under which truncating (i.e. not projecting) peripheral projections
above SUBJP in root clauses leaves the subject I as a root specifier, and this
allows material on the edge of a root projection to be unpronounced in the
PF component.
In Module 7.2, we went on to look at Auxiliary Drop in spoken English
sentences like <Are> you feeling OK? We saw that this is sometimes accom-
panied by Subject Drop in sentences like <Are you> feeling OK? I noted that
sentences involving Aux+Subject Drop can’t be handled in any straight-
forward way under a Truncation account, and can better be accounted for
by supposing that such sentences have the same full (FORCEP) syntactic
structure as their unreduced counterparts, but can be given an abbreviated
spellout via an ellipsis operation in the PF component which gives a null
spellout to the left edge of one or more peripheral projections at PF. I noted,
however, that the Left Edge Ellipsis account can’t handle abbreviated sen-
tences like <Has the pro>’fessor arrived yet? and that this requires a pro-
sodic account under which a string of one or more unstressed syllables at
the beginning of a prosodic phrase (e.g. a sentence) can be unpronounced.
In Module 7.3, we looked at Article Drop, and saw that this is restricted
to written registers, and is sporadic in some registers (e.g. social media like
Twitter), but systematic in others (e.g. newspaper headlines). I argued that
Article Drop involves a definite or indefinite article not being pronounced/
printed if it is the first word in the functional projection which it heads,
resulting in reduced sentences like <The> government imposes <a> ban on
<the> import of livestock.
In Module 7.4, I discussed the phenomenon of B e Drop whereby in some
abbreviated written registers such as headlinese, finite (indicative) forms of
be are systematically dropped, resulting in sentences like BoJo <is> losing
support. I outlined two alternative Truncation analyses of these. Under one,
e Drop clauses are truncated SUBJP constituents, headed by a SUBJ with
defective agreement properties (e.g. number but not person) which attracts
be to raise to adjoin to it, but (being defective) cannot attract the subject
to raise to spec‑SUBJP. Instead, the subject raises only as far as spec‑TP (in
the syntax, at least), and is assigned accusative case by default, and defec-
tive be receives a null spellout by default at PF. An alternative account is
that e Drop clauses are truncated SUBJP constituents with a tense deficit
 7.6 Summary 403

which means that b e receives a null spellout, and its subject receives default
case: on one implementation, they contain a TP whose head is unspecified
for tense; on another view, they lack both TP and DP projections, and have
NP subjects.
In Module 7.5, I discussed the phenomenon of Object Drop found in
some registers of written English, especially instructional registers. I be-
gan by outlining a Topic Drop analysis under which Object Drop clauses
are truncated TOPP structures which involve a topic being fronted and
moved to spec‑TOPP in the syntax, and then being given a null spell-
out in the phonology. I went on to outline an alternative analysis under
which Object Drop is treated as a particular instance of Article Drop, so
that both article-dropped nominals and dropped objects are DPs headed
by a null determiner, differing only in whether the NP complement of
the determiner is overt or null. Finally, I outlined a third analysis, under
which dropped objects are instances of an inherently null, specific, the-
matic, accusative, third person, singular/plural referential pronoun (i.e.
pronominal DP) pro.
Key analyses, principles, conditions and hypotheses introduced in this
chapter include the following:

(12) Syntactic truncation account of Subject Drop

(i) A sentence can have a truncated syntactic structure whose root is not
FORCEP, but some lower projection such as SUBJP
(ii) The edge of the highest projection in a sentence structure can be
unpronounced in the PF component, provided that it contains only
phonetically weak material with recoverable semantic content

(26) Principle of Unambitious Reverse Engineering (PURE)

When determining the identity of unpronounced material in the course
of reverse-engineering a speaker’s syntactic derivation, the language
system of the hearer considers only the minimally semantically
contentful possibilities compatible with the morphosyntactic
environment (Pesetsky 2020: 5)

(32) Left Edge Ellipsis account of Auxiliary (+Subject) Drop

(i) Reduced clauses have the same full syntactic structure as their
unreduced counterparts
(ii) The left edge of a continuous set of one or more projections at the
top of a tree in a root clause can undergo ellipsis in the phonology

(37) Prosodic account of Weak Syllable Drop

(i) Reduced clauses have the same full syntactic structure as their
unreduced counterparts
404 Abbreviated registers

(ii) A continuous sequence of one or more weak syllables at the

beginning of a prosodic phrase (e.g. a sentence) can undergo ellipsis
at PF

(50) Article Drop

An article which is the first word in the functional projection that it
heads can be dropped

(87) Generalised C‑command Condition on Article/B e Drop

Neither a null article nor a null counterpart of b e can be c‑commanded
by a nominal with an overt article

(88) Agreement Deficit analysis of B e Drop

e Drop clauses are truncated SUBJP constituents, headed by a SUBJ

with defective agreement properties (e.g. with number but not person).
SUBJ attracts e to raise to adjoin to it, but (being defective) cannot
attract the subject to move to spec‑SUBJP; and since the clause lacks
FIN, its subject cannot be assigned nominative case. Instead, the subject
raises only as far as spec‑TP, and is assigned accusative case by default.
e receives a null spellout by default at PF, because there is no overt

spellout for defective forms of b e .

(89) Tense Deficit analysis of B e Drop

e Drop clauses are truncated SUBJP constituents with a tense deficit,

which means that b e receives a null spellout by default, and its subject
receives default case. On one view, e Drop clauses contain a TP whose
head T contains an affix with an unvalued tense feature; on another
view, they lack both TP and DP projections, and have NP subjects.

(93) Topic Drop analysis of null objects

(i) Null objects arises when a pronoun is topicalised and thereby moves
to the edge of a TOPP projection at the root of a truncated clause
(ii) The edge of the highest projection in a sentence structure can be
unpronounced in the PF component, provided that it contains only
phonetically weak material with recoverable semantic content

(102) Article Drop analysis of null objects

Null objects are DPs headed by a null article with an NP complement
which can optionally be unpronounced at PF (thereby giving rise to a
null object)

(113) Pro analysis of null objects

Null objects in instructional registers are instances of a null, specific,
referential, thematic, third person, singular or plural, accusative
pronominal DP pro
 7.7 Bibliographical notes 405

7.7 Bibliographical notes

On registers and register variation, see Ferguson (1982), Zwicky & Zwicky
(1982), Biber (1995), Haegeman (2006c), Paesani (2006), and Massam &
Stowell (2017). On more specific registers, see research on diary styles
(Haegeman 1990a, 1990b, 1997, 2013, 2017, 2019; Matushansky 1995;
Horsey 1998; Haegeman & Ihsane 1999, 2001), newspaper headlines
(Simon-Vandenbergen 1981; Stowell 1990, 1991, 1996, 1999), recipe
books and instruction manuals (Haegeman 1987a; Massam 1987; Massam
& Roberge 1989; Culy 1996; Sigurðsson & Maling 2007; Paul & Massam
2020), note-taking (Janda 1985), telegrams and text messages (Barton
1998), telephone conversations (Hopper 1992), online blogs (Teddiman &
Newman 2010), and emails/postcards (Nariyama 2006).
On Subject Drop in spoken English see Morgan (1973), Schmerling
(1973), Thrasher (1973, 1974, 1977), Drachman (1974), Napoli (1982),
Zwicky & Pullum (1983), Quirk et al. (1985), Akmajian et al. (1985), Ihalain-
en (1991), Hopper (1992), Rizzi (1994, 1998, 2000a), Cote (1996, 1997), Har-
vie (1998), Mackenzie (1998), Kay (2002), Nariyama (2004, 2006), Zwicky
(2005), Oh (2005, 2006), Schirer (2008), Weir (2008, 2009a, 2009b, 2012,
2013, 2014, 2016, 2017, 2018, 2022), Mack & Fuerst (2009), Mack (2010),
Bailey (2011), Mack et al. (2012), S. Wagner (2012, 2016), Torres Cacoullos
& Travis (2014), Ikarashi (2015), Bae (2016), Sakamoto & Ikarashi (2017),
Shibata (2017), and Lindstrom (2020).
Subject Drop in written English has been reported to occur in a wide range
of registers, including informal writing (Zwicky & Zwicky 1981; Reiman
1994), diaries (Haegeman 1987b, 1990a, 1990b, 1997, 2000a, 2002b, 2007b,
2013, 2017, 2019; Matushansky 1995; Horsey 1998; Ihsane 1998, Haegeman
& Ihsane 1999, 2001; Becquet 2000; Scott 2004, 2010; Newman & Teddiman
2010; Teddiman & Newman 2010; Nanyan 2013; Haegeman & Starke 2021),
newspaper headlines (Straumann 1935; Mårdh 1980; Simon-Vandenbergen
1981; Stowell 1990, 1991, 1996, 1999; de Lange 2004, 2008; Weir 2009b,
2013, 2017), recipe books and instruction manuals (Haegeman 1987a;
Massam 1987, 1989; Massam & Roberge 1989; Culy 1996; Sigurðsson &
Maling 2007, 2010), note-taking (Janda 1985), telegrams (Barton 1998), text/
SMS messages (Crystal 2008; Tagg 2009; Cougnon & Fairon 2012), emails/
postcards/letters (Nariyama 2006), social media (Pérez-Sabatar 2012; Grant
2015; Daeninck 2019; Lansens 2019; Deneys 2020), TV dramas (Nariyama
2004), and cinema (Ito & Kashihara 2010).
On Auxiliary Drop in English, see Hendrick (1982), Andersen (1995),
Fitzpatrick (2005b, 2006), Schirer (2008), Totsuka (2015, 2017), and Pesetsky
(2020).
406 Abbreviated registers

On Article Drop see Straumann (1935), Mårdh (1980), Simon-Vanden-

bergen (1981), Bell (1991), Stowell (1990, 1991, 1996, 1999), de Lange
(2004, 2008), de Lange & Avrutin (2009), Weir (2009b, 2012, 2013, 2014,
2017, 2018), Wiljergård (2011), Sato & Kim (2012), Horch & Reich (2016),
Lemke et al. (2017), Oosterhof & Rawoens (2017), Paul (2017), Reich (2017),
Moncomble (2018), Nemes (2019), Eilander (2020), and Tel (2020).
B e Drop is found not only in headlinese (see e.g. Reich 2017, and
Kibsgaard 2019), but also in other varieties of English, most notably in
English-based creoles and pidgins. It is a common feature of African Amer-
ican English/AAE, but note that in AAE, the present tense forms is and are
are the only forms that can be null: see Labov (1969, 1995), Cukor-Avila
(1999), Bender (2000), and Green (2002). Null copulas are also found in
many other languages of the world (often restricted to present tense use):
see Stassen (2013).
On Object Drop in English found, inter alia, in the instructional reg-
ister (which includes recipes, instruction manuals, product labels and so
on) see Sadock (1974), Allerton (1975), Fillmore (1986), Fitzpatrick et
al. (1986), Haegeman (1987a, 1987b), Massam (1987, 1989, 1992), Mas-
sam & Roberge (1989), Ito (1993), Groefsema (1995), Culy (1996), Bender
(1999), García Velasco & Portero Muñoz (2002), Cummins & Roberge
(2004, 2005), Ohara (2007), Sigurðsson & Maling (2007, 2010), Ruppen-
hofer & Michaelis (2010), Wharton (2010), Ruda (2014, 2017), Massam
et al. (2017), and Weir (2017). On Object Drop in other languages, see
Chung (1984), Huang (1984, 1991), Zribi-Hertz (1984), Raposo (1986),
Rizzi (1986), Åfarli & Creider (1987), Cole (1987), Farkas (1987), Authi-
er (1988, 1989), Suñer & Yépez (1988), Bouchard (1989), Farrell (1990),
Gavarrò (1991), Roberge (1991), Kato (1993), Landa (1995), Lambrecht &
Lemoine (1996, 2005), Visser (1996), Cyrino (1997, 2004, 2021), Arteaga
(1998), Hoji (1998), Keller & Lapata (1998), Li (1998), Kim (1999), Larja-
vaara (2000), Schwenter & Silva (2002), Xu (2003), Cummins & Roberge
(2005), Schwenter (2006), Alamillo & Schwenter (2007), Sigurðsson &
Maling (2007, 2010), Milambiling (2011), Erteschik-Shir et al. (2013), Pao-
li (2014), Panitz (2015), Cyrino & Matos (2016), Funakoshi (2017), Lee
(2017), and Paul & Massam (2020).
On the syntax of imperatives, see Zwicky (1988), Beukema & Koopmans
(1989), Zhang (1990), Platzack & Rosengren (1997), Potsdam (1998, 2018),
Han (1998, 2000a, 2000b, 2001), Moon (2001), Flagg (2002), Jensen (2003,
2007), Rupp (1999, 2003), Portner (2004, 2007), Takahashi (2004), Mauck et
al. (2005), Bennis (2006), Koopman (2007), van der Wurff (2007), Zanuttini
(2008), Aikhenvald (2010), Kaufmann (2011), Medeiros (2013), Alcázar &
Saltarelli (2014), Isac 2015, and Shormani (2021).
 7.8 Workbook 407

7.8 Workbook

In the exercises below, an [M] after an example sentence indicates that a

model answer is provided in the free-to-download Students’ Answerbook
(and in the Teachers’ Answerbook as well). You will find it helpful to look at
this model answer before attempting to tackle other examples in the exer-
cise. Some other exercise examples have an [S] after them, indicating that
they are for self-study: this means that after reading the relevant module
(and those preceding it), and looking at the helpful hints and model answer
associated with the exercise, students should be able to analyse these self-
study examples on their own, and then compare their answers with those
provided in the Students’ Answerbook. The remaining exercise examples
(i.e. those not marked with [M] or [S] after them) are intended for teachers
to use as the basis for hands-on problem-solving work in seminars, classes,
or assignments: answers to all of these non-self-study examples are pro-
vided in the Teachers’ Answerbook. Note that Exercise 7.1 tests you on
Module 7.1, Exercise 7.2 on Module 7.2, and so on; and that each exercise
in a given chapter presupposes familiarity with material covered in earlier
chapters and in earlier modules of the same chapter. Unparenthesised num-
bers like 4 refer to material in the exercises, whereas parenthesised numbers
like (22) refer to examples and conditions in the main text of the book.

EXERCISE 7.1

This exercise is designed to be attempted after reading Module 7.1 and the
chapters preceding it.

Discuss the syntax of the dropped constituents in the following sentences

from Helen Fielding’s (1996) book Bridget Jones’s Diary (these examples
being cited by Haegeman & Ihsane 1999, 2001; I have added what I take
to be the missing/dropped constituents in <angle brackets>):

1 <I> was worried that <we> might split (Fielding 1996: 227) [M]
2 <I> start to wonder whether <I> am <a> really good friend (Fielding 1996: 261)
3 <I> realise <I> was using <the> telly remote control by mistake (Fielding
1996: 153) [S]
4 <I> understand where <I> have been going wrong (Fielding 1996: 97)
5 <I> think <I> will cross that bit out as <it> contains <a> mild accusation
(Fielding 1996: 25) [S]
6 <I> search for <my> hairbrush. <I> locate <it> in <my> handbag (Fielding
1996: 92) [S]
7 <I’m> feeling v. pleased with <my>self (Fielding 1996: 77; v. = very)
408 Abbreviated registers

8 <I> suddenly start thinking of <a> former boyfriend Peter with whom <I>
had <a> functional relationship for seven years, until <we> finished for <a>
heart-felt agonizing reason <I> can no longer remember (Fielding 1996: 190)

Helpful hints
Extensive research (particularly by Liliane Haegeman and her
collaborators: Haegeman 1987b, 1990a, 1990b, 1997, 2000a, 2002b,
2007b, 2013, 2017, 2019, 2021; Haegeman & Ihsane 1999, 2001) has
shown that (especially in fictional varieties of the diary register) a much
wider range of constituents than just root/main clause subjects can be
dropped (e.g. subjects can be dropped in non-root as well as root clauses,
and some non-subject constituents can be dropped as well – e.g. articles
like a/the). The relevant register has come to be known as Diary Drop.
Compare the four different accounts of Diary Drop outlined below and
say which can best handle sentences 1–8: the analyses in 10–12 are not
covered in Module 7.1, and are designed to test your ability to compare
and contrast unseen analyses of a given set of phenomena.
9 Subject Drop Truncation
(i) In the fictional diary register, a sentence can have a truncated
syntactic structure which projects only as far as SUBJP
(ii) The edge of the highest projection in a sentence can be
unpronounced at PF

10 Topic Drop Truncation

(i) In the fictional diary register, a sentence can have a truncated
syntactic structure which projects only as far as TOPP
(ii) Dropped constituents move to spec‑TOPP in the syntax, and thereby
(by virtue of being on the edge of the highest projection in the
sentence) can be unpronounced at PF

11 Left Edge Ellipsis

(i) In the fictional diary register, clauses have the same (untruncated)
syntactic structure as in standard registers
(ii) (A continuous string of one or more) weak items with little semantic
content can be given a null spellout on the left edge of a clausal or
nominal functional projection

12 Editing account
Diary Drop sentences have the same (untruncated) syntactic structure
as in standard registers, but the representation outputted by the PF
component is edited by omitting personal pronouns and articles

In relation to 9, consider whether the restriction on Truncation only

applying in sentences (i.e. root clauses) may be relaxed in the Diary Drop
register, so allowing Truncation to apply in root and non-root clauses
 7.8 Workbook 409

as well. In relation to 11, assume that weak items include monosyllabic

personal pronouns like for example, I/you/he and articles like a/the, and
that Left Edge Ellipsis results in items which are present in the syntax being
unpronounced in the phonology (or unprinted in written diary entries).
The account in 12 differs from the others in that the analyses in 9–11 posit
that Diary Drop is a natural (structure-dependent) operation, whereas the
account in 12 assumes that it is an artificial (structure-independent) process
which involves consciously editing out certain lexical items (personal
pronouns and articles) which are weak and have minimal semantic content.
In analysing the relevant diary entries, don’t try to draw tree diagrams
for whole sentences which have a complex internal structure: instead,
consider the internal structure of each of the component clauses
separately. You might want to bear in mind the following:
13 Clause Typing Conditions (from §1.4.3)
(i) A clause is interpreted as being of a given type (interrogative or
exclamative etc.) if it contains a peripheral clause-typing specifier of
the relevant type
(ii) An indicative clause is interpreted as declarative in type by default if
it contains no peripheral clause-typing specifier

14 COMP‑Trace Filter/CTF (from §1.5.2)

Any structure at the end of a syntactic derivation in which an overt
COMP/complementiser is immediately adjacent to and c‑commands a
trace (i.e. a gap/null copy left behind by a moved constituent) is filtered
out as ill-formed

15 Intervention Condition (from §1.5.1)

Likes cannot cross likes

16 Criterial Freezing Condition/CFC (incorporating a modification from

§6.1.2)
(i) A peripheral or subperipheral constituent must occupy a criterial
position to be interpretable at LF
(ii) A constituent which occupies its criterial position is frozen in place
17 Impenetrability Condition (from §4.4.1)
Anything below the head H of the highest projection HP in the periphery
of a complete clause (i.e. one containing FORCEP) is impenetrable to
anything above HP

18 Island Constraint (after Ross 1967)

Any clause which is not the complement of a verb/adjective is an island
(in the sense that it doesn’t allow anything to be extracted out of it)

Note that it follows from 18 that adjunct/adverbial clauses, clauses

which function as subjects or topics, and relative clauses etc. are islands/
410 Abbreviated registers

barriers to extraction. Look at the model answer in the Answerbook

before developing your analyses. To simplify your task, you might want
to follow the suggestions made below.
In 2, analyse the start clause and the whether clause separately: in
analysing the start clause, take start to wonder whether I am a really
good friend to be a VP, but don’t concern yourself with its internal
structure; in analysing the whether clause, assume (as suggested in §5.3.2)
that interrogative whether is a FORCE head which has an adverbial
null yes–no question operator OpYNQ as its specifier, and assume too
that am a really good friend is a TP, but don’t concern yourself with its
internal structure. In 3, bear in mind the suggestion made in §4.4.2 that
declarative complement clauses which are not introduced by that are
FINP constituents; treat was using the telly remote control by mistake as a
TP, but don’t concern yourself with its internal structure. In 4, take have
been going wrong to be a TP, but don’t concern yourself with its internal
structure; and take where to be a PP with a meaning paraphrasable as
‘in what respect’, but don’t concern yourself with its internal structure.
In 5, analyse the main+complement clause structure I think I will cross
that bit out first: take will cross that bit out to be a TP, but don’t concern
yourself with its internal structure; and (following the analysis outlined
in §4.4.2) assume that a that-less declarative complement clause (like
the cross clause here) projects only as far as FINP. Then analyse the
adjunct/adverbial clause headed by the subordinating conjunction as: for
the reasons outlined in Radford (2018: 197–215), take it to be a CONJP
headed by the CONJ/subordinating conjunction as, and take it contains a
mild accusation to be a FORCEP. In 6, analyse each of the two sentences
separately; in the second sentence, take locate it in my handbag to be
a VP, but do not attempt to analyse its internal structure (because this
is complex, and requires familiarity with work on split VPs which is
not covered in this book). In 7, take pleased to be an adjective/A which
has the PP with myself as its complement, and treat the ADVP very as
its specifier; consider the internal structure of myself, and whether it
may have a similar structure to a phrase like my car (with my and self
seemingly being able to function as separate words in a phrase like my
old self, in the same way as they do in my old car). In 8, analyse each of
the four clauses in the sentence separately. First analyse the main clause
I suddenly start thinking of a former boyfriend Peter: take start thinking
of a former boyfriend Peter to be a VP, but don’t concern yourself with
its internal structure. Then, analyse the adjunct/adverbial/subordinate
clause until we finished for a heart-felt agonizing reason: treat the until-
clause in the same way as the as-clause in sentence 5 (see the guidance
above), and take finished for a heart-felt agonizing reason to be a VP,
 7.8 Workbook 411

but don’t concern yourself with its internal structure. Next, analyse the
relative clause with whom I had a functional relationship for seven years:
take had a functional relationship for seven years to be a VP, but don’t
concern yourself with its internal structure. Finally, analyse the relative
clause I can no longer remember. Follow the analysis in §5.4.2 under
which relative clauses are RELP constituents in which an XP containing
or comprising a relative pronoun/pronominal DP like who/which (or a
null counterpart thereof) moves to spec‑RELP from its original position
(here, inside VP): assume that the REL head is null and has a FORCEP
complement (for the reasons set out in §5.4.2).

EXERCISE 7.2

This exercise is designed to be attempted after reading Module 7.2 and the
module and chapters preceding it.

Analyse the spoken English sentences below:

1a Did you ever visit Rome? [M]

1b You ever visit Rome? [M]
1c Ever visit Rome? [M]
2a A fine friend you turned out to be!
2b Fine friend you turned out to be! (Napoli 1982: 85)
3a I’m a bit ashamed of myself [S]
3b M’a bit ashamed of myself [S]
3c A bit ashamed of myself [S]
3d Bit ashamed of myself [S]
4a What time is it? [S]
4b Time is it? [S]
5 Intended as replies to the question ‘Do you play golf?’:
a. Of course I do
b. Course I do
c. *Course do
6a Have you been exercising much lately? [S]
6b You been exercising much lately? [S]
6c Been exercising much lately? [S]
6d Exercising much lately? (adapted from Napoli 1982: 109) [S]
7a Has the cat got your tongue?
7b The cat got your tongue?
7c Cat got your tongue? (Napoli 1982: 85)
8a Have you any idea where she is?
8b You any idea where she is?
412 Abbreviated registers

8c Any idea where she is?

8d Do you have any idea where she is?
8e You have any idea where she is?
8f Have any idea where she is?

Treat the shorter examples as reduced forms of the longer examples. In

relation to 8, note that my own (British) variety of English allows all the
sentences in 8; by contrast, Joseph Galasso and Jean Hannah (pc) tell me
that their (American) varieties of English allow 8c–8f, but not 8a, 8b.

Helpful hints
Discuss how well each of the three alternative accounts of Subject/
Auxiliary Drop outlined below can handle the relevant reduced clauses in
abbreviated registers of spoken English:

9 Truncation account
(i) A reduced clause has a truncated syntactic structure whose root is not
FORCEP, but some lower projection (e.g. TOPP, or FINP, or SUBJP).
(ii) The edge of the root/highest projection in a sentence structure can be
unpronounced in the PF component

10 Left Edge Ellipsis account

(i) Reduced clauses have the same syntactic structure as their unreduced
counterparts
(ii) The left edge of a continuous set of one or more functional
projections at the top of the tree in a root clause can undergo ellipsis
in the PF component

11 Prosodic account
(i) Reduced clauses have the same syntactic structure as their unreduced
counterparts
(ii) A continuous sequence of one or more weak syllables at the
beginning of a prosodic phrase (e.g. a sentence) can undergo ellipsis
(and be unpronounced) in the PF component

Under all three analyses, assume that only weak/unstressed material

can undergo ellipsis and that ellipsis cannot give a null spellout to (the
whole of) any word with substantive lexical content (in consequence of
the Recoverability Condition requiring material undergoing ellipsis to be
recoverable). In addition, assume that an indicative clause which contains
no specifier in its periphery typing it as interrogative/exclamative etc. is
interpreted as declarative in force by default.
Make the following assumptions about the categorial status of relevant
phrases, but don’t concern yourself with their internal structure. In 2a, a
fine friend is a focused ARTP/article phrase, and you turned out to be is a
 7.8 Workbook 413

SUBJP. In 3, a bit ashamed of myself is an AP comprising the head adjective

ashamed, the KP complement of myself, and the ARTP specifier a bit. In 4,
what time is a focused interrogative QP. In both 3 and 4, be originates as
the head V of VP, and from there can raise to T (and even higher to SUBJ)
in a finite clause. In 5, of course is a KP which is a clause modifier, and is
directly merged as the specifier of a MODP immediately below FORCEP.
In 6, exercising much lately is a VP, and been is the head of a PROGP/
progressive aspect projection. In 6, 7, have originates as the head of a
PERFP/perfect aspect projection, and from there raises stepwise into FIN.
In 8, have originates as the head V of the VP you have any idea where she
is (with any idea where she is being a QP); in British (but not American)
varieties, have can behave like an auxiliary and undergo Auxiliary
Inversion and thereby raise into FIN (in a succession of short steps).

EXERCISE 7.3

This exercise is designed to be attempted after reading Module 7.3 and the
modules and chapters preceding it.

Daeninck (2019) researches constituent drop on the social media

platform Twitter, and reports the four main types of dropped constituent
<bracketed> below:

1 Subject Drop
1a <I> can’t be bothered to go food shopping [M]
1b <I> might just eat McDonald’s for the rest of my life
1c <I> wrote this in 2015 but it was my most read blog post in 2018
1d <They> know everything about everyone
1e <It> was a pleasure to meet you

2 Subject+Aux Drop
2a <I’m> uploading a new video tonight
2b <I’ve> just come off a long flight and on the plane I watched a 1990s
classic I’d somehow never seen before
2c <We’ll> see how it goes
2d <It was> so lovely meeting you
2e <There’s> no rest for the curious

3 Aux Drop
3a <Does> anyone know if climbing Scafell Pike is dog friendly? [S]
3b <Do> you ever talk to someone so much you start adopting their
mannerisms and using their slang and you just sit back and think wtf
is happening to me?
414 Abbreviated registers

3c <Is> anyone else feeling really … off?

3d <Are> you guys in?

4 Aux+Subject Drop
4a <Did you> miss my latest briefing?
4b <Do you> want to know what it was like for my team during my
#MarsLanding?
4c <Does it> work for you?
4d <Have you> missed my nail-biting #MarsLanding?
4e <Are there> questions about my #MarsLanding?

(Note that wtf in 3b is an abbreviation of what the fuck; and the missing
word shown as … in 3c is presumably pissed.) Daeninck claims that the
abbreviated tweets in her corpus obey the same constraints as operate in
spoken English – namely that subjects and/or auxiliaries are only dropped
in root clauses (and not in complement clauses or other subordinate
clauses), and only when sentence-initial. Discuss the derivation of the first
example in each set – that is, 1a, 2a, 3a and 4a – comparing alternative
ways of analysing them (referring to other examples in each set where
appropriate).
In addition to the four sentence types in 1–4 above, Daeninck’s data
include some tweets which potentially don’t conform to the patterns in
1–4, including the following (where Frozen 2 is the name of a film):

5a Shit, <I> guess I care about Frozen 2 now huh? [S]

5b <The> entire cast was on fire [S]
5c <It’s the> last day here
5d If <she’s> not used to these sorts of climbs <it’s> maybe best to take
her next time or not try her on <the> tallest mountain in England for
<the> first try!

Say in what ways each of the sentences in 5 differs from the types of
ellipsis found in 1–4, and how each might be analysed.

Helpful hints
Discuss the extent to which each of the sentences you analyse can be
handled in terms of each of the three alternative accounts of ellipsis
outlined in the helpful hints for Exercise 7.2 (namely, the Truncation, Left
Edge Ellipsis, and Prosodic accounts); bear in mind too the possibility that
the Twitter register may allow for a separate type of ellipsis involving
the kind of Article Drop operation discussed in Module 7.3, whereby an
article which is the first word in the projection that it heads can be null.
For each set of sentences in 1–4, derive the first example from each set
(i.e. 1a, 2a, 3a, 4a), and refer to other examples in the set (if appropriate)
 7.8 Workbook 415

where they illustrate some relevant point. Which analysis handles the
widest range of data in 1–5 overall?
In analysing 1a, for the purposes of the exercise, make the simplifying
assumption that can’t is an inherently negative auxiliary in T, and that be
bothered to go food shopping is a VP, but don’t concern yourself with the
internal structure of this VP. In 2a, take uploading a new video tonight to
be a VP, but don’t concern yourself with its internal structure. In 3a, take
know if climbing Scafell Pike is dog friendly to be a VP, but don’t concern
yourself with its internal structure. In 4a, take miss my latest briefing to
be a VP, but don’t concern yourself with its internal structure. In 5a, take
guess I care about Frozen 2 now to be a VP (but don’t concern yourself
with its internal structure), and treat shit as an interjection projection/
INTJP which functions as the specifier of a speech act projection/ACTP
which forms part of what we could call the superperiphery of the clause
and hence is positioned above the FORCEP which is the highest projection
in the periphery: for the purposes of the exercise, ignore the discourse
particle huh at the end of the sentence. In 5b–5d, assume that the copula
be originates as the head V of VP, and from there raises to T (and can
in principle raise further into SUBJ). In 5b, treat the entire cast as a DP
headed by the determiner the, make the simplifying assumption that entire
cast is an NP (though don’t concern yourself with the structure of this
NP). In 5c, treat last day here as an NP for the purposes of this exercise,
but don’t concern yourself with its internal structure. In the complex
sentence in 5d, analyse the if-clause and the main clause (<it’s> maybe
best to …) separately; treat the if-clause as a CONJP (headed by CONJ if)
positioned on top of the main clause structure.

EXERCISE 7.4

This exercise is designed to be attempted after reading Module 7.4 and the
modules and chapters preceding it.

Discuss how ellipsis of the <bracketed> items works in the headlines

below (collected by me from British newspapers in April–May 2021, with
the exception of Example 5:

1 <The> Tory party paid <the> bill for <a> 58,000 flat makeover (Daily Mail,
27 April) [M]
2 <The> rota will mean <the> Queen is not alone (Daily Mirror, 19 April)
3 <The> EU <is> hopeful of <a> Northern Ireland solution (Financial Times,
19 April) [S]
416 Abbreviated registers

4 <The> young <are> turning their backs on learning to drive (Daily Telegraph,
26 April)
5 Me to blame for <the> Brexit fiasco, admits Boris (spoof headline, Mike Jones,
pc) [S]
6 <We’re> sick of his Bozwallop (Daily Star, 12 May)
7 <The> Prince is back. <He> Arrives in England ahead of <the> Duke’s funeral
(The Sun, 12 April) [S]
8 <Does> Anyone fancy a trip to the beer garden? (Daily Telegraph, 7 April) [S]
9 <Do you> Call this freedom? (Daily Mail, 6 April)
10 <Is it the> End of the nightmare? (Sunday Express, 2 May)

Note that Bozwallop in (6) is an invented noun meaning ‘codswallop/

nonsense characteristic of Bozo/Boris Johnson’.

Helpful hints
Bear in mind that more than one type of ellipsis may apply in some of
these sentences (Article Drop, e Drop, Subject Drop, Aux Drop: see the
accounts of Subject/Aux Drop in the helpful hints for Exercise 7.2). Bear
in mind, too, the following assumption made in the main text about
Article Drop in headlinese:

11 Article Drop
An article which is the first word in the functional projection that it
heads can be dropped

Say whether your analysis is in conformity with the following condition

on e Drop and Article Drop:

12 Generalised C‑command Condition on Article/B e Drop

Neither a null article nor a null counterpart of b e can be c‑commanded
by a nominal with an overt article

In relation to examples like 3 and 4, evaluate the following alternative

accounts of e Drop outlined in the main text:

13 Agreement Deficit analysis

e Drop clauses are truncated SUBJP constituents, headed by a SUBJ

with defective agreement properties (e.g. with number but not person).
SUBJ attracts be to raise to adjoin to it, but (being defective) cannot
attract the subject to move to spec‑SUBJP; and since the clause lacks
FIN, its subject cannot be assigned nominative case. Instead, the subject
raises only as far as spec‑TP, and is assigned accusative case by default.
e receives a null spellout by default at PF, because there is no overt

spellout for defective forms of b e .

 7.8 Workbook 417

14 Tense Deficit analysis

e Drop clauses are truncated SUBJP constituents with a tense deficit,

which means that be receives a null spellout by default, and its subject
receives default case. On one view, e Drop clauses contain a TP whose
head T contains an affix with an unvalued tense feature; on another
view, they lack both TP and DP projections, and have NP subjects.

For the purposes of this exercise, make the simplifying assumption that
complex nominals like Tory Party, 58,000 flat makeover, EU, Brexit
fiasco, Northern Ireland solution, beer garden are NPs, but don’t concern
yourself with their internal structure. In 5, concern yourself only with
the structure of the clause Me to blame for <the> Brexit fiasco. In 7, treat
ahead as a P with a KP complement headed by the K/case particle of.

EXERCISE 7.5

This exercise is designed to be attempted after reading Module 7.5 and the
modules and chapters preceding it.

Discuss the structure of the italicised clauses in the following extracts

from authentic recipes in cookbooks:

1 Chill dough, then roll to ¼″-thick and spread with date filling and turn over
on itself (¼″-thick = ‘a quarter inch thick’) [M]
2 Slice the mushrooms finely, and put in a large bowl with the oil [S]
3 Halve the avocado; scoop the flesh into a bowl, along with the lime juice,
then mash roughly, and season to taste
4 Start ladling in the hot stock, letting each ladleful become absorbed as you
stir, before adding the next one [S]
5 Spread the tortilla chips out on a foil-lined baking sheet and sprinkle the
cheese over
6 Put the dish on the baking sheet in the oven for 30 minutes. Let stand for
10 minutes before serving, and make sure you scoop out the sauce beneath
the sponge in the dish [S]
7 To serve the birds freshly stewed, let them stand 10 minutes before cutting
8 Make the compote, if using, by putting the berries into a small pan with the
sugar and lemon juice (if using = ‘if you are using it/the compote’)

Example 1 is from Ruppenhofer & Michaelis (2010), example 7 from

Bender (1999), and the remaining examples are from Ruda (2014), with all
examples sourced from authentic cookbooks. In a couple of cases, I have
418 Abbreviated registers

shortened the sentences by omitting a few words (but not in ways that
affect issues of concern here).

Helpful hints
Compare the following three accounts of Object Drop outlined in the
book:

9 Topic Drop account

(i) In the Recipe Register, a sentence/root clause can have a syntactic
structure whose root (i.e. highest) projection is TOPP
(ii) Dropped objects move to spec‑TOPP in the syntax, and thereby (by
virtue of being on the edge of the highest projection in the sentence)
can be unpronounced at PF

10 Null Article account

(i) Article Drop involves a DP/ARTP projection headed by a null article
with an overt NP as its complement
(ii) Dropped objects involve a DP/ARTP projection headed by a null
article with a null NP as its complement

11 Pro analysis of null objects

Null objects in the Recipe Register are instances of a null, specific,
referential, thematic, third person, singular/plural, accusative pronominal
DP pro

In addition, bear in mind a point not discussed in the main text, namely
that a restricted number of verbs (and prepositions) in English allow an
implicit null definite object (below denoted as an empty category/ec) – for
example, visit in 12a below but not see in 12b, with ec paraphrasable as
her in 12a and as him in 12b:

12a She was feeling lonely, because nobody had visited ec for a long time
12b *She missed her brother, because she hadn’t seen ec for a long time

Let’s call this kind of null object a ‘lexically licensed null object’, since
only certain lexical items allow its use.
Because many of the sentences in this exercise have long and complex
structures, analyse the specific clause or phrase containing each dropped
constituent separately, and avoid analysing the internal structure of
verb phrases (because some of these are complex): in some cases, you
may find it easier to use a labelled bracketing showing only key parts
of the structure. Treat imperative clauses containing and as coordinate
imperative root clauses, and analyse only clauses preceding/following
and. For the purposes of this exercise, treat nominals following articles/
determiners as NPs (e.g. treat nominals like lime juice/large bowl/tortilla
 7.8 Workbook 419

chips as NPs) and don’t concern yourself with their internal structure. In
relation to sentences like 4, 7, 8) bear in mind the following constraint
due to Ross (1967):

13 Adjunct Island Constraint

Adjuncts are islands – that is, they are structures out of which nothing
can be extracted by any movement operation that leaves behind a gap
(i.e. a null copy of the moved constituent)

In relation to 6 and 7, consider the implication of taking a phrase like

let the birds stand for 10 minutes to involve a structure in which the
transitive verb let takes a to-less infinitive complement – as below
(where the sentence is paraphrasable as ‘Ensure that the birds stand for
10 minutes’):

14 Let [the birds/them stand for 10 minutes]

Assume that the italicised infinitive subject in 14 is the thematic

argument of stand (its complement if stand is an unaccusative predicate),
but is assigned accusative case under c‑command by the transitive verb
let. In addition, bear in mind the discussion in the model answer in the
Answerbook. In 8 consider whether if using is a reduced form of the finite
clause if you are using ø, or whether it is a gerund clause with a PRO
subject (if PRO using ø) – where ø is a null object of some kind.
 Glossary and Abbreviations

Note: Bold print is used for headings, subheadings, and cross-references to other
entries; §3.4.2 denotes Chapter 3, Module 4, Section 2.

A/AP Abbreviation for adjective/ if it carries some unvalued

adjectival phrase. φ-feature/s, and becomes
A‑bar position/A‑bar Movement See inactive once its φ-feature/s have
A‑position/A‑Movement. been valued (Probe Condition)
abbreviated Abbreviated structures (ii) An XP can only be an
are structures in which certain active goal for Agreement,
functional categories are missing: A‑Movement, or Case-marking
for example, a newspaper headline if it has an unvalued case
like <The> president <is> intending feature, and becomes inactive
to host <a> peace conference, where once its case feature is valued
bracketed articles and a bracketed (Goal Condition)
auxiliary are missing. Abbreviated (iii) If a constituent is part of a
registers are varieties which make movement chain at a given
systematic use of abbreviated point in a derivation, only the
structures. See Chapter 7. highest copy in the chain can
Acc Abbreviation for accusative. See be active for operations like
case. Agreement, A‑Movement or
accessible A constituent is accessible Case-marking (Highest Copy
to a given operation if it is able Condition)
to undergo the relevant operation
Constituents which do not satisfy
without violating any constraints or
these conditions are inactive for
other requirements.
the relevant operations. On a
A‑chain See A‑position.
different use of active in relation
active A constituent is active for
to active voice, see passive.
(i.e. able to undergo) Agreement,
adjoin See adjunct.
A‑Movement, or Case-marking if it
adjunct/adjunction An adjunct
satisfies a set of Activity Conditions
is a constituent which adjoins/
which are given the following
attaches to another to form a larger
formulation in §3.1.1:
constituent of the same type via
(i) A head can only be an active an operation termed adjunction.
probe for Agreement or A head adjunct is a head which
A‑Movement or Case-marking adjoins to another head to form a
 Glossary and Abbreviations 421

complex head. A phrasal adjunct [usually ASPhabitual

is a phrase (more accurately: [again ASPrepetitive(I)
maximal projection) which adjoins [often ASPfrequentative(I)
to another phrase to form an even [intentionally MODvolitional
larger phrase of the same type: [quickly ASPcelerative(I)
for example, a VP adjunct is a [already Tanterior
constituent which adjoins to a VP [no longer ASPterminative
to form a larger VP; and a T-bar [still ASPcontinuative …
adjunct is a constituent which
The full hierarchy is listed in (36) of
adjoins to a T‑bar to form an even
§6.2.2.
larger T‑bar: see §1.2.3. The Adjunct
ADVP See ADV.
Island Constraint specifies that
Af See Affix.
adjuncts are islands which don’t
affix/affixal The term affix
allow anything to be extracted out
(abbreviated to Af) is typically
of them.
used to describe a grammatical
ADV The abbreviation ADV denotes
morpheme which cannot stand on
an adverb, and ADVP denotes an
its own as an independent word,
adverbial phrase. Some linguists
but which must be attached to a
treat certain ADVP constituents as
host word of an appropriate kind.
PPs headed by a null preposition:
An affix which attaches to the
for example, then is traditionally
beginning of a word (e.g. un- in
taken to be an ADVP, but might be
unhappy) is called a prefix: an
a PP with a more abstract structure
affix which attaches to the end of
paraphrasable as ‘at that time’.
a word (e.g. -s in chases) is called a
adverb hierarchy This refers to a
suffix. An affixal head is one which
template/schema used by Cinque
behaves like an affix in needing to
(1999) to show the relative
attach to a particular kind of host
ordering within clauses of different
word. See also clitic. Affix Hopping
functional projections housing
is an operation in the PF component
adverbs. Part of the hierarchy is
by which an unattached affix in T
listed below (where higher adverbs
is lowered onto the closest auxiliary
precede lower ones):
or non-auxiliary verb below T: see
[frankly MOODspeech act §1.3.1.
[fortunately MOODevaluative agent This is a term used to describe

[allegedly MOODevidential the semantic (= thematic) role which

[probably MODepistemic a particular type of argument plays
[once Tpast in a given sentence. It typically
[then Tfuture denotes a person who wilfully
[perhaps MOODirrealis causes some state of affairs to come
[necessarily MODnecessity about: hence for example, John
[possibly MODpossibility plays the thematic role of ag e nt
422 Glossary and Abbreviations

argument of the verb smashed in a but which must be bound by

sentence such as ‘John deliberately (i.e. take its reference from) an
smashed the bottle.’ See §2.2.2. appropriate antecedent that
agreement An operation by which c‑commands it within the same
(e.g. in a sentence like They are phrase or sentence. Hence, while
lying) the person/number features we can say John is deluding himself
of the auxiliary are come to be (where himself refers back to John),
assigned the same values as those we cannot say *For himself to have
of its subject they, so that are is to wait would be unthinkable, since
third person plural because it agrees the anaphor himself here has no
in person and number with its antecedent.
third person plural subject they. animate The term animate is used to
A language has rich agreement if denote (the gender of) an expression
it uses a wide range of distinctive which denotes a living being (e.g. a
person/number inflections on verbs, human being or animal), while the
and poor agreement if it does not: term inanimate is used in relation
see Chapter 3. to an expression which denotes
Agreement Deficit Hypothesis This is lifeless entities. For example, the
an account of the syntax of Be Drop relative pronoun who could be said
in newspaper headlines (outlined to be animate in gender and the
in Module 7.4) which makes the relative pronoun which inanimate –
following assumptions: hence we say someone who upsets
people and something which upsets
e Drop clauses are truncated
people.
SUBJP constituents, headed by a
antecedent An expression which is
SUBJ with defective agreement
referred to by another constituent
properties (e.g. with number but not
(typically a pronoun or anaphor)
person). SUBJ attracts be to raise
of some kind. For example, in John
to adjoin to it, but (being defective)
cut himself shaving, John is the
cannot attract the subject to move
antecedent of the anaphor himself,
to spec‑SUBJP; and since the clause
since himself refers back to John. In
lacks FIN, its subject cannot be
a sentence such as He is someone
assigned nominative case. Instead,
who we respect, the antecedent
the subject raises only as far as
of the relative pronoun who is
spec‑TP, and is assigned accusative
someone.
case by default. e receives a null
Antilocality Constraint A constraint
spellout by default at PF, because
which (in the formulation of Boeckx
there is no overt spellout for
(2007: 110) specifies that ‘Movement
defective forms of be .
internal to a projection counts as
A‑Movement See A‑position. too local, and is banned.’ This rules
anaphor This is an item (like the out (for example) movement from
reflexive anaphor himself) which complement to specifier position
cannot have independent reference, internally within a given phrase.
 Glossary and Abbreviations 423

AP Abbreviation for adjectival phrase. A‑bar chain is a movement chain

A‑position/A‑Movement The difference formed by a constituent undergoing
between an A‑position and an A‑bar A‑bar Movement; an A‑bar head
position can be illustrated in relation is the kind of head (like C) in the
to a clause such as that produced by clause periphery which allows as its
speaker b below: specifier either an argument or an
adjunct expression.
s p ea k e r a : What don’t you
argument This is a term borrowed
understand?
by linguists from philosophy
s p ea k e r b : [CP Why [TP you are
(more specifically, from predicate
blaming yourself]]
calculus) to describe the role played
An A‑position is one which can only by particular constituents in the
be occupied by an argument (not semantic structure of sentences. In
by an adjunct), which can control a sentence such as ‘John hit Fred’,
agreement, and from which an the overall sentence is said to be a
anaphor can be bound; by contrast, proposition (a term used to describe
an A‑bar position is one which the semantic content of a clause),
is not restricted to arguments (so and to consist of the predicate hit
can contain either an adjunct or and its two arguments John and
an argument), but cannot control Fred. The two arguments represent
agreement or bind an anaphor. the two participants in the act of
Given these criteria, the spec‑TP hitting, and the predicate is the
position occupied by you in speaker expression (in this case the verb hit)
b ’s utterance above is an A‑position which describes the activity in which
because the pronoun you that it they are engaged. By extension,
contains is an argument (of the verb in a sentence such as ‘John says
blaming), it controls/determines the he hates syntax’ the predicate in
agreement properties of the auxiliary the main clause is the verb says,
are, and it serves as the antecedent and its two arguments are John
of the anaphor yourself. By contrast, and the clause he hates syntax; the
the spec‑CP position occupied by argument he hates syntax is in turn a
why is an A‑bar position because proposition whose predicate is hates,
why is an adjunct, does not control and whose two arguments are he
agreement with are, and cannot and syntax. Since the complement
bind an anaphor like yourself. of a verb is positioned internally
A‑Movement is movement to an within V‑bar whereas the subject of
A‑position in the subperiphery (like a verb is positioned outside V‑bar,
spec‑TP or spec‑SUBJP); an A‑chain complements are also referred to as
is a movement chain formed by a internal arguments, and subjects as
constituent undergoing A‑Movement. external arguments. The argument
A‑bar Movement is movement to an structure of a predicate provides a
A‑bar position (in effect, a specifier description of the set of arguments
position in the clause periphery); an associated with the predicate, and
424 Glossary and Abbreviations

the theta role which each fulfils in example, CP is replaced by a set

relation to the predicate. A one-/ of functional projections including
two-/three-place predicate is a FORCEP, FINP and a range of
predicate which has one/two/three (optional) additional projections such
arguments. See Module 2.2. as TOPP, FOCP, and MODP.
ART(icle)/ARTP The term article is aspect A term typically used to
used in traditional grammar to denote the duration of the activity
describe a particular subclass of described by a verb (e.g. whether the
determiner/quantifier: the word the activity is ongoing or completed). In
is traditionally called the definite sentences such as:
article, and the word a(n) is termed
(i) He has taken the medicine
the indefinite article. In this book,
(ii) He is taking the medicine
I treat a phrase like a book as an
ARTP/article phrase which is headed the auxiliary has is said to be an
by the indefinite article, while auxiliary which marks perfect aspect
treating a phrase like the book as a (and hence to be a perfect auxiliary)
determiner phrase/DP headed by the in that it marks the perfection
determiner the. Article Drop denotes (in the sense of ‘completion’ or
a phenomenon whereby definite and ‘termination’) of the activity of
indefinite articles can be omitted taking the medicine; for analogous
in certain abbreviated registers of reasons, taken is said to be a perfect
English – for example, newspaper participle verb form in (i) (though is
headlines: it is characterised in sometimes referred to in traditional
Module 7.3 as involving an article grammars as a ‘past participle’).
which is the first word in the (DP/ Similarly, is functions as an auxiliary
ARTP) projection that it heads which marks progressive aspect in
being given a null spellout. The (ii), because it relates to an activity
C‑command Condition on Article which is ongoing or in progress; for
Drop (devised by Stowell 1999) this reason, is in (ii) is also referred
specifies that no nominal allows to as a progressive auxiliary. In
Article Drop if it is c‑commanded by the same way, the verb taking in
a nominal with an overt article: see (ii) is said to be the progressive
Module 7.3 and Module 7.4. participle form of the verb (though is
articulated Work in the Cartographic sometimes referred to in traditional
model (outlined in Chapters 4–6) sees grammars as a ‘present participle’).
clauses as having a more articulated associate An expression which
(i.e. more complex) structure represents the thematic argument
than assumed in the traditional in an expletive there construction,
CP+TP+VP model of clause structure. and which is associated with the
Hence, the Cartographic model expletive subject there: for example,
contains more functional projections an unfortunate accident in There
than the traditional model: for occurred an unfortunate accident.
 Glossary and Abbreviations 425

attract To say that a head H attracts being the head of an Auxiliary

a constituent C is to say that H Projection below T to adjoin to
triggers movement of C to some an unattached tense affix in T,
position on the edge of HP (so that as in §6.1.1; or moving from T
C may move to adjoin to H, or to into a higher position within the
become the specifier of H). subperiphery, as in §6.2.1).
attribute See value. bar When used as a suffix attached
AUX(iliary)/AUXP The term auxiliary to a category label such as N,
is used to categorise items such V, P, T etc. (as in N-bar, V‑bar,
as will/would/can/could/shall/ P-bar, T‑bar etc.), it denotes an
should/may/might/must/ought and intermediate projection which is
some uses of have/be/do/need/ larger than a word but smaller than
dare. Such items have a number of a phrase. Hence, in a phrase such
idiosyncratic properties, including as university policy on drugs, we
the fact that they can undergo might say that the string policy
(Subject–Auxiliary) Inversion (e.g. on drugs is an N-bar, since it is a
in questions like ‘Can you speak projection of the head noun policy,
French?’). An auxiliary projection/ but is an intermediate projection
AUXP is a phrase headed by an in that it has a larger projection
auxiliary: see §6.1.1. Auxiliary into the NP university policy on
Drop (see Module 7.2) is a drugs. For another use of bar in
phenomenon found in abbreviated the phrase ‘A‑bar position’, see
registers of English whereby an A‑position.
auxiliary can be omitted if it is bare A bare noun phrase/nominal
the first word in a sentence – as in is one used without any overt
<Are> you doing anything tonight? quantifier or determiner to modify
Auxiliary selection relates to the it (e.g. fish in Fish is expensive). A
type of verb which a given auxiliary bare clause is one not introduced
selects as its complement: for by an overt complementiser (like
example, in many languages (the he was tired in John said he was
counterpart of) be when used as tired).
a perfect auxiliary selects only a barrier A given structure is a barrier
complement headed by a verb with to movement if it doesn’t (readily)
no external argument, whereas allow other constituents to be
(the counterpart of) have selects a extracted/moved out of it. For
complement headed by a verb with example, the bracketed if-clause
an external argument. See §2.3.2. below is a barrier to extraction of
Auxiliary Inversion See Inversion. how in (ii):
Auxiliary Raising An operation by (i) She will be annoyed [if he
which an auxiliary verb moves from behaves badly]
a lower to a higher head position (ii) *How will she be annoyed [if he
within a clause (e.g. moving from behaves — ]?
426 Glossary and Abbreviations

In (ii), how originates in the gap bottom-up To say that a syntactic

position – that is, in the same structure is derived in a bottom-up
position as badly occupies in (i) – fashion is to say that the structure
and from there moves out of the is built up from bottom to top, with
bracketed if-clause to the front of lower parts of the structure being
the overall sentence. However, the formed before higher parts: see
if-clause is an adjunct to the main §1.2.2.
clause, and adjuncts are barriers to C Abbreviation for complementiser.
(i.e. resist) extraction, with the result See complementiser.
that (ii) is ungrammatical. canonical A term used to
B e Drop A phenomenon found in mean ‘usual’, ‘typical’ or
certain abbreviated registers of ‘normal’, as in ‘The canonical
English (especially newspaper word order in English is
headlines) involving omission specifier+head+complement.’
of indicative forms of be – as in Cartography/Cartographic An
Minnesota <is> hit by more unrest approach to syntax (associated with
(with omission of the bracketed the work of Luigi Rizzi and his
auxiliary <is>). See Module 7.4. collaborators) which aims to devise
binary A binary contrast is a two- a detailed ‘map’ of various types of
way contrast (e.g. between a syntactic structure (e.g. the clause
singular noun like man and its periphery and subperiphery). The
plural counterpart men). A binary- Cartographic model sees syntactic
branching structure/tree diagram structures as containing many more
is one in which each non-terminal functional projections than the
constituent (i.e. each constituent traditional CP+TP+VP model. See
not at the foot/bottom of the tree) Chapters 4–6.
branches down into two other case The different case forms of a
constituents immediately beneath it: pronoun are the different forms
see §1.2.2. which the pronoun has in different
bind/bound To say that one constituent sentence positions. It is traditionally
X binds another constituent Y (and said that English has three cases –
conversely that Y is bound by X) is nominative (abbreviated to NOM),
to say that X determines properties accusative (= ACC, sometimes
(usually, referential properties) of Y. also referred to as objective),
For example, in a sentence such as and genitive (= GEN). Personal
John blamed himself, the reflexive pronouns typically inflect overtly
anaphor himself is bound by John for all three cases, whereas noun
(equivalently, John binds himself), expressions overtly inflect only for
in the sense that the referential genitive case. The different case
properties of himself are determined forms of typical pronouns and noun
by John (i.e. himself refers back to expressions are given in the table
John). below:
 Glossary and Abbreviations 427

nominative I we you he she it they who the king

accusative me us you him her it them who(m) the king
genitive my our your his her its their whose the
mine ours yours hers theirs king’s

As is apparent, some pronouns have not fall within the domain of any
two distinct genitive forms: a weak case assigner are assigned case
(shorter) form used when they are by a Default Case Assignment
immediately followed by a noun operation which is given the
(as in ‘This is my car’), and a strong following formulation in §4.1.2:
(longer) form used when they are
A (pro)nominal which does not fall
not immediately followed by a
within the domain of (i.e. which
noun (as in ‘This car is mine’). The
is not c‑commanded by) a local
null subject PRO found in control
(i.e. nearby) case assigner receives
clauses is treated in this book as
accusative case by default.
carrying null case: see §1.3.2.
case particle Many linguists take of
case-marking Under the account of
in structures like destruction of the
case marking in §3.1.2 (revised
city or fond of him to belong to
in §4.3.2 by replacing C by FIN),
the category K of ‘case particle’,
nominal and pronominal XP
and the of-phrase to be a KP/case
constituents are assigned case by
phrase. The rationale for this is
an appropriate case assigner in
that of has to be used here because
accordance with the following Case
a noun like destruction and an
Conditions: A case-assigning probe
adjective like fond are not case
values an unvalued case feature on
assigners, and hence the nominal
an XP goal and assigns the goal
the city and the pronominal him
a case value that depends on the
would be unable to be case-marked
nature of the probe – namely:
if the case particle of were not used
(i) nominative if the probe is a finite to assign accusative case to them.
C/FIN with a TP complement The assumption here is all that
(where a finite C/FIN is one that (pro)-nominal constituents need to
is indicative, subjunctive or be assigned case in order to receive
imperative in mood) a spellout in the PF component: for
(ii) null if the probe is an example, the PF component needs
intransitive infinitival C/FIN to know whether a third person
with a TP complement singular masculine pronoun has
(iii) accusative if the probe is nominative, accusative or genitive
transitive case in order to determine whether
to spell it out as he, him or his.
(Pro)nominal constituents which c‑command In a structure containing
are in a position where they do two different constituents X
428 Glossary and Abbreviations

and Y, X c‑commands Y if X is and so is said to function as a

independent of Y (i.e. if neither complement clause in this type of
contains the other), and if the sentence.
mother of X contains Y. The clause type/Clause Typing
C‑command Condition on Article Conditions A clause is said to be
Drop in newspaper headlines is a declarative in type if used to make
condition (devised by Stowell 1999 a statement, interrogative in type if
and discussed in Module 7.3) which used to ask a question, imperative
specifies that no nominal allows in type if used to issue an order,
Article Drop if it is c‑commanded etc. The Clause Typing Conditions
by a nominal with an overt article; outlined in §1.4.3 specify that:
this condition is generalised in
(i) A clause is interpreted as being
Module 7.4.
of a given type (interrogative or
CED Abbreviation for Constraint on
exclamative etc.) if it contains
Extraction Domains: see extract.
a peripheral clause-typing
chain A set comprising a constituent
specifier of the relevant type
and any traces/copies associated
(ii) An indicative clause is
with it.
interpreted as declarative in
clause A clause is defined in
type by default if it contains
traditional grammar as an
no peripheral clause-typing
expression which contains a subject
specifier (where peripheral
and a predicate, and which may
means ‘within the clause
contain other types of constituent as
periphery’ – that is, within that
well (e.g. one or more complements
part of the clause structure
and/or adjuncts). In most cases,
positioned above the subject).
the predicate in a clause contains
a lexical (= non-auxiliary) verb, so cleft sentence A structure such as ‘It
that there will be as many different was syntax that he hated most’,
clauses in a sentence as there are where syntax is said to occupy focus
different lexical verbs. For example, position within the cleft sentence.
in a sentence such as She may think clitic(isation) The term clitic denotes
that you are cheating on her, there an item which is a reduced form
are two lexical verbs (think and of another word, and which has
cheating), and hence two clauses. the property that (in its reduced
The cheating clause is that you form) it must cliticise (i.e. attach
are cheating on her, and the think itself to) an appropriate kind of
clause is She may think that you are host word. For example, we could
cheating on her, so that the cheating say that the contracted negative
clause is one of the constituents of particle n’t is a clitic form of the
the think clause. More specifically, negative particle not, and that the
the cheating clause is the clitic n’t attaches itself to a finite
complement of the think clause, auxiliary verb, so giving rise to
 Glossary and Abbreviations 429

forms like isn’t, shouldn’t, mightn’t follow their heads in English.

etc. When a clitic attaches to the The choice of complement (and
end of another word, it is said to be the morphological form of the
an enclitic (and to encliticise to the complement) is determined by
relevant word). A clitic differs from properties of the head: for example,
an affix in that a clitic is generally an auxiliary such as will requires
a reduced form of a full word, and as its complement an expression
has a corresponding full form (so headed by a verb in the infinitive
that ’ll is the clitic form of will, for form (cf. ‘He will go/*going/*gone’).
example), whereas an affix (like Moreover, complements bear a
noun plural -s in cats) has no full- close semantic relation to their
word counterpart. Clitic doubling heads (e.g. in Kill him! the word
is a phenomenon which results in him is the complement of the
a structure containing a clitic copy verb kill and is an argument of
of some other constituent (e.g. of a the verb kill, representing the
subject or object): see §6.1.2. victim). Thus, a complement has
COMP/comp The label ‘COMP’ is an a close morphological, syntactic
abbreviation for complementiser. and semantic relation to its head.
By contrast, ‘comp’ is an A complement clause is a clause
abbreviation for complement, so which is used as the complement of
comp‑VP denotes ‘the complement some other word (typically as the
position in VP’. complement of a verb, adjective or
complement This is a term used to noun). Thus, in a sentence such as
denote a specific grammatical He never expected that she would
function (in the same way as come, the clause that she would
the term subject denotes a come serves as the complement
specific grammatical function). of the verb expected, and so is a
A complement is a constituent complement clause.
which is directly merged with (and complementiser A particular category
hence is the sister of) a head word, of clause-introducing word like
thereby projecting the head into a that/if/whether/for, as used in
larger structure of essentially the sentences like ‘I think [that you
same kind. For example, in close the should apologise]’, ‘I don’t know [if/
door, the door is the complement whether she will agree]’, ‘They’re
of the verb close; in after dinner, keen [for you to show up]’ where
dinner is the complement of the the italicised complementisers
preposition after; in good at physics, introduce the bracketed clauses. A
at physics is the complement of Complementiser Phrase/projection
the adjective good; in loss of face, (CP) is a clause headed by
of face is the complement of the a complementiser. The term
noun loss. As these examples complementiser is abbreviated
illustrate, complements typically to COMP in earlier work and to
430 Glossary and Abbreviations

C in later work. Complementiser conditions – for example, ‘If you

agreement is agreement between a should upset anyone, I’ll bar you’,
complementiser and the subject of or ‘Should you upset anyone, I’ll
its clause (this being marked overtly bar you’, where the italicised are
in certain varieties of Flemish, conditional clauses.
German and Dutch): see §3.1.2. CONJ/CONJP/Conjunction The term
complete clause A clause which has a conjunction denotes a word used
periphery that includes a FORCEP to join two or more constituents
projection. See §4.4.1. together. In a sentence like Although
Completeness Condition See Feature I was tired, I stayed up to watch
Valuation Conditions. the film, although is a conjunction
component A grammar is said to which links the subordinate clause I
have three main components: a was tired to the main clause I stayed
syntactic component which generates up to watch the film and so is
syntactic structures, a semantic/LF termed a subordinating conjunction
component which assigns each such (abbreviated to SUB in some works,
syntactic structure an appropriate but to CONJ here, in order to avoid
semantic/LF representation (= a confusion with SUBJ/subject);
representation of its logical form), consequently, subordinate clauses
and a phonological/PF component are treated as CONJP constituents
which assigns each syntactic structure here (but as SUBP constituents in
generated by the syntactic component some other works). In a sentence
an appropriate phonetic form/PF. like He stayed up and/but she
COMP‑Trace Filter A condition went to bed, the words and/but are
(introduced in §1.5.2) specifying coordinating conjunctions which
that any superficial structure in conjoin/link the two main clauses
which an overt complementiser he stayed up and she went to bed.
immediately precedes and constraint A grammatical principle
c‑commands a trace (i.e. a gap/ which prevents certain types
null copy left behind by a moved of grammatical operation from
constituent) is filtered out as applying to certain types of
ungrammatical at the end of the structure. The terms condition and
syntactic derivation. principle are used in much the same
concord A traditional term to describe way. See also specific constraints
an operation whereby a noun (e.g. the Constraint on Extraction
and any adjectives or determiners Domains, listed under Extract/
modifying it are assigned the same Extraction).
values for features such as number, content This term is generally used to
gender and case. refer to the semantic content (i.e.
condition See constraint. meaning) of a constituent (typically,
conditional A term used to represent of a word). However, it can also
a type of clause which sets out be used in a more general way to
 Glossary and Abbreviations 431

refer to the linguistic properties of copula/copular verb A ‘linking verb’,

an expression: for example, the used to link a subject with a non-
expression phonetic content is used verbal predicate. The main copular
to refer to the phonetic form of (e.g.) verb in English is be (though verbs
a word: hence, we might say that like beco e , re ain , stay etc. have
PRO is a pronoun which has no much the same linking function).
phonetic content (meaning that it is a In sentences such as They are lazy,
silent pronoun with no audible form). They are fools and They are outside,
contraction A process by which two the verb are is said to be a copula
different words are combined into (or copular verb) in that it links the
a single word, with either or both subject they to the adjectival predicate
words being reduced in form. For lazy, or the nominal predicate fools, or
example, by contraction, want to the prepositional predicate outside.
can be reduced to wanna, going to copy The Copy Theory of Movement
to gonna, he is to he’s, they have to is a theory developed by Chomsky
they’ve, did not to didn’t, etc. See which maintains that a moved
also clitic(isation). constituent leaves behind a copy
control/controller In a nonfinite of itself (called a trace) each time
clause with a PRO subject which it moves, with the copy generally
has an antecedent, the antecedent being given a null/silent spellout
is said to be the controller of PRO (and so being unpronounced) in
(or to control PRO), and conversely the phonology. Copy Raising is
PRO is said to be controlled by a type of movement operation in
its antecedent; and the clause which a raised constituent leaves
containing PRO is termed a control an overt copy behind (e.g. in ‘There
clause. So, in a structure like John look like there are no objections’,
tried [PRO to quit], the bracketed where expletive there raises from
structure is a control clause, John being subject of the are-clause to
is the controller of PRO, and becoming subject of the look-clause,
conversely PRO is controlled by and leaves an overt copy of itself
John (in the sense that PRO refers behind). See §3.4.2 and §3.4.3.
back to John). The term control coreferential Two expressions are
predicate/verb denotes a word coreferential if they refer to the
like try which takes an infinitive same entity. For example, in John
complement with a (controlled) PRO cut himself while shaving, himself
subject. See §2.5.2. and John are cofererential in the
coordinate A coordinate structure is sense that they refer to the same
a structure containing two or more individual.
expressions joined together by a count/countable A count(able) noun
coordinating conjunction such as is a noun which can be counted.
and/but/or/nor (e.g. John and Mary Hence, a noun such as chair is a
is a coordinate structure). count noun since we can say ‘One
432 Glossary and Abbreviations

chair, two chairs, three chairs, …’; (or semantic type) of a clause
but a noun such as furniture is a which is used to make a statement
non-count/uncountable/mass noun (as opposed to a clause which is
since we cannot say ‘*one furniture, interrogative, exclamative or
*two furnitures, …’. imperative in force).
CP Complementiser Phrase. See dedicated To say that an adverb like
Complementiser. On CP recursion, perhaps serves as the specifier of a
see recursion. dedicated functional head (a head
crash A derivation is said to crash expressing irrealis mood, according
if it violates some constraint/s or to Cinque 1999: see §6.2.2) is to say
interface condition/s, or results in that it is the specifier of a functional
a structure which can’t be assigned head which only allows adverbs
a phonetic form and/or a semantic of the same type as perhaps as its
interpretation. For example, if the specifier.
person or number features of have default A default value or property is
remain unvalued in a sentence one which obtains if all else fails
such as ‘He have left’, the resulting (i.e. if other conditions are not
sentence will crash at the phonetics satisfied). For example, if we say
interface, since the PF component that +d is the default past tense
will be unable to determine whether inflection in English, we mean that
h av e (if present tense) should be verbs which don’t have an irregular
spelled out as have or has. past tense form will form their past
Criterial Freezing Condition A tense by adding +d to their stem. On
condition which is given the Default Case Assignment, see case
following formulation in Module marking. On the Default Spellout
4.1 (with the italicised amendment Rule, see spellout rules.
added in §6.1.2): definite Expressions containing
determiners like the, this, that etc.
(i) A peripheral or subperipheral
are said to have definite reference
constituent must occupy
in that they refer to an entity
a criterial position to be
which is assumed to be known to
interpretable at LF
the addressee(s): for example, in a
(ii) A constituent which occupies
sentence such as ‘I hated the course’,
its criterial position is frozen in
the DP the course refers to a specific
place
course whose identity is assumed
cross-clausal Cross-clausal movement/ to be known to the hearer/reader.
agreement is an operation that takes In much the same way, personal
place across an intervening clause pronouns like he/she/it/they etc. are
boundary. See Module 3.4. said to have definite reference. By
D An abbreviation for Determiner. contrast, expressions containing the
declarative A term used as a article a are indefinite, in that (for
classification of the force example) if you say ‘I’m taking a
 Glossary and Abbreviations 433

course’, you don’t assume that the its complement of Utopia and its
hearer/reader knows what course specifier the. Since Abney (1987),
you are taking. such expressions have been taken
degraded A sentence is said to to have the status of DP/determiner
be degraded if it is less than phrase.
fully acceptable/less than fully DFCF See Doubly Filled COMP Filter.
grammatical. Diary Drop A register of written
deletion A type of operation by English used in diary entries
which a constituent present in the whereby sentences are abbreviated
syntactic structure of a sentence by omitting various types of
is given a silent spellout (i.e. is constituent, most notably subjects.
unpronounced) in the phonological/ See Module 7.1, and Exercise 7.1 of
PF component. the Workbook module in Chapter 7.
derive/derivation The derivation direct object See object.
of a phrase or clause is the set of discontinuous spellout See spellout.
syntactic (e.g. merge, movement, discourse Discourse factors are factors
adjunction, agreement, case relating to the extrasentential
marking etc.) operations used to setting in which an expression
form the relevant structure. To occurs (where extrasentential means
derive a structure is to say how it is ‘outside the immediate sentence
formed (i.e. specify the operations containing the relevant expression’).
by which it is formed). For example, to say that PRO has a
determiner/D A word like the/this/ discourse controller in a sentence
that used to modify a noun or noun such as It would be wise PRO to
phrase is a determiner in the sense prepare yourself for the worst
that it determines the referential means that PRO has no antecedent
properties of the noun (phrase). For within the sentence immediately
example, the in a sentence like Shall containing it, but rather refers
we take the car? serves to indicate to some individual(s) outside the
that the phrase the car is a definite sentence (in this case, the person
referring expression, in the sense being spoken to).
that it refers to a definite (specific) dislocate/dislocation In sentences
car which is assumed to be familiar such as:
to the hearer/addressee.
(i) The mayor of Trumpton, I doubt
Determiner Phrase/DP A phrase like
if he will win another term
the king of Utopia which comprises
(ii) I doubt if he will win another
a determiner the and a noun phrase
term, the mayor of Trumpton
complement king of Utopia. In work
before the mid-1980s, a structure the italicised constituent the mayor
like the king of Utopia would have of Trumpton is said to be dislocated,
been analysed as a noun phrase (= because it is in a separate intonation
NP), comprising the head noun king, group (set off by a comma) from
434 Glossary and Abbreviations

the rest of the sentence. Where the some?’, ‘He doesn’t want any’, and
dislocated constituent is to the left ‘He wants some, does he?’
of the rest of the sentence as in Doubly Filled COMP Filter A
(i), the phenomenon is termed Left condition specifying that at the
Dislocation; where the dislocated end of a syntactic derivation, any
constituent is to the right of the rest structure in which the edge of
of the sentence as in (ii), it is termed a projection headed by an overt
Right Dislocation. See §4.1.2. complementiser is doubly filled
Distinctness Condition A condition (i.e. contains some other overt
devised by Richards (2010: 5), constituent) is filtered out as
which is characterised informally as ungrammatical: see §1.5.2.
follows in §4.1.3: DP Abbreviation for determiner
phrase.
Any structure at PF which contains
DP Hypothesis The hypothesis that all
a sequence of two non-distinct
definite nominal and pronominal
constituents <α, α> is degraded
arguments have the status of DPs –
The condition marks the sequence not just nominals like the president
who who as degraded in a which contain an overt determiner,
sentence such as *Who who was but also proper names like John,
innocent would behave like that? and pronouns like he.
distribution The distribution of a ECM See Exceptional Case Marking.
constituent denotes the range of edge The edge of a given projection
positions that it can occupy in XP comprises the constituent(s)
phrases/sentences. which are contained in XP but are
domain The domain (or, more not c‑commanded by X (i.e. by the
fully, c‑command domain) of a head of XP). Typically, the edge of
head H is the set of constituents XP will include its head X and any
c‑commanded by H – namely specifier it may have.
its sister and all the constituents ellipsis/elliptical Ellipsis is an
contained within its sister. For operation by which an expression is
example, the domain of C includes omitted in order to avoid repetition.
its TP complement and all In this book, an ellipsed constituent
constituents of the relevant TP. is taken to be present in the syntax,
D o -support This refers to the use of but given a silent pronunciation in
the dummy/meaningless auxiliary the PF component. For example, in
d o to form questions, negatives a sentence such as I will do it if you
or tags in sentences which would will do it, we can ellipse (i.e. give a
otherwise contain no auxiliary. silent pronunciation to) the second
Hence, because a non-auxiliary verb occurrence of the verb phrase/
like want requires o -support in VP do it to avoid repetition, and
questions/negatives/tags, we have hence say I will do it if you will:
sentences such as ‘Does he want this type of ellipsis is referred to as
 Glossary and Abbreviations 435

VP ellipsis. An elliptical structure §3.1.1 specifies that an EPP feature

is one containing an ‘understood’ on a probe T attracts the closest
constituent which has undergone active XP goal that it agrees with
ellipsis (i.e. has been given a silent in person to undergo A‑Movement
pronunciation). Sentence-Initial to spec‑TP, and the EPP feature
Ellipsis is a phenomenon found is stripped from T (and thereby
in abbreviated registers/styles of removed from the derivation) once
English whereby a constituent can A‑Movement has taken place.
be omitted at the beginning of a Exceptional Case Marking/clause/
sentence, as with <I> couldn’t care ECM Accusative subjects of
less, where the <bracketed> subject infinitival complement clauses
I is ellipsed: see Module 7.1. (like him in the clause italicised
embedded An embedded clause is one in ‘I believe him to be innocent’)
which is contained internally within are said to carry exceptional case,
another constituent. For example, in in the sense that the case of the
a sentence such as He may suspect accusative subject is assigned by
that I hid them, the hid-clause (= the main clause verb believe, and
that I hid them) is embedded within it is exceptional for the case of the
(and is the complement of) the verb subject of one clause to be assigned
phrase headed by the verb suspect. by the verb in a higher clause.
Likewise, in The fact that he didn’t Transitive verbs (like believe) that
apologise is significant, the that- take an infinitival TP complement
clause (that he didn’t apologise) is with an accusative subject are said
an embedded clause in the sense to be ECM verbs. The infinitive
that it is embedded within a noun clause in such cases is said to be an
phrase headed by the noun fact. exceptional clause or ECM clause.
A clause which is not embedded See §3.3.2.
within any other expression is exclamative A type of structure
a root clause (see root) or main used to exclaim surprise, delight,
clause. annoyance, etc. In English syntax,
enclitic/encliticise See clitic(isation). the term is restricted largely
EPP This was originally an to clauses beginning with an
abbreviation for the Extended XP containing or comprising a
Projection Principle, which posited wh‑exclamative word like What! or
that every T constituent must be How! – for example, ‘What a fool
projected into a TP which has a I was!’ or ‘How blind I was!’ See
subject/specifier. In later work, the §5.4.1.
requirement for a T constituent like experien c e r A term used in the

will to have a specifier is said to analysis of semantic/thematic

be a consequence of T carrying an roles to denote an entity which
EPP feature requiring it to project experiences some emotional or
a specifier. The EPP Condition in cognitive state – for example,
436 Glossary and Abbreviations

John in John felt unhappy, or John extract/extraction Extract(ion) is

thought about his predicament. See another term for move(ment),
§2.2.2. and so denotes an operation by
expletive An expletive pronoun is one which one constituent is moved
with little or no inherent semantic out of another. For instance, in
content, such as there in sentences a structure such as Who do you
like There is no truth in the rumour, think [he saw — ] the pronoun
or impersonal it in sentences such who has been extracted out of the
as It is unclear why he resigned. bracketed clause (i.e. it is been
These pronouns are non-referential moved out of the gap position
and hence cannot be questioned by marked — ) and moved to the
where/what (cf. *Where is no truth front of the overall sentence.
in the rumour?/*What is unclear The extraction site for a moved
why he resigned?). In §3.2.3, the use constituent is the position which
of expletives is said to be subject to it occupied before undergoing
the following Expletive Conditions: movement (e.g. the gap position
in the above sentence). The
(i) External Argument Condition
Constraint on Extraction
An expletive can only be
Domains/CED (introduced in
merged as the highest argument
§1.5.1) specifies that extraction is
of a verb with no external
only possible out of a complement,
argument
not out of a specifier or adjunct.
(ii) Indefiniteness Condition
extrapose/Extraposition In a structure
Existential there can only
where a constituent which is part
be merged as the specifier
of a larger phrase is detached from
of an existential verb with
(and follows) the other constituents
an indefinite (pro)nominal
of the phrase, the detached
complement
constituent is said to have been
(iii) Inactivity Condition
extraposed or to have undergone
Expletive it can only be merged
Extraposition. So, for example, in:
with a constituent which does
not contain an active nominal (i) Scammers often ring me [who
or pronominal (i.e. one carrying claim to be from the bank]
an unvalued case feature)
the bracketed relative clause is
For detailed discussion of the syntax separated from its antecedent
of clauses with expletive subjects, scammers and positioned at the end
see Module 3.2. of the sentence (and so is said to
Extended Projection Principle See EPP. have been extraposed).
external argument See argument. On feature A device used to describe a
the External Argument Condition particular grammatical property. For
on the use of expletives, see example, the personal pronoun they
expletive. could be said to carry the features
 Glossary and Abbreviations 437

[third-person, singular-number, filter This is a device which rules

nominative-case]. See §1.1.3. out superficial syntactic structures
Feature Valuation Conditions A that fall foul of some grammatical
set of conditions which regulate requirement. For example, the
agreement, case-marking and Doubly Filled COMP Filter specifies
movement. These are given the that at the end of a syntactic
following formulation in §3.1.1: derivation, any structure in which
the edge of a projection headed by
(i) When an active probe like T
an overt complementiser is doubly
c‑commands an active XP
filled (i.e. contains some other
goal, each unvalued φ-feature
overt constituent) is filtered out as
on the probe will be valued by
ungrammatical. This filter rules out
the goal – that is, assigned a
the use of that in a clause like that
value which is a copy of the
bracketed in ‘I wonder [CP what
corresponding feature value on
[C ø/*that] he did].’
the goal (agreement)
FIN/FINP The abbreviation FIN
(ii) Only a φ-complete XP (i.e.
denotes an (overt or null) particle
one with person and number)
used to mark a clause as finite
can serve as a goal valuing
or nonfinite, like the infinitival
unvalued features on a
complementiser for in ‘She’s keen
probe, and conversely only a
for there to be a resolution.’ The
φ-complete probe (i.e. one with
italicised for-clause is analysed as
person and number) can value
a FINP/finiteness projection: see
unvalued features on a goal
§4.3.2.
(Completeness Condition)
finite The term finite verb/finite clause
(iii) Each feature on a constituent
denotes (a clause containing) an
must be valued (i.e. assigned
auxiliary or non-auxiliary verb
a single, appropriate value) in
which is inflected for tense/mood,
order for the constituent to be
and which can have a nominative
legible at the PF/LF interfaces;
subject like I/we/he/she/they. See
if not, the derivation will crash
§1.1.3.
(Legibility Condition)
floating quantifier A quantifier which
feminine This term is used in is separated from (and follows) the
discussion of grammatical gender to expression that it quantifies. For
denote pronouns like she/her/hers example, in a sentence such as ‘The
which refer to female entities. students should all have passed
filled To say that a given position in a their exams’, all quantifies (but is
structure must be filled is to say that separated from and follows) the
it cannot remain empty but rather students, so that all is a floating
must be occupied (usually by an quantifier here. On one view, all
overt constituent of an appropriate the students moves from being the
kind). subject of passed to becoming
438 Glossary and Abbreviations

the subject of have, and then the 4 and 5), such complementisers are
students moves on its own to said to constitute a FORCE head
become the subject of should, so which can project into a FORCEP/
leaving all stranded in front of have force projection.
(hence the term stranded quantifier formal A formal speech style denotes
is often used in place of floating a very careful and stylised form of
quantifier). See §6.4.2 and §6.4.3. speech (as opposed to the kind of
FOC/FOCP/focus(ed) A focused informal colloquial speech style used
constituent is one which is marked in a casual conversation in a bar).
(by its intonation/position) as Freezing Principle A constraint
containing new or unfamiliar specifying that the constituents of a
information. Focus can often be moved phrase are frozen internally
signalled by moving a constituent within (and so cannot be extracted
to the front of a clause or sentence, out of) the moved phrase. See §1.5.1.
as with the italicised constituents in front/fronting See preposing.
the sentences produced by speaker functional category/functional head/
b below: functor A word which has no
descriptive content and which
s p eaker a : What are the subjects
serves an essentially grammatical
you enjoy most and least?
function is said to be a function
s p eaker b : Syntax I enjoy most.
word or functor. A functional
Phonetics I enjoy least
category is a category whose
On one view, the italicised members are function words: hence,
constituents move to occupy the categories such as complementiser,
specifier position in a FOCP/focus auxiliary, infinitive particle, case
projection in the clause periphery, particle, and determiner are all
whose head FOC constituent is a functional categories – and so too
focus particle which is overt in are the expressions they head (e.g.
some languages, but null in English. C‑bar/CP, T‑bar/TP, D‑bar/DP etc.).
See Module 4.2. Functional Specifier Hypothesis A
foot The foot of a (movement) chain is hypothesis (derived from Cinque
the constituent which occupies the 1999) to the effect that every XP in
lowest position in the chain. the periphery or subperiphery of a
FORCE/FORCEP The complementisers clause is the specifier of a dedicated
that/if in a sentence such as I didn’t functional head. See §6.4.1.
know [that/if he was lying] are gap An empty position in a given
said to indicate that the bracketed structure which results from a
clauses are declarative/interrogative constituent undergoing ellipsis (as
in force (in the sense that they have with the second occurrence of ate in
the force of a question/a statement). John ate an apple and Mary ate a
In work on the clause periphery by pear), or movement (as with where
Luigi Rizzi (discussed in Chapters in Where has he gone where?).
 Glossary and Abbreviations 439

Gapping A form of ellipsis in which generic To say that an expression

the head word is omitted from like eggs in a sentence such as
one (or more) conjoined clauses ‘Eggs are fattening’ has a generic
in a coordinate structure in order interpretation is to say that it is
to avoid repetition. For example, interpreted as meaning ‘eggs in
the italicised second occurrence of general’.
bought can be gapped (i.e. omitted) genitive See case.
in a sentence such as ‘John bought gerund When used in conjunction
an apple and Mary bought a pear’, with the progressive aspect auxiliary
giving ‘John bought an apple, and be, verb forms ending in -ing are
Mary a pear.’ progressive participles; in other
gen In one use, an abbreviation uses they generally function as
for genitive case; in another, an gerunds (traditionally considered
abbreviation for gender. to be verbal nouns, by which is
gender A grammatical property meant ‘nouns derived from verbs’).
whereby words are divided into In particular, -ing verb forms are
different grammatical classes which gerunds when they can be used
play a role in agreement/concord as subjects, or as complements of
relationships. In French, for verbs or prepositions, and when
example, nouns are intrinsically (in literary styles) they can have
masculine or feminine in gender a genitive subject like my. Thus
(e.g. pommier ‘apple tree’ is writing is a gerund (verb form) in a
masculine, but pomme ‘apple’ sentence such as ‘She was annoyed
is feminine), and articles inflect at [my writing to her mother]’, since
for gender, so that un ‘a’ is the the bracketed gerund structure is
masculine form of the indefinite used as the complement of the
article, and une is its feminine form. preposition at, and has a genitive
Articles in French have to agree in subject my.
gender with the nouns they modify, goal /goal The term g oal is used in

hence we say un pommier ‘an apple the analysis of semantic/thematic

tree’, but une pomme ‘an apple’. In roles to denote the entity towards
contemporary English, nouns don’t which something moves – for
have inherent gender properties, example, Mary in ‘John sent Mary
and their modifiers don’t inflect a letter’: see §2.2.2. In a different
for gender either. Only personal sense, the term goal represents a
pronouns like he/she/it carry gender constituent which agrees with, or
properties in present-day English, is case-marked by, or is attracted
and these are traditionally said to by a higher head which serves as
carry masculine/feminine/neuter a probe: see §3.1.1. On the Goal
gender respectively (though the Condition, see active.
term inanimate is sometimes used head This term has two main uses. The
in place of neuter). head of a phrase is the key word
440 Glossary and Abbreviations

which determines the syntactic and 1998 and introduced in §1.5.1)

semantic properties of the phrase. specifying that anything below the
So, in a phrase such as fond of fast head C of a CP is impenetrable to
food, the head of the phrase is the (i.e. cannot undergo a syntactic
adjective fond, and consequently operation involving) anything above
the phrase is an adjectival phrase the CP. In §4.4.1, this is revised
(and hence can occupy typical to say: ‘Anything below the head
positions associated with adjectival H of the highest projection HP in
expressions – for example, as the a complete clause is impenetrable
complement of is in He is fond of to anything above HP’ (where a
fast food). In a different use of the complete clause is one containing
same word, the head of a movement FORCEP).
chain is the last-created/highest imperative A term employed to
copy in the chain. classify a type of sentence used to
headlinese The register/style of issue an order (e.g. Be quiet! Don’t
language found in newspaper say anything!), and also to classify
headlines. the type of verb form used in an
Head Movement Movement of an item imperative sentence (e.g. be is a
from one head position to another verb in the imperative mood in Be
(e.g. movement of an auxiliary from quiet!). Imperatives can begin with
T to C, or of the verb be from V to please (e.g. Please be quiet!).
T). See §1.4.3. inactive See active.
Highest Copy Condition See active. Inactivity Condition On this condition
identification/identify In a relative on the use of expletives, see
clause like that italicised in ‘I’m expletive.
looking for someone OpREL I can inanimate The term inanimate is
trust’ the null relative operator used to denote (the gender of) an
OpREL can be used in place of the expression which denotes a lifeless
overt relative pronoun whom entity. For example, the relative
because the contents of the null pronoun which could be said to
operator can be identified by its be inanimate in gender, because it
antecedent someone (in the sense requires an inanimate antecedent –
that the features carried by the hence we say something which
operator will match those of its upsets people but not *someone
antecedent). On the identification of which upsets people.
null objects in the recipe register of Indefiniteness Condition On this
English, see Module 7.5. condition on the use of expletives,
illegible See legible. see expletive.
impenetrable/Impenetrability indicative Indicative (auxiliary and
Condition. The Impenetrability main) verb forms are finite forms
Condition is a grammatical which are used (inter alia) in
principle (deriving from Chomsky declarative and interrogative clauses
 Glossary and Abbreviations 441

(i.e. statements and questions). Thus, interface levels Levels at which the
the italicised items are indicative in grammar interfaces (i.e. connects)
mood in the following sentences: with speech and thought systems
‘He is teasing you’, ‘Can he speak which lie outside the domain
French?’, ‘He had been smoking’, of grammar. Phonetic Form is
‘He loves chocolate’, ‘He hated the level at which the grammar
syntax.’ An indicative clause is a interfaces with speech systems, and
clause which contains an indicative Logical Form is the level at which
(auxiliary or non-auxiliary) verb. it interfaces with thought systems.
See mood. Interface conditions are conditions
infinitive/infinitival The infinitive/ which structures must meet in order
infinitival form of a verb is the to be processable by thought or
(uninflected) form which is used speech systems.
(inter alia) when the verb is the intermediate projection See
complement of a modal auxiliary project(ion).
like can, or of the infinitive internal argument See argument.
particle to. Accordingly, the interpretable A feature is
italicised verbs are infinitive/ (semantically) interpretable if
infinitival forms in sentences it contributes to meaning. For
like ‘He can speak French’, and example, in the case of a pronoun
‘He’s trying to learn French.’ An like they, the features [third person,
infinitive/infinitival clause is a plural-number] are interpretable,
clause which contains a verb in but the feature [nominative-case] is
the infinitive form. Hence, the not. See §3.1.3.
bracketed clauses are infinitive Interrogative Generalisation A
clauses in: He is trying [to help generalisation (outlined in §5.2.1)
her], and Why not let [him help specifying that:
her]? (In both examples, help is an
(i) A peripheral interrogative
infinitive verb form, and to when
XP that triggers Auxiliary
used with an infinitive complement
Inversion moves from
is said to be an infinitive particle.)
spec‑FINP (with FIN attracting
INFL A category devised by Chomsky
a finite auxiliary to adjoin to
(1981) whose members include
it) into its criterial position in
finite auxiliaries (which are
spec‑FOCP
INFLected for tense/agreement), and
(ii) The criterial position for a
the INFinitivaL particle to. INFL
peripheral interrogative XP
was abbreviated to I in Chomsky
that does not trigger Auxiliary
(1986b), and replaced by T (= tense
Inversion is spec‑FORCEP
marker) in later work.
in situ A constituent is said to remain Interrogative Inversion A
in situ (i.e. ‘in place’) if it doesn’t phenomenon whereby a focused
undergo movement. interrogative constituent triggers
442 Glossary and Abbreviations

Subject–Auxiliary Inversion, as in any movement operation. Island

‘What else would you like?’ See Constraints are a set of constraints
§5.1.2. devised by Ross (1967, 1986)
Intervention Condition A principle specifying that no constituent
of grammar specifying that ‘Likes inside an island can be extracted
cannot cross likes’ (Abels 2012: out of the island by any movement
247) – that is, no constituent operation that leaves a gap behind.
can move across an intervening For example, coordinate structures
constituent of the same kind. See like William and Harry are islands.
§1.5.1. Hence, in a sentence like They sent
INTP An interrogative projection in William and Harry to Eton, we can
the clause periphery which Shlonsky topicalise the whole coordinate
& Soare (2011) claim allows how structure William and Harry by
come as its specifier. See §5.2.2. moving it to the front of the
intransitive See transitive. overall sentence (as in ‘William and
Inversion A term used to denote a Harry, they sent to Eton’), but we
movement process by which the cannot topicalise Harry alone (as
relative order of two expressions we see from the ungrammaticality
is reversed. It is most frequently of *‘Harry, they sent William and
used in relation to the more specific to Eton’), nor William alone (cf.
operation by which an auxiliary *‘William, they sent and Harry to
verb comes to be positioned in Eton’).
front of its subject, for example, in K/KP Case particle/case phrase. See
questions such as ‘Can you speak case particle.
Swahili?’, where can is positioned in label A notational device used to
front of its subject you. In Module represent linguistic (particularly
5.1, Auxiliary Inversion is said to categorial) properties of
involve a null FIN head which has constituents. For example, if we say
a focused interrogative/negative XP that the word man belongs to the
as its specifier attracting a finite category N of noun, we are using N
auxiliary in T to adjoin to FIN. as a label to indicate the categorial
irrealis An infinitive complement properties of the word man (i.e. to
clause like that italicised in ‘They tell us what grammatical category
would prefer (for) you to abstain’ is man belongs to).
said to denote an irrealis (a Latin Left Edge Ellipsis A type of ellipsis
word meaning ‘unreal’) event in found in abbreviated registers of
the sense that the act of abstention English which is given the following
is a hypothetical event which has characterisation in §7.2.2:
not yet happened and may never
(i) Reduced clauses have the same
happen.
full syntactic structure as their
island A structure out of which
unreduced counterparts
no subpart can be extracted by
 Glossary and Abbreviations 443

(ii) The left edge of a continuous form of some phrase or clause (i.e.
set of one or more projections of linguistic aspects of its meaning).
at the top of a tree in a root The LF component of a grammar
clause can undergo ellipsis in is the (semantic) component which
the phonology converts syntactic structures into
LF-representations. The LF interface
legible A structure is legible at the PF
is the point at which syntactic
interface if it can be assigned an
structures are transferred to the LF
appropriate phonetic representation,
(i.e. semantic) component of the
and legible at the LF interface if
grammar.
it can be assigned an appropriate
license/licensing The licensing
semantic representation (and
conditions for a given type of
conversely, illegible if not). On the
constituent specify the requirements
Legibility Condition, see Feature
that it needs to satisfy in order to
Valuation Conditions.
be used. For example, to say that
lexical The word lexical is used in two
not licenses any in a sentence like
different ways. Since a lexicon is a
He has not made any comment
dictionary (i.e. a list of all the words
is to say that the presence of not
in a language and their idiosyncratic
allows any to be used – that is, any
linguistic properties), the term lexical
can only occur if a licenser like
property means ‘property of some
not is present (cf. *He has made
individual word(s)’. However, the
any comment). On the licensing
word lexical is also used in a second
conditions for the use of null
sense, in which it is contrasted with
objects in the recipe register of
functional (and hence means ‘non-
English, see Module 7.5.
functional’). In this second sense, a
Like-Adjoins-to-Like Constraint A
lexical category is a category whose
constraint which specifies that
members are words with descriptive
likes can only adjoin to likes:
content: hence, categories such as
hence adjunction can only involve
noun, verb, adjective or adverb are
one head adjoining to another, or
lexical categories in this sense. So,
one phrase/maximal projection
for example, the term lexical verb
adjoining to another. See §6.1.2.
means ‘main verb’ (i.e. a non-
linear/linearisation Linear ordering
auxiliary verb like go, find, hate,
is the left-to-right ordering of
want etc.).
words in a structure. On one view,
lexicalise To say that (for example)
linearisation is an operation
a declarative FORCE head is
by which unordered syntactic
lexicalised as that in a complement
structures are assigned a linear
clause means that the FORCE head
ordering by linearisation conditions
is spelled out/pronounced as that.
in the PF component.
LF/Logical Form An LF representation
link A constituent (or position) which
is a representation of the logical
is part of a movement chain.
444 Glossary and Abbreviations

local/locality A number of linguistic to denote pronouns like he/him/

operations (e.g. Agreement, Case- his which refer to male entities. See
marking, Movement etc.) are said to gender.
be local in the sense that they can matrix In a sentence such as ‘I think
only apply when one constituent he lied’, the (italicised) lied clause is
is sufficiently close to another. an embedded/complement clause
For example, Auxiliary Inversion (by virtue of being embedded as
involves adjoining a T auxiliary the complement of the verb think),
to the closest head above it (i.e. and the think clause is the matrix
to C/FIN). Locality requirements clause, in the sense that it is the
are imposed by constraints like clause immediately containing the
the Minimality Condition and the lied clause.
Impenetrability Condition. maximal projection See projection.
locative This is a term which denotes merge An operation by which two
the semantic/thematic function of a constituents are combined together
constituent which denotes place: see to form a larger constituent:
§2.2.2. So, for example, there/where for example, by merging the P/
are locative pronouns in sentences preposition to with the QP/
such as ‘I found it there?’ or ‘Where quantifier phrase several people we
did you find it?’ form the PP/prepositional phrase
locus To say that T is the locus of to several people. See §1.2.1. The
tense is to say that the tense Merge Hypothesis specifies that a
properties of (present or past tense) maximal projection/XP can only
clauses originate on (an auxiliary/ be introduced into a structure
affix in) T. containing it via (first- or second-)
Logical Form See LF/Logical Form. merge with a head: see §6.4.1.
long(-distance) movement Movement Minimality Condition A constraint
across an intervening clause (termed the Relativised Minimality
boundary. See §2.4.2 and §2.4.3. Condition in Rizzi 1990) specifying
M M denotes a modal auxiliary which that a constituent can only enter
originates as the head of an MP/ into a syntactic relation with the
modal projection. See modal. minimal (= closest) constituent
main clause See root clause. of the relevant type above/
main verb A non-auxiliary verb. See below it (where above means
auxiliary. ‘c‑commanding’, and below means
marginal A marginal sentence is ‘c‑commanded by’). See §1.3.2.
one which is of questionable MOD/MODP Constituents in the clause
grammaticality, marked by one or periphery which modify the clause
more preposed question marks (e.g. containing them (e.g. circumstantial
?‘I ought to apologise, shouldn’t I?’). adverbials) are treated as serving
masc(uline) This term is used in as specifiers of a MODP/modifier
discussions of grammatical gender phrase projection with a null MOD/
 Glossary and Abbreviations 445

modifier head under the analysis in Morphological properties are

§4.3.1. properties relating to the form
modal/modality A modal item is of words (e.g. relating to the
one which expresses modality inflections or affixes they carry).
(i.e. notions such as possibility, For example, it is a morphological
futurity or necessity). The set of property of regular count nouns that
modal auxiliaries found in English they have a plural form ending in -s
is usually assumed to include will/ (e.g. cat has the plural cats).
would/can/could/shall/should/may/ morphosyntactic A morphosyntactic
might/must/ought, and need/dare property is a grammatical
when followed by a ‘bare’ (to-less) property, that is, a property
infinitive complement. The set of which affects (or is affected by)
modal adverbs found in English relevant aspects of morphology
includes possibly, perhaps, maybe, and syntax. For instance, case is
probably, conceivably, definitely and a morphosyntactic property in
certainly. that (for example) pronouns have
mood This is a term describing different morphological forms and
inflectional properties of finite occupy different syntactic positions
verbs. Finite (auxiliary and non- according to their case: for example,
auxiliary) verbs in English can be the nominative form of the first
in the indicative mood, subjunctive person plural personal pronoun is
mood, or imperative mood. we and its accusative form is us;
Examples of each type of mood the two occupy different syntactic
are given by the italicised verb positions in that the nominative
forms in the following: ‘He hates form occurs as the subject of a finite
[= indicative] spaghetti’; ‘The court verb (as in ‘We agreed’), whereas
ordered that he be [= subjunctive] the accusative form occurs as the
detained indefinitely’; ‘Please keep complement of a transitive verb or
[= imperative] quiet!’ The mood of preposition (as in ‘Join us’).
the verb determines aspects of the MP A modal projection – that is, a
interpretation of the relevant clause, phrase headed by a modal auxiliary:
so that for example, subjunctive see §6.1.1.
verbs occur in irrealis clauses. multiple agreement Agreement
morpheme The smallest unit of between a probe and more than one
grammatical structure. Thus, a goal. See §3.2.2, and §3.3.1.
plural noun such as cats comprises N Abbreviation for noun.
two morphemes, namely the stem native A native speaker of English
cat and the plural suffix -s. is someone who has acquired and
morphology/ used English as a first language in
morphological Morphology an English-speaking environment
studies how morphemes are from birth (or early childhood), and
combined together to form words. who speaks the language fluently.
446 Glossary and Abbreviations

A person’s native language is a However, the term is sometimes

language which they have acquired extended to mean ‘expression
and used from birth, and which they containing or comprising a noun or
speak fluently. pronoun’.
natural language A language nominative See case.
acquired in a natural setting by non-auxiliary verb A lexical/main
human beings (hence, excluding verb (like want, try, hate, smell, buy
for example, computer languages, etc.) which requires Do-support to
animal communication systems, form questions, negatives and tags.
etc.). nonfinite See finite.
NEG See NEGP. noun/N A category of word (whose
Negation A process or construction in members include items such as boy/
which some proposition is said to friend/thought/sadness/computer)
be false. Negation involves the use which typically denotes an entity
of some negative item such as not, of some kind. In traditional
n’t, nobody, nothing, never, etc. – grammar, a distinction is drawn
though most discussions of negation between common nouns and
in this book are about the not/n’t: proper nouns. Proper nouns are
see NEGP. names of individual people (e.g.
Negative Inversion A phenomenon Chomsky), places (e.g. Colchester,
whereby a focused negative Essex, England), dates (e.g. Tuesday,
constituent triggers Subject– February, Easter), magazines (e.g.
Auxiliary Inversion, as in ‘Not a Cosmopolitan) etc. whereas common
single word did he utter’. See §5.1.1. nouns (e.g. boy, table, syntax etc.)
NEGP The abbreviation NEGP denotes are nouns denoting general (non-
a negative projection (i.e. a phrase individual) entities. Proper nouns
whose edge contains n’t/not). In have the semantic property of
this book, n’t is treated as a clitic having unique reference, and the
which originates as the head NEG syntactic property that (unless
constituent of NEGP and encliticises themselves modified) they generally
(i.e. attaches to the end of) a finite can’t be modified by a determiner in
auxiliary in T; and not is treated as English (cf. *the London). On counts
an ADVP occupying the specifier nouns, see count.
position in a NEGP with a null NEG Noun Phrase/NP A phrase whose
head. head is a noun. In work prior to
neuter See gender. the mid-1980s, a structure such as
Nom An abbreviation for nominative. the king of Utopia was taken to be
See case. a noun phrase/NP comprising the
nominal This is the adjective head noun king, its complement of
associated with the word noun, so Utopia, and its specifier the. In more
that a nominal (constituent) is one recent work, such expressions are
containing or comprising a noun. taken to be Determiner Phrases/DPs
 Glossary and Abbreviations 447

comprising the head determiner the ‘this book’, ‘these books’), in some
and a noun phrase/NP complement pronouns (cf. it/they), and in finite
king of Utopia, with the NP in turn (auxiliary or main) verbs (cf. ‘It
comprising the head noun king and smells’, ‘They smell’).
its complement of Utopia. NUMP See NUM.
NP See Noun Phrase. OBJ Abbreviation for object or object
null A null constituent is one which agreement affix.
is ‘silent’ or ‘unpronounced’ and so object The complement of a transitive
has no overt phonetic form. Such item (e.g. in Help me! the pronoun
a constituent is said to receive a me is the object of the transitive
null spellout in the phonological verb help; and in for me, me is the
component. object of the transitive preposition
null case The case carried by PRO. See for). The term object is generally
case. restricted to complements which
null object See Object Drop. carry accusative case – that
null subject A subject pronoun which is, to nominal or pronominal
has grammatical and semantic complements: hence, nothing would
properties but no overt phonetic be the object (and complement) of
form. There are a variety of different said in He said nothing, but the
types of null subject, including the clause that he was tired would
null pro subject which can be used be the complement (but not the
in any finite clause in a language object) of said in He said that he
like Italian, the null counterpart of was tired – though some traditional
you found in English imperative grammars extend the term object
clauses like Look after yourself!, the to cover clausal complements as
null PRO subject found in nonfinite well as (pro)nominal complements.
control clauses like that bracketed In sentences such as She gave him
in The prisoners tried [PRO to them, the verb give is traditionally
escape], and the null subject found said to have two objects, namely
in abbreviated sentences like Can’t him and them: the first object
find my pen. Must be on my desk at (representing the recipient) is termed
home. the indirect object, and the second
NUM An abbreviation for numeral. object (representing the gift) is
Hence a phrase like ten people is a termed the direct object; the relevant
NUMP (numeral projection) headed construction is known as the double
by the NUM/numeral constituent object construction. Where a verb
ten. has a single object (e.g. nothing in
number A term used to denote the He said nothing), this is the direct
contrast between singular and plural object of the relevant verb.
forms. In English, we find number Object Drop A phenomenon found
contrasts in nouns (cf. ‘one dog’, in certain abbreviated registers
‘two dogs’), in some determiners (cf. of English (e.g. on product labels,
448 Glossary and Abbreviations

in recipe books, in instruction if it has no phonetic content (and so

manuals etc.) whereby a third is unpronounced/silent). Thus, in a
person accusative pronoun such as structure such as:
it/them can be dropped/omitted –
(i) He wants [PRO to leave]
as in the following instructions
on a nasal spray Shake <it> he is an overt pronoun, but PRO is
gently before use (where the third a null pronoun.
person pronoun it is omitted). See P Abbreviation of preposition.
Module 7.5. participle A nonfinite verb form
One Feature One Head Principle A which encodes aspect or voice.
principle which specifies that English has three types of
each functional head carries only participle: progressive participles
one interpretable feature. See (ending in -ing) used in conjunction
Module 4.1. with the progressive aspect
one-place predicate A predicate auxiliary b e in sentences like ‘It
which has only one argument. See is raining’; perfect participles
argument. (generally ending in -d or -n) used
Op/operator This term is used in in conjunction with the perfect
syntax to denote a constituent (e.g. aspect auxiliary h av e in sentences
negative, interrogative, relative, like ‘The referee has shown/showed
conditional, or imperative) whose him a red card’; and passive
semantic function is to convert participles (also generally ending
a proposition into, for example, in -d or -n) used in conjunction
a negative or interrogative with the passive voice auxiliary b e
or relative or conditional or in sentences like ‘He was shown/
imperative clause. So, for example, showed a red card by the referee.’
a yes–no question like Have you particle/PRT This is an informal
eaten anything? can be analysed term used to describe a range of
as containing a null/silent yes–no (typically monosyllabic) items which
question operator (OpYNQ) which are invariable in form, and which
types the clause as a yes–no don’t fit easily into traditional
question, which triggers Auxiliary systems of grammatical categories.
Inversion, and which licenses the For example, infinitival to (e.g.
polarity item anything: see Module in Try to be nice!) is said to be an
5.3. Likewise, an interrogative infinitive particle; of as used in
word such as what in a sentence expressions like loss of face is said
like What has anyone ever done for to be a case particle; and not and
me? is a wh‑question operator, n’t are said to be negative particles.
since it types the relevant clause as The term is sometimes extended to
a wh‑question. include prepositions used without
overt An expression is overt if it has a complement (e.g. down in He fell
an audible phonetic form, but null down).
 Glossary and Abbreviations 449

passive/Passivisation A contrast ‘Syntax, why do students struggle to

is traditionally drawn between understand it?’, all three italicised
sentence pairs such as (i) and (ii) items that precede the underlined
below: subject are positioned in the clause
periphery. See Chapters 4 and 5.
(i) The thieves stole the jewels
Pers An abbreviation of person.
(ii) The jewels were stolen by the
person In traditional grammar, English
thieves
is said to have three grammatical
(i) is said to be an active clause (or persons. A first person expression
sentence), and (ii) to be its passive (e.g. I/we) is one whose reference
counterpart; similarly, the verb includes the speaker(s); a second
stole is said to be an active verb person expression (e.g. you) is one
(or a verb in the active voice) in (i), which excludes the speaker(s) but
whereas the verb stolen is said to includes the addressee(s) (i.e. the
be a passive verb (or a verb in the person or people being spoken to);
passive voice – more specifically, a a third person expression (e.g. he/
passive participle) in (ii); likewise, she/it/they) is one whose reference
the auxiliary were in (ii) is said to excludes both the speaker(s) and the
be a passive auxiliary. Passivisation addressee(s) – that is, an expression
is a movement operation whereby which refers to someone or
(typically) an expression which is something other than the speaker(s)
the complement of a verb in the or addressee(s).
passive participle form becomes the personal pronouns These are pronouns
subject of the same clause (as in which carry inherent person
‘The jewels were stolen — ’). In cases properties – for example, first person
of long Passivisation, the subject pronouns such as I/we, second
of a complement clause becomes person pronouns such as you, and
the subject of a higher clause third person pronouns such as he/
containing a passive participle (as she/it/they. In the framework used
in ‘The minister was said — to have here, personal pronouns are treated
lied to Parliament’). See Module 2.4. as pronominal DPs. See person.
past tense See tense. PF/Phonetic Form A PF
PERF(ect)/PERFP The abbreviation representation is a representation of
PERF denotes an auxiliary marking the phonetic form of an expression.
perfect aspect, and PERFP denotes The PF component of a grammar is
a phrase headed by a perfect aspect the component which converts the
auxiliary (e.g. have lied to her in ‘He structures generated by the syntactic
may have lied to her’). See aspect. component of the grammar into
periphery The periphery of a clause PF representations, via a series of
is that part of the clause structure morphological and phonological
which is positioned above/to the left operations. The PF interface is the
of the subject. So, in a sentence like point at which syntactic structures
450 Glossary and Abbreviations

are transferred to the PF component pl/plural A plural expression is one

of the grammar: the phrase ‘at PF’ which denotes more than one
means ‘at the PF interface’. entity (e.g. these cars is a plural
phi-features/φ-features Person and expression, whereas this car is a
number features. A phi-complete singular expression).
(or φ-complete) constituent is one polarity A polarity item/expression
which carries both person and is a word/phrase (e.g. a word
number. See §3.1.1. like ever or a phrase like at all or
phonetic form See PF and interface care a damn) which is restricted
levels. to occurring within the scope
phonetic representation See of a negative, interrogative or
representation. conditional constituent.
phrasal adjunct See adjunct. possessive/possessor/possessum A
phrase The term phrase is used possessive structure is one which
to denote an expression larger indicates possession, like John’s
than a word which is a maximal book. In such structures, the
projection: see projection. In person possessing the item in
traditional grammar, the term question (here, John) is termed
refers strictly to non-clausal the possessor, and the possessed
expressions (Hence, reading a book object (here, book) is termed the
is a phrase, but He is reading a possessum. On the structure of
book is a clause, not a phrase). phrases like John’s book/your book
However, in the framework used see §7.3.1.
here, clauses are analysed as PP See prepositional phrase.
types of phrases: for example, pragmatics The study of how non-
in the classic analysis of clause linguistic knowledge is integrated
structure outlined in Chapter 1, with linguistic knowledge in our use
he will resign is a tense phrase of language.
(TP), and that he will resign is a precede(nce) To say that one
complementiser phrase (CP). constituent precedes another is
pied-piping A process by which a to say that it is positioned to its
moved constituent drags one or left (on the printed page) and that
more other constituents along with neither constituent contains the
it when it moves. For example, if other. Precedence is left-to-right
we compare a sentence like Who linear ordering.
were you talking to? with To whom predicate See argument.
were you talking? we can say that premodifier A premodifier is a
in both cases the pronoun who(m) is constituent which precedes and
moved to the front of the sentence, modifies some other constituent
but that in the second sentence the (e.g. fervently is a premodifier of the
preposition to is pied-piped along VP believe in equality in ‘We all do
with the pronoun whom. fervently believe in equality’).
 Glossary and Abbreviations 451

prepose/preposing Preposing is minimally semantically contentful

an informal term to denote a possibilities compatible with the
movement operation by which a morphosyntactic environment’: see
given expression is fronted – that is, §7.2.1.
moved to the front of some phrase, PRO A null case pronoun (known
clause or sentence. informally as ‘big PRO’, because it
preposition/P A preposition is a word is written in capital letters) which
used to indicate notions such as represents the understood subject
direction (as in ‘He went to Paris’), of an infinitive complement of a
time (as in ‘I met her on Friday’), control predicate, for example, in a
place (as in ‘She lives in a big structure such as John decided PRO
house’), or means (as in ‘He was to leave, where PRO refers back to
killed with an axe’). In English, John.
it is a characteristic property of pro A null pronoun (known informally
prepositions that they are invariable, as ‘little pro’ because it is written
and that they can be modified in lower case letters). One type of
by straight/right if they have an pro is the understood null subject
appropriate meaning (as in ‘She of a finite clause in an Italian
went straight to the police’). Where sentence like ‘pro pioverà domani’
a preposition has a nominal or (= ‘It [Link] tomorrow’). Another
pronominal complement, it is said type of pro is the null third person
to be transitive; in other uses, it is object found in recipe registers, for
said to be intransitive. Hence down example, ‘Divide pro into four’: see
is a transitive preposition in He fell §7.5.3.
down the stairs, but an intransitive probe When a head is active, it serves
preposition in He fell down. as a probe which searches for a
prepositional phrase/PP A phrase suitable goal within its complement
whose head is a preposition – for (where the goal is a constituent
example, in town, on Sunday, to the which it can agree with, or case-
market, for someone else, etc. mark, or attract to move to the edge
pres Abbreviation for present tense: of its projection): see §3.1.1. On the
See tense. Probe Condition, see active.
principle See constraint. PROG(ressive)/PROGP PROG
Principle of Unambitious Reverse denotes an auxiliary marking
Engineering/PURE A principle progressive aspect: PROGP denotes
devised by Pesetsky (2020: 5) a progressive aspect phrase (like
specifying that ‘When determining be lying in ‘He may be lying’). See
the identity of unpronounced aspect; see also §6.1.1.
material in the course of reverse- project(ion) A projection is a
engineering a speaker’s syntactic constituent containing or
derivation, the language system comprising a head word. For
of the hearer considers only the example, a noun phrase/NP such
452 Glossary and Abbreviations

as students of Linguistics is a is traditionally said to be a word

projection of its head noun students used in place of a noun expression.
(equivalently, we can say that the Pronouns differ from nouns in that
noun students here projects into they have no intrinsic descriptive
the NP students of linguistics). A content, and so are function words.
minimal projection is a constituent There are a range of different
which is not a projection of some types of pronoun found in English,
other constituent: hence, heads (e.g. including the pronominal NP one(s)
words) are minimal projections. used in sentences like I’ll take the
An intermediate projection is a red one(s), pronominal QPs like
constituent which is larger than a any in I couldn’t find any, and
word, but smaller than a full phrase pronominal DPs like this in This
(e.g. is working in He is working). A is hard. The term pronoun is most
maximal projection is a constituent frequently used to indicate a class of
which is not contained within any items (like he/him/his) traditionally
larger constituent with the same referred to as personal pronouns
head. So, for example, in a sentence (though they are analysed in this
like ‘I’ve heard several accounts of book as pronominal DPs).
what happened’, the italicised noun proposition This is a term used to
phrase accounts of what happened describe the semantic content
is a maximal projection (an NP), (i.e. meaning) of a sentence.
since it is a projection of the noun For example, we might say
accounts but is not contained within that the sentence Does John
any larger projection of the noun smoke? questions the truth of the
accounts (if we assume that several proposition that ‘John smokes’. See
accounts of what happened is a argument.
quantifier phrase/QP headed by prosodic account This is an account
the quantifier several). By contrast, of abbreviated registers of English
in a sentence such as ‘I’ve heard which (as outlined in Module 7.2)
several accounts’, the italicised makes the following assumptions
noun accounts is both a minimal about reduced/abbreviated clauses:
projection (by virtue of not being
(i) Reduced clauses have the same
a projection of some other head)
full syntactic structure as their
and a maximal projection (by virtue
unreduced counterparts
of not being contained within any
(ii) A continuous sequence of one
larger structure which has the same
or more weak syllables at the
head noun): hence accounts here
beginning of a prosodic phrase
can be analysed as an NP.
(e.g. a sentence) can undergo
pronoun The word pronoun is
ellipsis at PF
composed of two morphemes –
namely pro (meaning ‘on behalf pseudo-cleft sentence A sentence such
of’) and noun: hence, a pronoun as ‘What he hated most was syntax’,
 Glossary and Abbreviations 453

where syntax is said to occupy focus A‑Movement operation (sometimes

position within the overall sentence. called Subject Raising) by which an
PURE An abbreviation of the expression is moved from being the
Principle of Unambitious Reverse subject of one clause to becoming
Engineering. the subject of another. The term
Q/QP See quantifier. raising predicate denotes a word
quantifier A quantifier/Q is a type like seem whose subject is raised out
of word used to denote quantity. of subject position in an infinitival
Typical quantifiers include the complement clause to become
universal quantifiers all/both, subject of the (TP constituent in the)
the free choice quantifiers any/ seem clause. See Module 2.5.
whatever (as in You can choose realis A realis clause is one used to
any/whatever book you like), the describe a real state of affairs, and
distributive quantifiers each/every, typically contains an (auxiliary
the partitive quantifiers some/any, or non-auxiliary) verb which
etc. A quantifier phrase/QP is a is indicative in mood, like the
phrase whose head is a quantifier – italicised verbs in ‘He cheats/cheated
for example, an expression such as at cards.’
many people, or few of the students. Recoverability Condition This is a
The QP Hypothesis of §1.3.1 condition proposed in Chomsky
posits that all indefinite nominal (1964b: 41) and Katz & Postal
arguments are QPs, headed by an (1964: 79) to the effect that a
overt or null quantifier. In this constituent which is present in the
book, quantificational pronouns syntax can only be deleted/ellipsed/
like someone/anything/everyone are unpronounced/null if its semantic
taken to be pronominal QPs. content is recoverable.
quantifier floating/stranding See recursion A structure involves
floating quantifier. recursion if it contains more
question This refers to a type of than one instance of a particular
sentence which is used to ask category. For example, CP recursion
whether something is true, or to structures are clauses which contain
ask about the identity of some more than one CP projection in
entity. See yes–no question and their periphery, like the clause
wh‑question. italicised in ‘He said [that [unless
raising The term raising is used in she does as he says [that he will tell
two senses. In its most general her father]]]’, which contains three
sense, it denotes any movement bracketed CPs. See §4.1.1.
operation which involves moving reduced See abbreviated.
some constituent from a lower to reflexive See anaphor.
a higher position in a structure. REL/relative/RELP In a sentence
However, it also has a more specific such as ‘He’s someone [who you
sense, indicating a particular kind of can trust]’, the bracketed clause is
454 Glossary and Abbreviations

said to be a relative clause because restrictive relative clause See relative

it ‘relates to’ (i.e. modifies) the clause.
pronoun someone. The pronoun resumptive A sentence such as:
who which introduces the clause
(i) Behaviour like that, we simply
is said to be a relative pronoun,
cannot tolerate it in this school
since it ‘relates to’ the expression
someone (in the sense that who involves an italicised topic followed
refers back to someone). Restrictive by a comment. The relevance of the
relative clauses serve the function comment to the topic is ensured by
of restricting the class of entities using the resumptive pronoun it to
referred to by the antecedent to ‘resume’ or ‘reprise’ (i.e. refer back
those which have the property to) the topic this kind of behaviour,
described in the relative clause: so ensuring that the comment is
for example, in a sentence like ‘I relevant to the topic.
need to work with people who I root The root of a tree is the topmost
can trust’, the italicised restrictive node in the tree. A root clause
relative clause restricts the class of is a free-standing clause, that is,
people referred to in the sentence a clause which is not contained
to those trusted by me. Under the within another clause (or phrase).
analysis in §5.4.2, a relative clause In traditional grammar, a root
like that italicised in ‘the film which clause is termed a principal clause,
I watched’ is treated as a RELP/ independent clause, or main clause.
relative projection, with a REL head By contrast, an embedded clause
which has a relative operator (here, is a clause which is contained
the relative pronoun which) as its within some larger expression;
specifier. and a complement clause is an
Relativised Minimality Condition See embedded clause which is used as
Minimality Condition. the complement of some item. So,
representation A syntactic in a sentence such as I think he
representation is a notation/device loves you, the think clause (i.e. the
(typically, a tree diagram or labelled expression I think he loves you) is a
bracketing) used to represent the root clause, whereas the loves clause
syntactic structure of an expression (i.e. the expression he loves you) is
(i.e. the way in which it is structured an embedded clause. Moreover, the
out of words and phrases): a loves clause is also a complement
semantic representation is a clause since it serves as the
representation of linguistic aspects complement of the verb think.
of the meaning of an expression; scope The scope of an expression
a PF representation or phonetic is the set of constituents which
representation is a representation of it modifies or which fall within
the phonetic form of an expression, (what we might informally call) its
specifying how it is pronounced. ‘sphere of influence’. For example,
 Glossary and Abbreviations 455

a sentence like He cannot be semantics/semantic component

telling the truth has a meaning Semantics is the study of linguistic
paraphrasable as ‘It is not possible aspects of meaning. The semantic
that he is telling the truth’, and in component of a grammar is the
such a sentence the negative not is component which maps/converts
said to have scope over the modal syntactic structures into semantic
auxiliary can (or to have wide scope representations. See representation.
with respect to can), and conversely sentence This term is usually used
can is said to fall within the scope to denote a root clause – that is,
of not, or to have narrow scope a free-standing clause which is
with respect to not. By contrast, a not contained within some larger
sentence such as You mustn’t tell expression. See root.
lies has a meaning paraphraseable sg See singular.
as ‘It is necessary that you not tell silent See null.
lies’, and in such a sentence, the singular A singular expression is one
auxiliary must is said to have scope which denotes a single entity (e.g.
over the negative particle n’t (or this car is a singular/sg expression,
to have wide scope with respect whereas these cars is a plural/pl
to n’t). expression).
second person See person. Sluicing A form of ellipsis in
select(ion)/selectional/selector When which the material following an
a word has a particular type of interrogative wh‑constituent is
complement, it is said to select given a silent pronunciation – as
(i.e. ‘take’ or ‘allow’) the relevant with the material following when
type of complement, and the (marked by strikethrough) in ‘I
relevant phenomenon is referred remembered that he was leaving,
to as (complement-) selection. but forgot when he was leaving.’
For example, we can say that the spec See specifier. Terms like spec-
word expect has the selectional CP/spec-TP (etc.) denote the
property that it can select an specifier of CP/TP.
infinitive complement (e.g. in specifier/spec The grammatical
structures like ‘They expect to win’). function fulfilled by certain types
In this structure, the verb expect of constituent which precede the
is the selector for the infinitive head of their containing phrase.
complement to win, in the sense For example, in a preposition
that expect is the item which selects phrase/PP such as straight through
to win as its complement. Selection the window, the preposition
restrictions are restrictions which through is the head of the PP,
predicates impose on their choice the window is the complement
of arguments: for example, the verb of through, and straight is the
think requires an animate/rational specifier of the (PP headed by the)
subject. preposition through.
456 Glossary and Abbreviations

spellout The pronunciation of an comprising a number of separate

expression: for example, to say phrases. For example, the traditional
that an item has a null spellout CP constituent is split up into (inter
is to say that it is ‘silent’ and so alia) FORCEP, TOPP, FOCP, MODP
is unpronounced. Discontinuous and FINP constituents in more
spellout (also termed split spellout) recent Cartographic work. See
is a phenomenon which arises when Chapters 4–7.
part of a moved phrase is spelled split spellout See spellout.
out on a lower copy of a moved stack/stacking Stacking is a
constituent, and the remainder on process by which more than one
the highest copy – as in ‘How much constituent of the same type can
do you believe of what he tells be used to modify some other
you?’, where the wh‑phrase how constituent. For example, an NP
much of what he tells you moves like head of state can be modified
to the front of the sentence, with by numerous different APs/
how much being spelled out in the adjectival phrases in an expression
position it moves to, and of what like ‘a very sensitive, rather tall,
he tells you being spelled out in the really dark, incredibly handsome
position in which it originates. The head of state’. In this particular
Default Spellout Rule of §1.4.1 example, four different (italicised)
specifies that where the spellout of APs are said to have been stacked
a constituent is not determined in on top of/to the left of the NP/noun
some other way, the highest copy phrase head of state.
of the constituent is pronounced at stem The stem form of a word is the
PF, and any lower copies are silent. base form to which inflectional
The Low Spellout Rule of §2.3.3 affixes are added. So, a verb form
specifies that a KP/PP/CP which is like going comprises the stem go and
the lowest/rightmost constituent the inflectional suffix -ing.
of a larger phrase that undergoes stranded quantifier See floating
movement can be spelled out on a quantifier.
lower copy of the moved phrase. string A continuous sequence of
The Floating Quantifier Spellout words contained within the same
Rule of §6.4.2 specifies that phrase or sentence. For example,
floatable quantifiers can be spelled in the sentence ‘They hate syntax’,
out on any non-initial link of an the sequences They hate, hate
A‑Movement chain in standard syntax and They hate syntax are all
varieties of English. strings – but They syntax is not.
split projection A split projection Structure Dependence Principle A
analysis is one in which a structure traditional grammatical principle
that was traditionally analysed as (mentioned in §1.3.2) requiring that
comprising a single phrase becomes syntactic operations be sensitive
reanalysed in subsequent work as to hierarchical relations between
 Glossary and Abbreviations 457

constituents – and to the relation nominative subject he, and hence is

c‑command in particular. a finite verb form.
SUBJ/SUBJP In work discussed in subperiphery/subperipheral The
§6.1.2, the criterial position for a subperiphery of a clause comprises
subject is taken to be the specifier that part of the clause structure
position in a SUBJP/subject positioned above the verb phrase
projection with an abstract SUBJ but below the periphery: in the
head. traditional CP+TP+VP model of
subject The subject of a clause is a clause structure, it corresponds
nominal or pronominal expression to the TP layer. A subperipheral
which is normally positioned constituent is one positioned in the
between a complementiser and an subperiphery of the clause (i.e. on
(auxiliary or non-auxiliary) verb. the edge of TP in the traditional
Syntactic characteristics of subjects CP+TP+VP model): see Chapter 6.
include the fact that they can successive-cyclic movement
trigger agreement with auxiliaries Movement in a succession of short
(as in ‘The president is lying’, where steps (e.g. one clause at a time). See
the auxiliary is agrees with the §1.5.2.
subject the president), and they suffix See affix.
can be inverted with auxiliaries T This denotes a tense-marking
in main clause questions (as in ‘Is constituent containing either a
the president lying?’, where the tensed auxiliary, or an abstract
auxiliary is has been inverted with tense affix, or a nonfinite tense
the subject the president). particle like infinitival to. A phrase
Subject Drop A phenomenon whereby headed by a T constituent is labelled
the subject of a finite clause can TP. T‑to‑C Movement is movement
be dropped/omitted in abbreviated of an auxiliary verb from the head
registers of English. For example, in T position of TP into the head
spoken English, we frequently find C position of CP – as with the
sentences like can’t find my phone, italicised inverted auxiliary in ‘Is it
where the bracketed subject I is raining?’ See §1.4.3.
omitted. See Module 7.1. tag A string usually consisting of an
subjunctive In a (formal style) auxiliary and a subject pronoun
sentence such as ‘The judge ordered which is ‘tagged’ (i.e. added) onto
that he be detained indefinitely’, the end of a sentence. Thus, the
the passive auxiliary verb be is italicised string is the tag in the
traditionally said to be in the following: ‘The president is working,
subjunctive mood, since although is(n’t) he?’, and the overall sentence
it has exactly the same form as the is known as a tag question/tag
infinitive form be (e.g. in infinitive sentence. The tag usually contains
structures such as ‘To be or not to a copy of the auxiliary in the main
be – that is the question’), it has a clause (which may or may not
458 Glossary and Abbreviations

carry the negative suffix n’t), and a subject receives default case. On
pronominal copy of the subject. one view, e Drop clauses contain a
tense/tensed (Auxiliary and TP whose head T contains an affix
main) verbs which are in the with an unvalued tense feature; on
indicative mood in English show another view, they lack both TP
a binary (two-way) tense contrast, and DP projections, and have NP
traditionally said to be between subjects.
present tense forms and past tense
theme The name of a specific theta
forms. Thus, in ‘John hates syntax’,
role (sometimes also termed pat i e nt )
hates is a present tense verb form,
representing the entity undergoing
whereas in ‘John hated syntax’,
the effect of some action (e.g. Harry
hated is a past tense verb form.
in ‘William teased Harry’). See
This present/past tense distinction
§2.2.2.
correlates with time reference, so
theta/θ A theta role/θ‑role/thematic
that (for example) past tense verbs
role is the semantic role played
typically describe an event taking
by an argument in relation to its
place in the past, whereas present
predicate (e.g. agent , the e , goa ,
tense verbs typically describe an
etc.). For example, in a sentence
event taking place in the present (or
like William teased Harry, the verb
future). However, the correlation is
tease assigns the θ‑role agent to its
an imperfect one, since for example,
subject William and the theta role
in a sentence such as ‘If you went
the e to its complement Harry.
to Paris, you’d have a great time’,
A theta‑position/θ‑position is a
the verb went carries the past tense
position to which a theta role is
inflection -t (found on past tense
assigned; a theta‑bar/θ‑bar position
verbs like left) but does not denote
is a position to which no theta role
past time (but instead denotes
is assigned. To say that a predicate
a hypothetical event). A tensed
theta-marks its arguments is to
(auxiliary or non-auxiliary) verb
say that it determines the theta role
form is one which carries (present/
played by each of its arguments.
past) tense – for example, is, could,
The Theta Criterion/θ‑Criterion is
hates, went, etc. By extension, a
a principle of Universal Grammar
tensed clause is one containing a
which specifies that each argument
tensed auxiliary or main verb.
should bear one and only one
Tense Deficit Hypothesis An account
theta role, and that each theta role
of Be Drop which (as outlined in
associated with a given predicate
§7.4.3) posits that:
should be assigned to one and only
e Drop clauses are truncated one argument. See Module 2.2.
SUBJP constituents with a tense third person See person.
deficit, which means that be receives TOP/TOPP TOP denotes a topic
a null spellout by default, and its particle; TOPP denotes a topic
 Glossary and Abbreviations 459

projection in the periphery of a transitive A word is traditionally

clause which contains a topic as its said to be transitive (in a given
specifier. See topic. use) if it assigns accusative case
topic/Topicalisation/Topic Drop In a to a noun or pronoun expression
dialogue such as the following: which it immediately precedes and
c‑commands: See case. So, likes in
s p ea k e r a : I’ve been having
‘John likes him’ is a transitive verb,
problems with the Fantasy
since it assigns accusative case
Syntax seminar
to its complement him. Likewise,
s p ea k e r b : That kind of course, very
in is a transitive preposition in
few students seem to get their
‘people have little faith in him’.
heads round
And similarly, infinitival for is a
the italicised expression that kind transitive complementiser, since
of course serves as the topic of the it assigns accusative case to the
sentence produced by speaker b , in subject of its infinitive complement
the sense that it refers back to the (as in ‘I’m keen [for him to
Fantasy Syntax seminar mentioned by participate more actively]’). A verb
the previous speaker: more generally, or complementiser is intransitive
an expression which represents ‘old’ (in a particular structure) if it
or ‘familiar’ information is said to be does not assign accusative case
a topic. Topicalisation is a movement to any constituent in the relevant
operation by which a constituent is structure.
moved into the periphery of a clause truncated/Truncation A truncated
to mark it as a topic: see Module clause is one which does not project
4.1 and Module 4.2. Topic Drop is all the way up to FORCEP, but
an account of how a third person rather only as far as (for example)
pronoun can undergo Topicalisation TOPP, or FINP, or SUBJP, and
in the syntax and move into the consequently lacks one or more of
specifier position in a TOPP/topic the peripheral projections found in
projection, and thereafter be given a complete clauses (i.e. clauses that
silent spellout in the phonology: see contain FORCEP). In Module 7.1,
§7.5.1. Truncation is given the following
TP Tense projection/tense phrase – characteristion (based on work by
that is, a phrase headed by a tense- Rizzi (1994, 1998, 2000a):
marked auxiliary, or by an abstract
(i) A sentence can have a truncated
tense affix (Af), or by an infinitival
syntactic structure whose root
tense particle.
is not FORCEP, but some lower
trace A trace of a moved constituent
projection such as SUBJP
is a (usually null) copy left behind
(ii) The edge of the highest
(as a result of movement) in each
projection in a sentence
position out of which a constituent
structure can be unpronounced
moves. See also COMP‑Trace Filter.
460 Glossary and Abbreviations

in the PF component, UTAH See Uniform Theta Assignment

provided that it contains only Hypothesis.
phonetically weak material with V/VP See verb/V and verb phrase/VP.
recoverable semantic content value In relation to a feature such as
[Singular-Number], number is said
T‑to‑C movement See T. to be an attribute (and represents
two-place predicate A predicate which the property being described) and
has two arguments – for example, singular its value.
tease in William teased Harry where variety A particular (e.g. geographical
the two arguments of the predicate or social) form of a language.
tease are William and Harry. See verb/V A category of word which
argument. has the morphological property
UG See Universal Grammar. that it can carry a specific range of
unaccusative An unaccusative inflections (e.g. the verb show can
predicate is a word like come whose carry past tense -d, third person
superficial subject originates as its singular present tense -s, perfect -n
complement. See Module 2.3. and progressive -ing, giving rise to
unergative An unergative predicate is shows/showed/shown/showing), and
a verb like groan in a sentence such the syntactic property that it can
as ‘He was groaning’ which has an head the complement of infinitival
ag e nt subject but no overt object: to (as in ‘Do you want to show me?’)
see §2.3.2. verb phrase/VP A phrase which is
Uniform Theta Assignment headed by a verb – for example, the
Hypothesis/UTAH A hypothesis italicised phrase in ‘They will help
(developed by Baker 1988) which you’.
maintains that each theta role voice See passive. A VOICEP
assigned by a particular kind of constituent is a voice phrase headed
predicate is canonically associated by a VOICE constituent like the
with a specific syntactic position: passive auxiliary b e : see §6.1.1.
for example, the complement VP See verb phrase. A VP-adverb is
position is the canonical position an adverb which adjoins to a VP to
for a the e argument. See §2.3.2. form an even larger VP (e.g. in He
uninterpretable Features are may have completely forgotten what
uninterpretable if they have no you asked, the adverb completely
semantic content and hence play no is adjoined to the VP forgotten
role in determining meaning. See what you asked). On VP Ellipsis,
§3.1.3. see ellipsis. On VP adjunct, see
Universal Grammar/UG This term adjunct.
denotes (a theoretical model of, VP-Internal Subject Hypothesis/
or the study of) those aspects of VPISH The hypothesis that subjects
grammar which are hypothesised to originate internally within the verb
be universal. phrase, and from there move to the
 Glossary and Abbreviations 461

criterial subject position (in spec‑TP/ wh‑question A question which

spec‑SUBJP): see Module 2.1. contains an interrogative wh‑word,
wh This is widely used as a feature for example, What are you doing?
carried by constituents which On wh‑question operator, see
undergo Wh‑Movement – for operator.
example, the relative pronoun who wh‑word A word which begins with
in someone who I think is lying can wh (e.g. who/what/which/where/
be described as a wh‑pronoun, as when/why), or which has a similar
can the interrogative pronoun who syntax to wh‑words (e.g. how).
in Who are you waiting for? and X A symbol used to denote any type
the exclamative quantifier what of constituent which is a head (e.g.
in What a great time we had! See P, A, ADV, D, Q, T, C, etc.)
Chapter 5. X‑bar Syntax A theory of syntactic
wh‑constituent A constituent structure which posits that there are
containing or comprising a X‑bar constituents larger than words
wh‑word. but smaller than phrases: see bar.
Wh‑Movement A type of XP A symbol used to denote any type
movement operation whereby a of constituent which is a phrase/
wh‑constituent is moved into the maximal projection (e.g. PP, ADVP,
periphery of some clause (e.g. to AP, DP, QP, TP, CP, etc.).
the front of the overall sentence in yes–no question operator See
‘Where has he gone?’). operator.
wh‑operator See operator. YNQ An abbreviation of yes–no
wh‑phrase A phrase containing a question. For example OpYNQ
wh‑word (e.g. what kind of person denotes an abstract yes–no question
in What kind of person is she?). operator: see operator and §1.4.3.
 References

The references below include all items referred to in the book and the accompany-
ing Answerbooks.

Abbott, B. 1993. A pragmatic account L. Brugé, G. Giusti, N. Munaro,

of the definiteness effect in W. Schweikert & G. Turano (eds.)
existential sentences. Journal of 265–85. Venice: Cafoscarina.
Pragmatics 19: 39–55. ———2006. Complementation in
Abeillé, A. & Godard, D. 1999. A Saramaccan and Gungbe: The case
lexical approach to quantifier of C‑type modal particles. Natural
floating in French. In Lexical and Language and Linguistic Theory 24:
Constructional Aspects of Linguistic 1–55.
Explanation, G. Webelhuth, J.-P. ———2007. Focused versus non-focused
Koenig & A. Kathol (eds.) 81–96. wh‑phrases. In Focus Strategies
Stanford, CA: CSLI Publications. in African Languages, E. O. Aboh,
Abels, K. 2007. Deriving selectional K. Hartmann & M. Zimmermann
properties of ‘exclamative’ (eds.) 287–314. Berlin: Mouton de
predicates. In Interfaces and Gruyter.
Interface Conditions, A. Späth (ed.) ———2010. Information structuring
115–40. Berlin: De Gruyter. begins with numeration. Iberia 2:
———2010. Factivity in exclamatives is a 12–42.
presupposition. Studia Linguistica ———2016. Information structure: A
64: 141–57. Cartographic perspective. In The
———2012. The Italian left periphery: Oxford Handbook of Information
A view from locality. Linguistic Structure, C. Féry & S. Ishihara
Inquiry 43: 229–54. (eds.) 147–64. Oxford: Oxford
Abney, S. P. 1987. The English noun University Press.
phrase in its sentential aspect. PhD Aboh, E. & Pfau, R. 2010. What’s a
diss. MIT. wh‑word got to do with it? In
Aboh, E. O. 2004 The Morphosyntax The Cartography of Syntactic
of Complement-Head Sequences. Structures, Vol. 5: Mapping the left
Oxford: Oxford University Press. periphery, P. Benincà & N. Munaro
———2005. Deriving relative and (eds.) 91–124. Oxford: Oxford
factive constructions in Kwa. In University Press.
Proceedings of the XXX Incontro Ackema, P. 2010. Restrictions
di Grammatica Generativa, on subject extraction: A PF
 References 463

interface account. In Interfaces Akiyama, M. 1994. On quantifier

in Linguistics: New Research floating. English Linguistics 11:
Perspectives, R. Folli & C Ulbrich 100–22.
(eds.) 225–41. Oxford: Oxford Akiyama, T. 2002. The infinitival
University Press. relative clause in English: An
Ackema, P. & Neeleman, A. 2003. analysis based on the British
Context-sensitive spell-out. Natural National Corpus. PhD diss.
Language and Linguistic Theory 21: Lancaster University.
681–735. Akmajian, A. 1979. Aspects of the
Ackerman, L., Frazier, M. & Yoshida, Grammar of Focus in English. New
M. 2014. Resumptive pronouns York: Garland.
salvage island violations in Akmajian, A., Demers, R. A. &
forced-choice tasks. Poster Harnish, R. H. 1985. Linguistics,
presented to the 27th Annual 2nd ed. Cambridge, MA: MIT
CUNY Conference on Human Press.
Sentence Processing, Ohio State Alami, L. 2011. Word order, agreement
University, 14 March. and resumption in the Tashelhit
Adger, D. 2003. Core Syntax: A variety of Berber. PhD diss.
minimalist approach. Oxford: University of Essex.
Oxford University Press. Alamillo, A. R. & Schwenter, S. A.
———2007. Three domains of finiteness: 2007. Null objects and neuter
A minimalist perspective. In lo: A cross-dialectal variationist
Finiteness: Theoretical and analysis. In Selected Proceedings
empirical foundations, I. Nikolaeva of the Third Workshop on Spanish
(ed.) 23–58. Oxford: Oxford Sociolinguistics, J. Holmquist,
University Press. A. Lorenzino & L. Sayahi (eds.)
Adger, D. & Ramchand, G. 2005. Merge 113–21. Somerville, MA: Cascadilla
and Move: Wh‑dependencies Proceedings Project.
revisited. Linguistic Inquiry 36: Al-Bataineh, H. 2020. Exclamatives
161–93. are nonsententials: Evidence
Åfarli, T. 1989. Passive in Norwegian from Arabic and other languages.
and in English. Linguistic Inquiry Ms. Memorial University of
20: 101–8. Newfoundland.
———1994. A promotion analysis of Alcázar, A. & Saltarelli, M. 2014. The
restrictive relative clauses. The Syntax of Imperatives. Cambridge:
Linguistic Review 11: 81–100. Cambridge University Press.
Åfarli, T. & Creider, Ch. 1987. Alexiadou, A. 1994. Issues in the
Nonsubject Pro-Drop in Norwegian. syntax of adverbs. PhD diss.
Linguistic Inquiry 18: 339–45. University of Potsdam.
Aikhenvald, A. Y. 2010. Imperatives ———1997. Adverb Placement: A case
and commands. New York: Oxford study in asymmetric syntax.
University Press. Amsterdam: Benjamins.
464 References

Alexiadou, A. & Anagnastopoulou, E. Movement in the Determiner

1998. Parametrizing Agr: Word- Phrase. Amsterdam: Benjamins.
order, V‑movement, and EPP- Alexopoulou, T. 2006. Resumption in
checking. Natural Language and relative clauses. Natural Language
Linguistic Theory 16: 491–539. and Linguistic Theory 24: 57–111.
———1999. Raising without infinitives ———2010. Truly intrusive: Resumptive
and the nature of agreement. pronominals in questions and
Proceedings of the West Coast relative clauses. Lingua 120:
Conference on Formal Linguistics 485–505.
18: 15–25. Alexopoulou, T. & Keller, F. 2002.
———2001. The subject-in-situ Resumption and locality: A
generalization and the role of case crosslinguistic experimental
in driving computations. Linguistic study. Chicago Linguistic Society
Inquiry 32: 193–231. Proceedings 38: 1–14.
Alexiadou, A., Anagnostopoulou, ———2007. Locality, cyclicity and
E. & Schäfer, F. 2015. External resumption: At the interface
Arguments in Transitivity between the grammar and the
Alternations: A layering approach. human sentence processor.
Oxford: Oxford University Press. Language 83: 110–60.
———2018. Passive. In Syntactic Al Khalaf, E. 2019. Floating quantifiers
Structures After 60 Years: are autonomous phrases: A
The impact of the Chomskyan movement analysis. Glossa 4(1):
revolution in linguistics, 89, 1–23.
N. Hornstein, H. Lasnik, Allerton, D. J. 1975. Deletion and
P. Patel-Grosz & C. Yang (eds.). proform reduction. Journal of
Berlin: De Gruyter Mouton. Linguistics 11: 213–37.
Alexiadou, A., Iordăchioaia, G. & Ambar, M. 1999. Aspects of the syntax
Marchis, M. 2012. In support of focus in Portuguese. In The
of long-distance agree. In Grammar of Focus, G. Rebuschi &
Local Modeling of Non-local L. Tuller (eds.) 23–54. Amsterdam:
Dependencies, A. Alexiadou, G, Benjamins.
Müller & T. Kiss (eds.) 55–81: ———2002. Wh‑questions and
Tübingen: Niemeyer. wh‑exclamatives unifying mirror
Alexiadou, A., Law, P., Meinunger, A. effects. In Romance Languages
& Wilder, C. (eds.) 2000. The Syntax and Linguistic Theory 2000:
of Relative Clauses. Amsterdam: Selected papers from ‘Going
Benjamins. Romance 2000’, C. Beyssade, R.
Alexiadou, A., & Schäfer, F. 2013. Non- Bok-Bennema, F. Drijkoningen,
Canonical Passives. Amsterdam: & P. Monachesi (eds.) 15–40.
Benjamins. Amsterdam: Benjamins.
Alexiadou, A. & Wilder, C. (eds.) ———2007. Verb Movement and tense:
1998. Possessors, Predicates and EPP and T‑completeness. In
 References 465

Proceedings of the XXXII Incontro of Linguistic Evidence 2020:

di Grammatica Generativa, Linguistic theory enriched by
M. C. Picchi & A. Pona (eds.) 1–20. experimental data, R. Hörnig, S.
Alessandria: Edizioni dell’Orso. von Weitersheim, A. Konietsko
Amritavalli, R. 2013. Separating tense & S. Featherston (eds.). 519–40.
and finiteness: Anchoring in Tübingen: University of Tübingen.
Dravidian. Natural Language & Arnold, D. J. 2004. Non-restrictive
Linguistic Theory 32: 283–306. relative clauses in construction-
Anagnastopoulou, E., van Riemsdijk, H. based HPSG. In Proceedings of
& Zwarts, F. 1997 (eds.). Materials the 11th International Conference
on Left Dislocation. Amsterdam: on Head-Driven Phrase Structure
Benjamins. Grammar, S. Müller (ed.) 27–47.
Andersen, G. 1995. Omission of the Stanford, CA: CSLI publications.
primary verbs BE and HAVE ———2007. Non-restrictive relatives are
in London teenage speech: A not orphans. Journal of Linguistic
sociolinguistic study. MA thesis. 43: 271–309.
University of Bergen. Arregi, K. 2016. Focus projection
Anderson, J. 1977. On Case Grammar. theories. In The Oxford Handbook
London: Croom Helm. of Information Structure, C. Féry &
———2007. Finiteness, mood and S. Ishihara (eds.) 185–202. Oxford:
morphosyntax. Journal of Oxford University Press.
Linguistics 43: 1–32. Arregi, K. & Nevins, A. 2012.
Angelopoulos, N., Collins, C. & Terzi, Morphotactics: Basque auxiliaries
A. 2020. Greek and English and the structure of spellout.
passives and the role of by‑phrases. Dordrecht: Springer.
Glossa 5: 90, 1–29. Arteaga, D. 1998. On null objects in
Aoun, J. 2000. Resumption and last Old French. In Romance Linguistics:
resort. Documentação de Estudos Theoretical perspectives, A.
em Lingüística Teórica e Aplicada Schwegler, B. Tranel & M. Uribe-
16: 13–43. Etxebarria (eds.) 1–12. Amsterdam:
Aoun, J., Choueiri, L. & Hornstein, N. Benjamins.
2001. Resumption, movement, and Asudeh, A. 2002. Richard III. Chicago
derivational economy. Linguistic Linguistic Society Proceedings 38:
Inquiry 32: 371–403. 31–46.
Aoun, J. & Li, Y. A. 2003. Essays on the ———2004. Resumption as resource
Representational and Derivational management. PhD diss. Stanford
Nature of Grammar. Cambridge, University.
MA: MIT Press. ———2011a. Towards a unified theory
Armenante, G. & Braun, J. 2022. of resumption. In Resumptive
Fake past in conditionals and Pronouns at the Interfaces, A.
attitude reports: A crosslinguistic Rouveret (ed.) 121–87. Amsterdam:
correlation. In Proceedings Benjamins.
466 References

———2011b. Copy Raising and formal Authier, J.-M. & Reed, L. 2005. The
variation. Handout for talk diverse nature of non-interrogative
at University of Edinburgh, Wh. Linguistic Inquiry 36: 635–47.
24 November. ———2009. False negatives on
———2012. The Logic of Pronominal Agentivity: The case of
Resumption. Oxford: Oxford get‑passives. Unpublished paper,
University Press. Pennsylvania State University.
Asudeh, A. & Toivonen, I. 2006. Aygen-Tosun, G. 1998. The split INFL
Expletives and the syntax and hypothesis in Turkish. MA thesis.
semantics of Copy Raising. Boğaziçi University.
In Proceedings of the LFG08 Badan, L. 2007. High and low
Conference, M. Butt & T. H. King periphery: A comparison between
(eds.) 68–88. Stanford, CA: CSLI Italian and Chinese. PhD diss.
Publications. University of Padua.
———2007. Copy Raising and its Badan, L. & Cheng, L. 2015.
consequences for perception Exclamatives in Mandarin Chinese.
reports. In Architectures, Rules, Journal of East Asian Linguistics
and Preferences: Variations on 24: 383–413.
themes by Joan W Bresnan, A. Badan, L. & Crocco, C. 2019. Focus
Zaenen, J. Simpson, T. H. King, J. in Italian echo wh‑questions:
Grimshaw, J. Maling & C. Manning An analysis at syntax-prosody
(eds.) 49–67. Stanford, CA: CSLI interface. Probus 31: 29–73.
Publications. Badan, L. & Del Gobbo, F. 2010. On
———2012. Copy Raising and Perception. the syntax of topic and focus
Natural Language and Linguistic in Chinese. In The Cartography
Theory 30: 321–80. of Syntactic Structures, Vol. 5:
Authier, J.-M. 1988. Null object Mapping the left periphery, P.
construction in Kinande. Natural Benincà & N. Munaro (eds.) 63–90.
Language and Linguistic Theory 6: Oxford: Oxford University Press.
19–37. Bae, T.-W. 2016. Subject ellipsis in
———1989. Arbitrary null objects and spoken English. SNU Working
unselective binding. In The Null Papers in English Linguistics and
Subject Parameter, O. Jaeggli & K. Language 14: 1–22.
Safir (eds.) 45–67. Dordrecht: Reidel. Bailey, L. 2011. Null subjects in
———1992. Iterated CPs and embedded Northeast English. Newcastle
Topicalization. Linguistic Inquiry Working Papers in Linguistics 17:
23: 329–36. 23–45.
Authier, J.-M. & Haegeman, L. 2016. Baker, C. L. 1970. Notes on the
On the syntax and semantics of description of English questions:
Mirative Focus Fronting in French. The role of an abstract question
Ms. Pennsylvania State University morpheme. Foundations of
and Ghent University. Language 6: 197–219.
 References 467

Baker, M. 1988. Incorporation. Chicago: Bayer, J. 1984a. Towards an explanation

Chicago University Press. of certain that-t phenomena:
———1997. Thematic roles and syntactic The COMP node in Bavarian. In
structure. In Elements of Grammar, Sentential Complementation, W.
L. Haegeman (ed.) 73–137. de Geest & Y. Putseys (eds.) 23–32.
Dordrecht: Kluwer. Dordrecht: Foris.
———2008. The Syntax of Agreement ———1984b. COMP in Bavarian syntax.
and Concord. Cambridge: The Linguistic Review 3: 209–74.
Cambridge University Press. ———2010. Wh‑drop and recoverability.
Baker, M., Johnson, K. & Roberts, I. In Structure Preserved: Studies
1989. Passive arguments raised. in syntax for Jan Koster, J. W.
Linguistic Inquiry 20: 219–51. Zwart & M. de Vries (eds.) 31–40.
Baltin, M. 1978. Toward a theory of Amsterdam: Benjamins.
movement rules. PhD diss. MIT. ———2014. Syntactic and phonological
———1982. A landing site theory of properties of wh‑operators and
movement rules. Linguistic Inquiry wh‑movement in Bavarian. In
13: 1–38. Bavarian Syntax: Contributions to
———1995. Floating quantifiers, PRO the theory of syntax, G. Grewendorf
and predication. Linguistic Inquiry & H. Weiss (eds.) 23–50.
26: 199–248. Amsterdam: Benjamins.
———2010. The nonreality of doubly ———2015. Doubly-filled Comp, wh‑head
filled Comps. Linguistic Inquiry 41: movement, and derivational
331–5. economy. In Representing Structure
Banfield, A. 1982. Unspeakable in Phonology and Syntax, M. van
Sentences: Narration and Oostendorp & H. van Riemsdijk
representation in the language (eds.) 7–40. Berlin: Mouton de
of fiction. Boston: Routledge and Gruyter.
Kegan Paul. ———2016. Doubly-filled comp,
Bárány, A. 2015. Hungarian object wh‑head‑movement, and the
agreement with personal pronouns. doubly-filled-comp‑filter. Ms.
Proceedings of the West Coast Universität Konstanz.
Conference on Formal Linguistics Bayer, J. & Brandner, E. 2008.
32: 208–17. On wh‑head‑movement and
Barbiers, S., Bennis, H. & de Vogelaar, the doubly-filled-comp‑filter.
G. 2005. Syntactic Atlas of the Proceedings of the West Coast
Dutch Dialects, Vol. 1. Amsterdam: Conference on Formal Linguistics
Amsterdam University Press. 26: 87–95.
Barton, E. 1998. The grammar of Bayer, J. & Dasgupta, P. 2016.
telegraphic structures: Sentential Emphatic topicalization and the
and nonsentential derivation. structure of the left periphery:
Journal of English Linguistics 26: Evidence from German and Bangla.
37–67. Syntax 19: 1–45.
468 References

Becquet, K. 2000. L’omission du sujet Beltrama, A. 2013. Intrusive but

grammatical dans les journaux not intruders. The processing of
intimes. MA thesis. Université resumptive pronouns in Italian and
Charles de Gaulle, Lille. English. Ms., University of Chicago.
Beijer, F. 2002. The syntax and Beltrama, A. & Xiang, M. 2016.
pragmatics of exclamations Unacceptable but comprehensible:
and other expressive/emotional The facilitation effect of resumptive
utterances. Working Papers in pronouns. Glossa: A Journal of
Linguistics 2: 1–22, Department of General Linguistics 1: 1–24.
English, University of Lund. Bender, E. 1999. Constituting
Beck, S. 2006. Intervention effects context: Null objects in English
follow from focus interpretation. recipes revisited. University of
Natural Language Semantics 14: Pennsylvania Working Papers in
1–56. Linguistics 6: 53–68.
Beck, S. & Kim, S.-S. 2006. ———2000. Syntactic variation and
Intervention effects in alternative linguistic competence: The case of
questions. Journal of Comparative AAVE copula absence. PhD diss.
German Linguistics 9: 165–208. Stanford University.
Becker, M. 2006. There began to be Benincà, P. 1995. Il tipo esclamativo.
a learnability puzzle. Linguistic In Grande Grammatica Italiana
Inquiry 37: 441–56. di Consultazione, Vol. 3, L. Renzi,
Bell, A. 1991. The Language of the G. Salvi & A. Cardinaletti (eds.)
News Media. Oxford: Blackwell. 127–52. Bologna: Il Mulino.
Belletti, A. 1990. Generalized Verb ———1996. La struttura della frase
Movement: Aspects of verb syntax. esclamativa alla luce del dialetto
Turin: Rosenberg and Sellier. padovano. In Italiano e Dialetti nel
———2001. Agreement projections. In Tempo. Saggi di Grammatica per
The Handbook of Contemporary Giulio C. Lepschy, P. Benincà, G.
Syntactic Theory, M. Baltin & C. Cinque, T. De Mauro & N. Vincent
Collins (eds.) 483–510. Oxford: (eds.) 23–43. Roma: Bulzoni.
Blackwell. ———2001. The position of topic and
———2004a. Aspects of the low IP area. focus in the left periphery. In
In The Cartography of Syntactic Current Studies in Italian Syntax:
Structures, Vol. 2: The structure Essays Offered to Lorenzo Renzi,
of IP and CP, L. Rizzi (ed.) 16–51. G. Cinque & G. Salvi (eds.) 39–64.
Oxford: Oxford University Press. Amsterdam: Elsevier-North
———2004b (ed). The Cartography Holland.
of Syntactic Structures, Vol. 3: ———2003. La frase relativa in
Structures and beyond. Oxford: fiorentino antico. Paper presented
Oxford University Press. to the V° Incontro di Dialettologia,
———2009. Structures and Strategies. University of Bristol, 26–27
London: Routledge. September.
 References 469

———2006. A detailed map of the left Benincà, P. & Poletto C. 2004. Topic,
periphery of medieval Romance. Focus and V2: Defining the CP
In Crosslinguistic Research in sublayers. In The Cartography
Syntax and Semantics: Negation, of Syntactic Structures, Vol. 2:
tense, and clausal architecture, R. The structure of IP and CP, L.
Zanuttini, H. Campos, E. Herburger Rizzi. (ed.) 52–75. Oxford: Oxford
& P. Portner (eds.) 53–86. University Press.
Washington, DC: Georgetown Benmamoun, E. 1999. The syntax of
University Press. quantifiers and quantifier float.
———2010. La periferia sinistra. In Linguistic Inquiry 30: 621–42.
Grammatica dell’Italiano Antico, Benmamoun, E., Bhatia, A. & Polinsky,
G. Salvi & L. Renzi (eds.) 27–59. M. 2009. Closest conjunct
Bologna: Il Mulino. agreement in head-final languages.
———2012a. Lexical complementisers Linguistic Variation Yearbook 9:
and headless relatives. In 67–88.
The Cartography of Syntactic Bennis, H. 2006. Agreement, pro and
Structures, Vol. 7: Functional imperatives. In Arguments and
heads, L. Brugè, A. Cardinaletti, Agreement, P. Ackema, P. Brandt,
G. Giusti, N. Munaro & C. Poletto M. Schloorlemmer & F. Weerman
(eds.) 29–41. Oxford: Oxford (eds.) 101–23. Oxford: Oxford
University Press. University Press.
———2012b. Frasi relative e strutture Bennis, H. & Haegeman, L. 1994.
copulari. In Per Roberto Gusmani: On the status of agreement and
Studi in ricordo, V. Orioles & P. relative clauses in West Flemish.
Borghello (eds.) 251–67. Udine: In Sentential Complementation, W.
Forum editrice. de Geest & Y. Putseys (eds.) 33–55.
Benincà, P. & Cinque, G. 2010. La Dordrecht: Foris.
frase relativa. In Grammatica Berger, M. 2015. The subject position
dell’Italiano Antico, G. Salvi & L. in Spanish nominalized infinitives.
Renzi (eds.) 469–507. Bologna: Il University of Pennsylvania Working
Mulino. Papers in Linguistics 21: 1–10.
———2014. Kind-defining relative Berizzi, M. 2010. Interrogatives and
clauses in the diachrony of Italian. relatives in some varieties of English.
In Diachrony and Dialects, P. PhD diss. University of Padua.
Benincà, A. Ledgeway & N. Vincent Berizzi, M. & Rossi, S. 2010.
(eds.) 257–78. Oxford: Oxford ‘Something here what made me
University Press. think.’ Some new views on relative
Benincà, P. & Munaro, N. 2010 (eds.). what in the dialects of English.
The Cartography of Syntactic LangUE Proceedings 2009: 14–26,
Structures, Vol. 5: Mapping the University of Essex.
left periphery. Oxford: Oxford Bernstein, J. B. 1993. Topics in the
University Press. syntax of nominal structures across
470 References

Romance. PhD diss. City University conjunctions. Natural Language and

of New York. Linguistic Theory 31: 951–1013.
———2001. The DP hypothesis: Bianchi, V. 1999. Consequences of
Identifying clausal properties in the Antisymmetry: Headed Relative
nominal domain. In The Handbook Clauses. Berlin: Mouton de Gruyter.
of Contemporary Syntactic Theory, ———2000. The raising analysis of
M. Baltin & C. Collins (eds.) relative clauses: A reply to Borsley.
536–61. Oxford: Blackwell. Linguistic Inquiry 31: 123–40.
Bertollo, S. & Cavall, G. 2012. The ———2003. On finiteness as logophoric
syntax of Italian free relative anchoring. In Tense and Point of
clauses: An analysis. Generative View, J. Guéron & L. Tasmovski
Grammar in Geneva 8: 59–76. (eds.) 213–46. Nanterre: Université
Beukema, F. & Koopmans, P. 1989. A Paris X.
Government-Binding perspective ———2008. Resumptives and LF chains.
on the imperative in English. Ms. University of Siena.
Journal of Linguistics 25: 417–36. Bianchi, V., Bocci, G. & Cruschina,
Bever, T. G., Carroll, J. M. & Hartig, R. S. 2016. Focus fronting,
1976. Analogy of ungrammatical unexpectedness and evaluative
sequences that are utterable and implicatures. Semantics and
comprehensible are the origins Pragmatics 9: 1–54.
of new grammars in language Bianchi, V. & Cruschina, S. 2016. The
acquisition and linguistic derivation and interpretation of
evolution. In An Integrated Theory polar questions with a fronted
of Linguistic Disability, T. G. Bever, focus. Lingua 170: 47–68.
J. J. Katz & D. T. Langendoen (eds.) Bianchi, V. & Frascarelli, M. 2010. Is topic
149–82. New York: T. Y. Crowell a root phenomenon? Iberia 2: 43–88.
Press. Biber, D. (1995) Dimensions of Register
Beyssade, C. 2009. Exclamation and Variation: A Cross-linguistic
presupposition. MIT Working comparison. Cambridge: Cambridge
Papers in Linguistics 60: 19–34. University Press.
Bhatt, R. 2002. The raising analysis Biber, D., Johansson, S., Leech, G.,
of relative clauses: Evidence from Conrad C. & Finegan E. 1999.
adjectival modification. Natural Longman Grammar of Spoken and
Language Semantics 10: 43–90. Written English. Harlow: Pearson.
———2015. Relative clauses and Biberauer, T. 2008. Semi null-
correlatives. In Syntax: Theory subject languages, expletives
and analysis volume I, T. Kiss & A. and expletive pro reconsidered.
Alexiadou (eds.) 708–49. Berlin: De Cambridge Occasional Papers in
Gruyter. Linguistics 4: 1–45.
Bhatt, R. & Walkow, M. 2013. Biloa, E. 2013. The Syntax of Tuki:
Locating agreement in grammar: A Cartographic Approach.
An argument from agreement in Amsterdam: Benjamins.
 References 471

Birner, B. J. & Ward, G. 1998a. perspective with evidence from

Information Status and Non- Italian. Amsterdam: Benjamins.
canonical Word Order in English. Boeckx, C. 2000. A note on
Amsterdam: Benjamins. contraction. Linguistic Inquiry 31:
———1998b. Discourse and information 357–66.
structure. Ms. Northwestern ———2001. Scope reconstruction and
University, IL. A‑movement. Natural Language
Bjorkman, B. & Zeijlstra, H. 2019. and Linguistic Theory 19: 503–48.
Checking up on (ϕ)-Agree. ———2003. Islands and Chains:
Linguistic Inquiry 50: 527–69. Resumption as derivational residue.
Bloomfield, L. 1935. Language. London: Amsterdam: Benjamins.
George Allen and Unwin. ———2007. Understanding Minimalist
Blythe, H. 2016. Resumption in English: Syntax. Oxford: Blackwell.
An investigation of usage and ———2008. Aspects of the Syntax of
acceptability. Cambridge Occasional Agreement. New York: Routledge.
Papers in Linguistics 9: 156–83. ———2012. Syntactic Islands.
Bobaljik, J. 1999. Adverbs: The Cambridge: Cambridge University
hierarchy paradox. GLOT Press.
International 4: 27–28. Boeckx, C. & Grohmann, K. K. 2005.
———2003. Floating quantifiers: Handle Left dislocation in Germanic. In
with care. In The Second GLOT Focus on Germanic Typology, W.
International State-of-the-Article Abraham (ed.) 131–44. Berlin:
Book, L. Cheng & R. Sybesma (eds.) Akadamie-Verlag.
107–48. Berlin: Mouton de Gruyter. Boeckx, C. & Hornstein, N. 2003.
———2009. Where’s phi? Agreement Reply to Control is not movement.
as a post-syntactic operation. In Linguistic Inquiry 34: 269–80.
Phi-Theory: Phi-features across ———2004. Movement under control.
interfaces and modules, D. Harbour, Linguistic Inquiry 35: 431–52.
D. Adger & S. Béjar (eds.) 295–328. ———2006a. The virtues of control as
Oxford: Oxford University Press. movement. Syntax 9: 118–30.
Bocci, G. 2004. Contrastive focalisation ———2006b. Control in Icelandic and
on topics and preverbal subjects theories of Control. Linguistic
in Italian. Rivista di Grammatica Inquiry 37: 591–606.
Generativa 29: 3–60. Boeckx, C. & Jeong, Y. 2004. The fine
———2007. Criterial positions and structure of intervention in syntax.
left periphery in Italian. Nanzan In Issues in Current Linguistic
Linguistics 3: 35–70. Theory: A Festschrift for Hong
———2009. On syntax and prosody in Bae Lee, C. Kwon & W. Lee (eds.)
Italian. PhD diss. University of 83–116. Seoul: Kyungjin.
Siena. Boeckx, C. & Lasnik, H. 2006.
———2013. The Syntax-Prosody Intervention and repair. Linguistic
Interface: A Cartographic Inquiry 37: 150–55.
472 References

Boef, E. 2013. Doubling in English Language and Linguistics

relative clauses: Aspects of 13: 409–31.
morphosyntactic microvariation in Bošković, Z. 2001. Floating quantifiers
Dutch. PhD diss. Meertens Instituut and θ‑role assignment. Proceedings
(KNAW), University of Utrecht. of the North East Linguistic Society
Bolinger, D. 1977. Meaning and Form. 31: 59–78.
London: Longman. ———2002 A‑Movement and the EPP.
Bondaruk, A. 1995. Resumptive Syntax 5: 167–218.
pronouns in English and Polish. In ———2004. Be careful where you float
Licensing in Syntax and Phonology, your quantifiers. Natural Language
E. Gussmann (ed.) 27–55. Lublin: and Linguistic Theory 22: 681–742.
Folium. ———2007. On the locality and
Borer, H. 1984. Restrictive relatives in motivation of Move and Agree:
modern Hebrew. Natural Language An even more minimal theory.
and Linguistic Theory 2: 219–60. Linguistic Inquiry 38: 589–644.
Borise, L. 2019. Phrasing is key: The ———2009. On relativization strategies
syntax and prosody of focus and resumptive pronouns. In
in Georgian. PhD diss. Harvard Studies in Formal Slavic Phonology,
University. Morphology, Syntax, Semantics,
Börjeson, K. & Müller, G. 2020. Long- and Information Structure:
distance agreement and locality: A Proceedings of Formal Description
reprojection approach. In Agree to of Slavic Languages 7, Leipzig
Agree: Agreement in the minimalist 2007, G. Zybatow, U. Junghanns,
program, P. W. Smith, J. Mursell & D. Lenertová & P. Biskup (eds.)
K. Hartmann (eds.) 307–46. Berlin: 79–93. Berlin: Peter Lang.
Language Science Press. ———2011a. Rescue by PF deletion,
Borsley, R. D. 1984. Free relatives in traces as (non)interveners, and the
Polish and English. In Contrastive that-trace effect. Linguistic Inquiry
Linguistics: Prospects and Problems, 42: 1–44.
J. Fisiak (ed.) 1–18. Berlin: Mouton. ———2011b. On unvalued
———1992. More on the difference uninterpretable features.
between English restrictive and Proceedings of the Annual Meeting
non-restrictive relative clauses. of the North East Linguistics
Journal of Linguistics 28: 139–48. Society 39: 109–20.
———1997. Relative clauses and ———2012. On NPs and clauses. In
the theory of phrase structure. Discourse and Grammar: From
Linguistic Inquiry 28: 629–47. sentence types to lexical categories,
———2001. More on the raising analysis G. Grewendorf & T. E. Zimmerman
of relative clauses. Ms. University (eds.) 179–242. Berlin: de Gruyter.
of Essex. ———2016. On the timing of labeling:
———2010. On so-called transitive Deducing Comp‑trace effects, the
expletives in Belfast English. Subject Condition, the Adjunct
 References 473

Condition, and tucking in from Breivik, L. E. 1983. Existential there:

labeling. The Linguistic Review 33: A synchronic and diachronic
17–66. study. Bergen: Norwegian Research
———2020. On smuggling, the Council.
freezing ban, labels and tough- Bresnan, J. 1970. On complementizers:
constructions. In Smuggling in Toward a syntactic theory of
Syntax, A. Belletti & C. Collins complement types. Foundations of
(eds.) 53–95. Oxford: Oxford Language 6: 297–321.
University Press. ———1972. Theory of complementation
Bosque, I. 2017 (ed.) Advances in the in English syntax. PhD diss. MIT
Analysis of Spanish Exclamatives. (published as Bresnan 1979).
Columbus, OH: Ohio State ———1976. Evidence for a theory
University Press. of unbounded transformations.
Bouchard, D. 1989. Null objects and Linguistic Analysis 2: 353–93.
the theory of empty categories. In ———1979. Theory of Complementation
Studies in Romance Linguistics: in English Syntax. New York:
Selected papers from the XVIIth Garland.
linguistic symposium on Romance Bresnan, J. & Grimshaw, J. 1978. The
languages at Rutgers University, syntax of free relatives in English.
C. Kirschner & J. Decesaris (eds.) Linguistic Inquiry 9: 331–91.
33–49. Amsterdam: Benjamins. Brillman, R. J. & Hirsch, A. 2016. An
Bowers, J. 1973. Grammatical relations. anti-locality account of English
PhD diss. MIT. subject/non-subject asymmetries.
———2002. Transitivity. Linguistic Chicago Linguistic Society
Inquiry 33: 183–224. Proceedings 50: 73–88.
Branan, K. 2018. Attraction at a Brisson, C. M. 1998. Distributivity,
distance: A’‑movement and case. maximality and floating quantifiers.
Linguistic Inquiry 49: 409–40. PhD diss. Rutgers University.
Brandi, L. & Cordin, P. 1989. Two ———2000. Floating quantifiers as
Italian dialects and the null subject adverbs. In Proceedings of the
parameter. In The Null Subject Fifteenth Eastern States Conference
Parameter, O. Jaeggli & K. Safir on Linguistics, R. Daly & A. Riehl
(eds.) 111–43. Dordrecht: Kluwer. (eds.) 13–24. Ithaca, NY: Cornell
Brandner, E. 2010. On the syntax of University.
verb-initial exclamatives. Studia Brody, M. 1990. Some remarks on
Linguistica 64: 81–115. the focus field in Hungarian. UCL
Branigan, P. 1992. Subjects and Working Papers in Linguistics 2:
complementisers. PhD diss. MIT. 1–25.
Branigan, P. & MacKenzie, M. 2001. Brook, M. 2016. Syntactic categories
Altruism, A′‑movement, and informing variationist analysis: The
object agreement in Innu-aimûn. case of English copy‑raising. PhD
Linguistic Inquiry 33: 385–407. diss. University of Toronto.
474 References

———2018. Taking it up a level: Sandt (eds.) 142–65. Cambridge:

Copy‑raising and cascaded tiers of Cambridge University Press.
morphosyntactic change. Language ———2003. On D‑trees, beans and
Variation and Change 30: 231–60. B-accents. Linguistics and
Browning, M. A. 1996. CP recursion Philosophy 26: 511–45.
and that-t effects. Linguistic ———2004. Negative Inversion. Draft of
Inquiry 27: 237–55. paper presented to a meeting of the
Brownlow, O. 2011. Towards a unified North East Linguistic Society. Ms.
analysis of the syntax and UCLA.
semantics of get constructions. PhD ———2009. Towards a typology of
diss. Queen Mary University of focus realization. In Information
London. Structure, M. Zimmermann &
Bruening, B. 2001. Syntax at the edge: C. Féry (eds.) 177–205. Oxford:
Cross-clausal phenomena and the Oxford University Press.
syntax of Passamaquoddy. PhD ———2016. (Contrastive) topic. In The
diss. MIT. Oxford Handbook of Information
———2009. Selectional asymmetries Structure, C. Féry & S. Ishihara
between CP and DP suggest that (eds.) 64–85. Oxford: Oxford
the DP hypothesis is wrong. University Press.
University of Pennsylvania Working Burke, I. 2017. Wicked which: The
Papers in Linguistics 15: 27–36. linking relative in Australian
———2013. By phrases in passives and English. Australian Journal of
nominals. Syntax 16: 1–41. Linguistics 37: 356–86.
———2014. Precede and command Burton, S & Grimshaw, J. 1992.
revisited. Language 90: 342–88. Coordination and VP‑internal
Bruening, B., Dinh, X. & Kim. L. 2015. subjects. Linguistic Inquiry 23:
Selection, idioms and the structure 305–13.
of nominal phrases with and Burzio, L. 1986. Italian Syntax.
without classifiers. Ms. University Dordrecht: Reidel.
of Delaware. Cable, S. 2005. Free relatives in Tlingít
Brugé, L., Cardinaletti, A., Giusti, G., and Haida: Evidence that the
Munaro, N. & Poletto, C. 2012. mover projects. Ms. University of
Functional Heads. Oxford: Oxford Massachusetts.
University Press. ———2007. The grammar of Q: Q‑particles
Brunetti, L. 2004. A Unification of and the nature of wh‑fronting, as
Focus. Padua: Unipress. revealed by the wh‑questions of
Büring, D. 1997. The Meaning of Topic Tlingit. PhD diss. MIT.
and Focus: The 59th Street Bridge ———2008. Question particles and
accent. London: Routledge. the nature of wh‑fronting. In
———1999. Topic. In Focus: Linguistic, Quantification: A cross-linguistic
cognitive and computational perspective, L. Matthewson (ed.)
perspectives, P. Bosch & R. van der 105–78. Bingley: Emerald.
 References 475

———2010a. The Grammar of Q: and Semantics of Spatial P, A.

Q‑particles, Wh‑Movement, and Asbury, J. Dotlačil, B. Gehrke & R.
pied-piping. Oxford: Oxford Nouwen (eds.) 365–85. Amsterdam:
University Press. Benjamins.
———2010b. Against the existence of ———2009. The nominal nature of when,
pied-piping: Evidence from Tlingit. where, and how: Evidence from
Linguistic Inquiry 41: 563–94. free relatives. Linguistic Inquiry 40:
———2012. The optionality of movement 155–75.
and EPP in Dholuo. Natural Caponigro, I. & Schütze, C.T. 2003.
Language and Linguistic Theory 30: Parameterizing passive participle
651–97. movement. Linguistic Inquiry 34:
Camilleri, M. 2018. On Raising and 293–308.
Copy Raising in Maltese. In The Caponigro, I., Torrence, H. & Cisneros,
Languages of Malta, P. Paggio C. 2013. Free relative clauses in two
& A. Gatt (eds.) 171–201. Berlin: Mixtec languages. International
Language Science Press. Journal of American Linguistics 79:
Cann, R., Kaplan, T. & Kempson, 61–96.
R. 2005. Data at the grammar- Cardinaletti, A. 1997. Subjects and
pragmatics interface: The case of clause structure. In The New
resumptive pronouns in English. Comparative Syntax, L. Haegeman
Lingua 115: 1551–78. (ed.) 33–63. London: Longman.
Capelle, B. 2005. Particle patterns ———2004. Towards a cartography
in English: A comprehensive of syntactic positions. In The
coverage. PhD diss. University of Structure of CP and IP, L. Rizzi
Leuven. (ed.) 115–65. Oxford: Oxford
Caponigro, I. 2002. Free relatives University Press.
as DPs with a silent D and a CP ———2009 On a (wh‑)moved topic in
complement. In Proceedings of Italian, compared to Germanic.
WECOL 2000, V. Samiian (ed.) Linguistics Today 141: 3–40.
140–50. Fresno, CA: California Cardinaletti, C., Cinque, G. & Endo,
State University. Y. (eds.). 2014. On Peripheries:
———2004. The semantic contribution Exploring clause initial and clause
of wh‑words and type shifts: final positions. Tokyo: Hituzi Syobe
Evidence from free relatives Publishing.
crosslinguistically. In Proceedings Cardinaletti, A. & Starke, M. 1999. The
of the 14th Semantics and typology of structural deficiency:
Linguistic Theory Conference, A case study of the three classes of
SALT XIV, R. B. Young (ed.) 38–55. pronouns. In Clitics in the Languages
Ithaca, NY: Cornell University. of Europe, H. van Riemsdijk (ed.)
Caponigro, I. & Pearl, L. 2008. Silent 145–233. Berlin: Mouton de Gruyter.
prepositions: Evidence from Carlson, G. 1977. Amount relatives.
free relatives. In The Syntax Language 53: 520–42.
476 References

Carroll, S. 1983. On for and the ———2008. Deconstructing

constituency of infinitives. exclamations. Catalan Journal of
Toronto Working Papers in Linguistics 7: 41–90.
Linguistics, 4. ———2010. An expressive answer: Some
Carstens, V. 2003. Rethinking considerations on the semantics
complementizer agreement: Agree and pragmatics of wh‑exclamatives.
with a case-checked goal. Linguistic Chicago Linguistic Society
Inquiry 34: 393–412. Proceedings 44: 3–17.
———2016. Delayed valuation: ———2019. On wh‑exclamatives and
A reanalysis of ‘upwards’ gradeability. An argument from
complementizer agreement and the Romance. Journal of Linguistics
mechanics of case. Syntax 19: 1–42. 57: 41–82.
Carstens, V. & Diercks, M. 2013. Cattell, R. 1976. Constraints on
Parameterizing case and activity: movement rules. Language 52:
Hyper‑raising in Bantu. Proceedings 18–50.
of the North East Linguistic Society Cecchetto, C. 2005. Reconstruction in
40: 99–118. relative clauses and the copy theory
Carter, R. & McCarthy, M. 1995. of traces. Linguistic Variation
Grammar and the spoken Yearbook 5: 73–103.
language. Applied Linguistics 16: Cecchetto, C. & Donati, C. 2010. On
141–58. labelling: Principle C and Head
Casielles-Suárez, E. 2004. The Syntax- Movement. Syntax 13: 241–78.
Information Structure Interface: ———2015. (Re)labeling. Cambridge, MA:
Evidence from Spanish and English. MIT Press.
New York: Routledge. Cerron-Palomino, A. 2015. Resumption
Castillo, C. 2009. English to‑infinitive or contrast? Non-standard
relative clauses. Folia Linguistica pronouns in Spanish relative
Historica 27: 135–53. clauses. Spanish in Context 12:
Castillo, J. C. C., Drury, J. E., & 349–72.
Grohmann, K. K. 2009. Merge over Chaiphet, K. 2017. Aspects of quantifier
Move and the Extended Projection float in Thai. MA thesis. City
Principle: MOM and EPP revisited. University of New York.
Iberia 1: 53–114. Chao, W. & Sells, P. 1983. On the
Castroviejo, M.E. 2006. Wh‑exclamatives interpretation of resumptive
in Catalan. PhD diss. University of pronouns. Proceedings of the North
Barcelona. East Linguistic Society 13: 47–61.
———2007. A degree-based account of Chapman, C. 2013. A revised account
wh‑exclamatives in Catalan. In of the EPP in French: A semantics-
Proceedings of Sinn und Bedeutung, based analysis. McGill Working
E. Puig-Waldemüller (ed.) 134–49. Papers in Linguistics 23: 1–23.
Barcelona: Universitat Pompeu Chaves, R. 2012. On the grammar
Fabra. of extraction and co-ordination.
 References 477

Natural Language and Linguistic August 27–31, 1962. H. G. Lunt

Theory 30: 465–512. (ed.). The Hague: Mouton.
———2013. An expectation-based ———1964b. Current Issues in Linguistic
account of subject islands and Theory. The Hague: Mouton.
parasitism. Journal of Linguistics ———1965. Aspects of the Theory of
49: 285–327. Syntax. Cambridge, MA: MIT Press.
Checa-García, I. 2019. Resumptive ———1968. Noam Chomsky and Stuart
elements in Spanish relative Hampshire discuss the study of
clauses and processing difficulties: language. The Listener, 30 May,
A multifactorial analysis. Folia 79(2044): 687–91 .
Linguistica 53: 479–517. ———1970. Remarks on nominalization.
Cheng, L. 1997. On the Typology In Readings in English
of Wh‑Questions. New York: Transformational Grammar, R. A.
Garland. Jacobs & P. S. Rosenbaum (eds.)
Cheng, L. & Downing, L. 2012. Against 184–221. Waltham, MA: Ginn.
FocusP: Arguments from Zulu. ———1972 Language and Mind (enlarged
In Contrasts and Positions in edition). New York: Harcourt Brace
Information Structure, I. Kučerová Jovanovich.
& A. Neeleman (eds.) 247–66. ———1973. Conditions on
Cambridge: Cambridge University transformations. In A Festschrift
Press. for Morris Halle, S. R. Anderson
Chernilovskaya, A. 2014. Exclamativity & P. Kiparsky (eds.) 232–86. New
in discourse: Exploring the York: Holt, Rinehart and Winston.
exclamative speech act from a ———1975 The Logical Structure of
discourse perspective. PhD diss. Linguistic Theory. New York:
Utrecht University. Plenum Press.
Chernilovskaya, A., Condoravdi, C. & ———1977. On Wh‑movement. In
Lauer, S. 2012. On the discourse Formal Syntax, P. W. Culicover,
effects of wh‑exclamatives. T. Wasow & A. Akmajian (eds.)
Proceedings of the West Coast 71–132. New York: Academic
Conference on Formal Linguistics Press.
30: 109–19. ———1980. On binding. Linguistic
Chomsky, N. 1955. The Logical Inquiry 11: 1–46.
Structure of Linguistic Theory. ———1981. Lectures on Government and
Mimeo, MIT (published as Chomsky Binding. Dordrecht: Foris.
1975). ———1982. Some Concepts and
———1957. Syntactic Structures. The Consequences of the Theory
Hague: Mouton. of Government and Binding.
———1964a. The logical basis of Cambridge, MA: MIT Press.
linguistic theory. In Proceedings of ———1986a. Knowledge of Language:
the Ninth International Congress Its nature, origin and use. Praeger,
of Linguists, Cambridge, Mass., New York.
478 References

———1986b. Barriers. Cambridge, MA: Essays in honor of Jean-Roger

MIT Press. Vergnaud, R. Friedin, C. Otero &
———1989. Some notes on economy of M.-L. Zubizarreta (eds.) 133–65.
derivation and representation. MIT Cambridge, MA: MIT Press.
Working Papers in Linguistics 10: ———2013. Problems of projection.
43–74. Reprinted as chapter 2 of Lingua 130: 33–49.
Chomsky 1995. ———2014. Minimal recursion:
———1993. A minimalist program for Exploring the prospects. Studies in
linguistic theory. In The View from Theoretical Psycholinguistics 43:
Building 20: Essays in honor of 1–15.
Sylvain Bromberger, K. Hale & S. J. ———2015. Problems of projection:
Keyser (eds.) 1–52. Cambridge, MA: Extensions. In Structures,
MIT Press. Reprinted as chapter 3 Strategies, and Beyond: Studies
of Chomsky 1995. in honour of Adriana Belletti, E.
———1995. The Minimalist Program. Di Domenico, C. Hamann & S.
Cambridge, MA: MIT Press. Matteini (eds.) 3–16. Amsterdam:
———1998. Minimalist inquiries: The Benjamins.
framework. MIT Occasional Papers ———2021. Minimalism: Where are we
in Linguistics 15. Republished as now, and where can we hope to
Chomsky 2000. go? Gengo Kenkyu 160: 1–41.
———1999. Derivation by phase. MIT Chomsky, N. & Lasnik, H. 1977. Filters
Occasional Papers in Linguistics 18. and control. Linguistic Inquiry 8:
Republished as Chomsky 2001. 425–504.
———2000. Minimalist inquiries: The ———1993. The theory of principles
framework. In Step by Step: Essays and parameters. In Syntax:
on minimalism in honor of Howard An international handbook of
Lasnik, R. Martin, D. Michaels contemporary research, J. Jacobs,
& J. Uriagereka (eds.) 89–155. A. von Stechow, W. Sternefeld &
Cambridge, MA: MIT Press. T. Venneman (eds.) 506–69. Berlin:
———2001. Derivation by phase. In Mouton de Gruyter. Reprinted as
Ken Hale: A life in language, M. chapter 1 of Chomsky 1995.
Kenstowicz (ed.) 1–52. Cambridge, Chung, S. 1984. Identifiability and
MA: MIT Press. null objects in Chamorro. Berkeley
———2005. Three factors in language Linguistics Society Proceedings: 10:
design. Linguistic Inquiry 36: 1–22. 116–30.
———2007. Approaching UG from Cinque, G. 1977. The movement nature
below. In Interfaces+Recursion = of left dislocation. Linguistic
Language? U. Sauerland & H.-M. Inquiry 8: 397–412.
Gärtner (eds.) 1–29. Berlin: Mouton ———1978. Towards a unified
de Gruyter. treatment of island constraints.
———2008. On phases. In Foundational In Proceedings of the Twelfth
Issues in Linguistic Theory: International Congress of Linguists,
 References 479

W. U. Dressler & W. Meid (eds.) Cinque, G. & Rizzi. L. 2008. The

344–48. Innsbrücker Beiträge zur cartography of syntactic
Sprachwissenschaft. structures. Studies in Linguistics
———1983. Topic constructions in 2: 42–58.
some European languages and ———2010a. The cartography of
connectedness. Tilburg Studies in syntactic structures. In The
Language and Literature 4: 7–41. Oxford Handbook of Grammatical
Reprinted in Anagnastopoulou, van Analysis, B. Heine & H. Narrog
Riemsdijk & Zwarts, 1997. (eds.) 51–65. Oxford: Oxford
———1999. Adverbs and Functional University Press.
Heads. Oxford: Oxford University ———2010b (eds.). The Cartography
Press. of Syntactic Structures, Vol. 6:
———2002 (ed.) The Cartography of Mapping spatial PPs. Oxford:
Syntactic Structures, Vol. 1: The Oxford University Press.
structure of DP and IP. Oxford: Cirillo, R. J. 2009. The syntax of
Oxford University Press. Floating Quantifiers: Stranding
———2004. Issues in adverbial syntax. revisited. Utrecht: LOT.
Lingua 114: 683–710. ———2012. A fresh look at the debate on
———2008. Two types of non-restrictive floating quantifiers. Language and
relatives. Empirical Issues in Linguistics Compass 6: 796–815.
Syntax and Semantics 7: 99–137. Citko, B. 2000. Parallel merge and the
———2011. On double-headed relative syntax of free relatives. PhD diss.
clauses. Revista de Estudos State University of New York at
Linguisticos da Universidade do Stony Brook.
Porto 1: 67–91. ———2001. Deletion under identity in
———2013. Typological Studies: relative clauses. Proceedings of the
Word order and relative clauses. North East Linguistic Society 31:
Abingdon: Routledge. 131–45.
———2017a. On the status of functional ———2002. (Anti)reconstruction effects
categories (heads and phrases). in free relatives: A new argument
Language and Linguistics 18: against the Comp account.
521–76. Linguistic Inquiry 33: 507–11.
———2017b. On the double-headed ———2004. On headed, headless, and
analysis of ‘headless’ relative light-headed relatives. Natural
clauses. Ms. University of Venice. Language and Linguistic Theory 22:
———2020. The Syntax of Relative 95–126.
Clauses: A unified analysis. ———2005. On the nature of Merge:
Cambridge: Cambridge University External Merge, Internal Merge,
Press. and Parallel Merge. Linguistic
———2023. On Linearization: Toward a Inquiry 36: 475–96.
restrictive theory. Cambridge, MA: ———2008. An argument against
MIT Press.. assimilating appositive relatives to
480 References

coordinate structures. Linguistic ———2021. Principles of Argument

Inquiry 39: 633–55. Structure: A Merge-based
Claridge, C. 2012. The origins of how approach. Ms. New York University.
come and what … for. In English Collins, C. & Postal, P. M. 2014.
Historical Linguistics 2010: Classical NEG Raising: An essay on
Selected Papers from the Sixteenth the syntax of negation. Cambridge,
International Conference on English MA: MIT Press.
Historical Linguistics, I. Hegedüs & Collins, C. & Radford, A. 2015. Gaps,
A. Fodor (eds.) 177–95. Amsterdam: ghosts and gapless relatives in
Benjamins. spoken English. Studia Linguistica
Clemens, L. E., Morgan, A. M., 69: 191–235.
Polinsky, M. & Xiang, M. 2012. Collins, C. & Stabler, E. 2016. A
Listening to resumptives: An formalization of minimalist syntax.
auditory experiment. Poster Syntax 19: 43–78.
presented to the 25th Annual CUNY Collins, P. 2004. Exclamative clauses:
conference on Human Sentence A corpus-based account. Word 56:
Processing, City University of New 1–17.
York. Comrie, B. 1999. Relative clauses:
Cole, P. 1987. Null objects in Universal Structure and typology on the
Grammar. Linguistic Inquiry 18: periphery of standard English. In
597–612. The Clause in English, P. Collins
Collins, C. 1991. Why and how come. & D. Lee (eds.) 81–91. Amsterdam:
MIT Working Papers in Linguistics Benjamins.
15: 1–45. ———2002. Typology and language
———2005. A smuggling approach to acquisition: The case of relative
the passive in English. Syntax 8: clauses. In Typology and Second
81–120. Language Acquisition, A. G. Ramat
———2007. Home sweet home. NYU (ed.) 19–37. Berlin: Mouton de
Working Papers in Linguistics 1: Gruyter.
1–34. Conroy, A. 2006. The semantics of
———2015. Relative clause deletion. MIT how come: A look at how factivity
Working Papers in Linguistics 77: does it all. University of Maryland
57–69. Working Papers in Linguistics 14:
———2017. Merge(X,Y) = {X,Y}. In 1–24.
Labels and Roots, L. Bauke & A. Constant, N. 2014. Contrastive topic:
Blümel (eds.) 487–68. Berlin: De Meanings and realizations. PhD
Gruyter. diss. University of Massachusetts,
———2018. History of the by‑phrase in Amherst.
generative syntax (1957 to 2005). Contreras, H. 1976. A Theory of Word
Ordinary Working Grammarian. Order with Special Reference to
Blog. 17 September. [Link] Spanish. Amsterdam: North-
.ly/3FQ5qHS Holland.
 References 481

———1986. Spanish bare NPs and ———1997. Elaboration: A function

the ECP. In Generative Studies and a form. Ms. University of
in Spanish Syntax, I. Bordelois, Pennsylvania. Text of paper
H. Contreras & K. Zagona (eds.) presented to the 23rd annual
25–49. Dordrecht: Foris. meeting of the Berkeley Linguistics
———1987. Small clauses in Spanish Society.
and English. Natural Language and Cougnon, L-A. & Fairon, C. (eds.)
Linguistic Theory 5: 225–44. 2012. SMS Communication: A
———1991. On resumptive pronouns. linguistic approach. Amsterdam:
In Current Studies in Spanish Benjamins.
Linguistics, H. Campos & F. Cowart, W. 1997 Experimental Syntax:
Martinez-Gil (eds.) 143–63. Applying objective methods to
Washington, DC: Georgetown sentence judgments. Thousand
University Press. Oaks, CA: Sage Publications.
Corbett, G. 2006. Agreement. Cowper, E. 2016. Finiteness and
Cambridge: Cambridge University pseudofiniteness: In Finiteness
Press. Matters: On finiteness-related
Cormack, A. & Smith, N. 2000. phenomena in natural languages,
Fronting: The syntax and K. M. Eide (ed.) 47–78. Amsterdam:
pragmatics of ‘focus’ and ‘topic’. Benjamins.
UCL Working Papers in Linguistics Craenenbroeck, J. van 2020. Expletives,
12: 387–416. locatives and subject doubling.
Cornilescu, A. 1981. Non-restrictive In Linguistic Variation: Structure
relative clauses: An essay in and interpretation, L. Franco & P.
semantic description. Revue Lorusso (eds.) 661–90. Berlin: De
Roumaine de Linguistique 26: 41–67. Gruyter.
Costa, J. 2000. Adverbs as adjuncts Craenenbroeck, J. van & den Dikken,
to non-universal functional M. 2006. Ellipsis and EPP repair.
categories: Evidence from Linguistic Inquiry 37: 653–64.
Portuguese. In Adverbs and Cresswell, C. 2002. Resumptive
Adjunction, A. Alexiadou & P. pronouns, wh‑island violations,
Svenonius (eds.) 103–122. Potsdam: and sentence production. In
Institut für Linguistik, Universität Proceedings of the 6th International
Potsdam. Workshop on Tree-Adjoining
———2004. A multifactorial approach Grammar and Related Frameworks
to adverb placement: Assumptions, (TAG+6), 101–9. Venice: Università
facts and problems. Lingua 114: di Venezia.
711–53. Cruschina, S. 2006. Informational focus
Cote, S. 1996. Grammatical and in Sicilian and the left periphery.
discourse properties of null In Phases of Interpretation, M.
arguments in English. PhD diss. Frascarelli (ed.) 363–85. Berlin:
University of Pennsylvania. Mouton de Gruyter.
482 References

———2008. Discourse-related features Cukor-Avila, P. 1999. Stativity

and the syntax of peripheral and copular absence in AAVE:
positions: A comparative study Grammatical constraints at the
of Sicilian and other Romance subcategorical level. Journal of
languages. PhD diss. University of English Linguistics 27: 341–55.
Cambridge. Culicover, P. & Jackendoff, R. 2001.
———2010a. Fronting as focalization in Control is not movement. Linguistic
Sicilian. In Syntactic Variation: The Inquiry 30: 483–512.
Dialects of Italy. R. D’Alessandro, ———2005. Simpler Syntax. Oxford:
A. Ledgeway, I. Roberts (eds.) Oxford University Press.
247–60. Cambridge: Cambridge ———2006. Turn over control to the
University Press. semantics. Syntax 9: 131–52.
———2010b. Aspetti morfologici e Culicover, P. & Rochemont, M. 1983.
sintattici degli avverbi in siciliano. Stress and focus in English.
In Quaderni di Lavoro ASIt n.11: Language 59: 123–65.
Studi sui Dialetti della Sicilia, Culy, C. 1996. Null objects in English
J. Garzonio (ed.) 21–42. Padua: recipes. Language Variation and
Unipress. Change 8: 91–124.
———2011a. On the syntactic status Cummins, S. & Roberge, Y. 2004. Null
of sentential adverbs and modal objects in French and English.
particles. Sprachtypologie und In Contemporary Approaches to
Universalienforschung (STUF) 63: Romance Linguistics, J. Auger, C.
345–57. Clements & B. Vance (eds.) 121–38.
———2011b. Grammaticalization and Amsterdam: Benjamins.
the expression of evidentiality ———2005. A modular account of null
and epistemicity in Italian and in objects in French. Syntax 8: 44–64.
Sicilian. Unpublished handout, Cyrino, S. 1997. O Objeto Nulo no
University of Manchester. Português Brasileiro: Um estudo
———2012. Discourse-Related Features sintático-diacrônico. Londrina:
and Functional Projections. Oxford: Editora da UEL.
Oxford University Press. ———2004. Null complement
Cruschina, S. & Remberger, E.-M. anaphora and null objects
2008. Hearsay and reported speech: in Brazilian Portuguese. Ms.
Evidentiality in Romance. Rivista University of Campinas.
di Grammatica Generativa 33: [Link]
99–120. publication/242082650
———2017. Focus fronting. In Manual ———2021. On the syntax of null
of Romance Morphosyntax and objects in Brazilian Portuguese.
Syntax, E. Stark & A. Dufter (eds.) Handout for a conference on Null
502–35. Berlin: de Gruyter. Objects from a Crosslinguistic
Crystal, D. 2008. Txting: the Gr8 Db8. and Developmental Perspective,
Oxford: Oxford University Press. University of Campinas.
 References 483

Cyrino, S. & Matos, G. 2016. Null ———2016. Different languages –

objects and VP Ellipsis in European different sentence types? On
and Brazilian Portuguese. In The exclamative sentences. Language
Handbook of Portuguese Linguistics, and Linguistics Compass 10:
W. L Wetzels, S. Menuzzi & J. Costa 159–75.
(eds.). London: Wiley-Blackwell. Deal, A.R. 2009. The origin and content
Daeninck, B. 2019. Subject omission of expletives: Evidence from
on Twitter: An explorative case ‘selection’. Syntax 12: 285–323.
study of the syntactic constraints ———2015. Interaction and satisfaction
of subject omission in finite clauses in φ-agreement. Proceedings of the
in English Twitter posts. MA thesis. North East Linguistic Society 45:
Ghent University. 179–92.
Dąbrowska, E. 2010. Naive v. expert ———2016. Cyclicity and connectivity
intuitions: An empirical study in Nez Perce relative clauses.
of acceptability judgments. The Linguistic Inquiry 47: 427–70.
Linguistic Review 27: 1–23. De Cat, C. 2000. Towards a unified
Dalrymple, M. & Nikolaeva, I. 2011. analysis of French floating
Objects and Information Structure. quantifiers. Journal of French
Cambridge: Cambridge University Language Studies 10: 1–25.
Press. De Cesare, A. M. 2015. Defining focus
Danckaert, L. 2011. On the left modifiers in a cross-linguistic
periphery of Latin embedded perspective: A discussion based
clauses. PhD diss. University of on English, German, French and
Ghent. Italian. In Adverbs: Functional and
———2012. Latin Embedded Clauses: diachronic aspects, K. Pittner, D.
The left periphery. Amsterdam: Elsner & F. Barteld (eds.) 47–81.
Benjamins. Amsterdam: Benjamins.
Danon, G. 2011. Agreement and DP- Déchaine, R.-M. & Wiltschko, M. 2002.
internal feature distribution. Syntax Decomposing pronouns. Linguistic
14: 297–317. Inquiry 33: 409–42.
Davies, W. D. & Dubinsky, S. 2004. The de Clerk, K. 2010. NEG-shift in English:
Grammar of Raising and Control: A Evidence from PP adjuncts. In
course in syntactic argumentation. Proceedings of the 12th Seoul
Oxford: Blackwell. International Conference on
Davison, A. 1984. Syntactic Generative Grammar, 2010:
markedness and the definition of Movement in minimalism, D.-H. An
the sentence topic. Language 60: & J. T. Kim (eds.) 231–51. Seoul:
797–864. Hankuk.
d’Avis, F.-J. 2002. On the interpretation Deen, K. U. 2002. The acquisition of
of wh‑clauses in exclamative Nairobi Swahili: The morphosyntax
environments. Theoretical of inflectional prefixes and
Linguistics 28: 5–31. subjects. PhD diss. UCLA.
484 References

Dehé, N. 2002. Particle Verbs in Spanish complementizer system.

English: Syntax, information Probus 21: 23–49.
structure, and intonation. ———2013. El que citativo en español
Amsterdam: Benjamins. y otros elementos de la periferia
Dehé, N., Feldhausen, I. & Ishihara, S. oracional: Variación inter e
2011. The prosody-syntax interface: intralingüística. In Autour de ‘que’,
Focus, phrasing, language D. Jakob & K. Plooj (eds.) 47–69.
evolution. Lingua 121: 1863–69. Frankfurt: Peter Lang.
Delahunty, G. 1983. But sentential ———2014. Evidentiality and illocutionary
subjects do exist. Linguistic force: Spanish matrix que at
Analysis 12: 379–98. the syntax-semantics interface.
de Lange, J. 2004. Article omission in Linguistics Today 214: 217–51.
child speech and headlines. In OTS Deneys, A.-S. 2020. Subject omission
Yearbook 2004, A. Kerkhoff, J. de on Facebook: A quantitative study
Lange & O.S. Leicht (eds.) 109–19. of subject omission in finite clauses
Utrecht Institute of Linguistics. in English Facebook posts. MA
———2008. Article omission in headlines thesis. Ghent University.
and child language: A processing Déprez, V. 1992. Raising constructions
approach. PhD diss. Utrecht in Haitian Creole. Natural Language
University. and Linguistic Theory 10: 191–231.
de Lange, J. & Avrutin, S. 2009. ———2000. Parallel (a)symmetries and
Reading between the (head)lines: the internal structure of negative
A processing account of article expressions. Natural Language and
omissions in newspaper headlines Linguistic Theory 18: 253–342.
and child speech. Lingua 119: de Vries, L. 1993. Forms and Functions
1523–40. in Kombai, an Awyu Language
Delfitto, D. & Fiorin, G. 2014. of Irian Jaya. Pacific Linguistics
Exclamatives: Issues of syntax, series B-108, Australian National
logical form and interpretation. University, Canberra.
Lingua 152: 1–20. de Vries, M. 2002. The Syntax of
del Gobbo, F. 2003 Appositives at the Relativization. Utrecht: LOT.
interface. PhD diss. University of ———2006. The syntax of appositive
California, Irvine. relativization: On specifying co-
Delsing, L.-O. 2010. Exclamatives in ordination, false free relatives and
Scandinavian. Studia Linguistica promotion. Linguistic Inquiry 37:
64: 16–36. 229–70.
Demirdache, H. 1991. Resumptive Dickey, M. W. 1996. Constraints
chains in restrictive relatives, on the sentence processor and
appositives, and dislocation the distribution of resumptive
structures. PhD diss. MIT. pronouns. In University of
Demonte, V. & Fernández Soriano, O. Massachusetts Occasional Papers in
2009. Force and finiteness in the Linguistics 19: 157–92.
 References 485

Diercks, M. 2010. Agreement with Doherty, C. 1994. The syntax of subject

subjects in Lubukusu. PhD diss. contact relatives. Chicago Linguistic
Georgetown University. Society Proceedings 29: 55–65.
———2013. Indirect Agree in Lubukusu Dominguez, L. 2014. Mapping Focus:
Complementizer Agreement. The syntax and prosody of focus in
Natural Language and Linguistic Spanish. Boston: Proquest.
Theory 31: 357–407. Donati, C. 2006. On wh‑head
Diercks, M., van Koppen, M., & movement. In Wh‑Movement:
Putnam, M. 2020. Agree probes Moving on, L. Cheng & N. Corver
down: Anaphoric feature valuation (eds.) 21–46. Cambridge, MA: MIT
and phase reference. In Agree to Press.
Agree: Agreement in the minimalist Donati C. & Cecchetto, C. 2011.
program, P. W. Smith, J. Mursell & Relabeling Heads: A unified
K. Hartmann (eds.) 347–89. Berlin: account for relativization
Language Science Press. structures. Linguistic Inquiry 42:
Dikken, M. den 1995. Binding, 519–60.
expletives and levels. Linguistic Doran, D. 2015. The semantics of
Inquiry 26: 347–54. copy raising. MA thesis. McMaster
———2001. Pluringulars, pronouns and University.
quirky agreement. The Linguistic Dowty, D. & Brodie, B. 1984. A
Review 18: 19–41. semantic analysis of floating
———2006. Either float and the quantifiers in transformationless
syntax of co-or-dination. Natural grammar. Proceedings of the
Language and Linguistic Theory 24: West Coast Conference on Formal
689–749. Linguistics 3: 75–90.
———2017. Predication in the syntax of Douglas, J. 2016. The syntactic
hyperraising and copy raising. Acta structures of relativization. PhD
Linguistica Academica 64: 3–43. diss. University of Cambridge.
———2018. Secondary predication and ———2017. Unifying the that-trace and
the distribution of raising to object. anti-that-trace effects. Glossa 2:
Acta Linguistica Academica 65: 1–28.
87–117. Drachman, G. 1974. The syntax of
Di Sciullo, A. M., & Isac, D. 2008. The casual speech. Salzlburger Beiträge
asymmetry of Merge. Biolinguistics zur Linguistik I: 273–95.
2: 260–90. Drubig, H. B. 2003. Toward a typology
Doetjes, 1992. J.S. Rightward of focus and focus constructions.
quantifiers float to the left. The Linguistics 41: 1–50.
Linguistic Review 9: 313–32. Dubinsky, S. 1997. Infinitival
———1997. Quantifiers and selection: relative clauses in English: An
On the distribution of quantifying antisymmetric approach to
expressions in French, Dutch and discontinuous constituency.
English. Utrecht: LOT. Proceedings of the Eastern States
486 References

Conference on Linguistics 96: 82– ———1979. Appositive relatives have no

93. Ithaca, NY: Cornell University. properties. Linguistic Inquiry 10:
Duffield, N. 2015. Where not to put 211–43.
why, and why not? The Journal of ———2004. Unspecified categories as
Konan University Faculty of Letters the key to root constructions.
165: 57–68. In Peripheries: Syntactic Edges
Durrleman, S. 2008. The Syntax of and their Effects, D. Adger, C. de
Jamaican Creole: A Cartographic Cat & G. Tsoulas (eds.) 75–120.
perspective. Amsterdam: Benjamins. Dordrecht: Kluwer.
Eckardt, R. 2007. Inherent focus on Endo, Y. 2007a. Locality and
wh‑phrases. Sinn und Bedeutung Information Structure. A
11: 209–28. Cartographic approach to Japanese.
Eide, K. M. 2016. Introduction. In Amsterdam: Benjamins.
Finiteness Matters: On finiteness- ———2007b. Locality and floating
related phenomena in natural quantifiers. Generative Grammar in
languages, Κ. M. Eide (ed.) 1–44. Geneva 5: 35–50.
Amsterdam: Benjamins. ———2014. Why, what … for, how come
Eilander, S. 2020. Front page and why the hell. Papers from the
information: Article omission International Spring Forum of the
in Italian headlines in light of English Linguistic Society of Japan
information density. MA thesis. 7: 10–16.
Utrecht University. ———2015a. Two ReasonPs: What
Elenbaas, M. 2009. The Synchronic and are(*n’t) you coming to the US for?
Diachronic Syntax of the English In The Cartography of Syntactic
Verb-Particle Combination. Utrecht: Structures, Vol. 10: Beyond
LOT. functional sequence, U. Shlonsky
Elgin, S. H. & Haden, R. 1991. A (ed.) 220–31. Oxford: Oxford
Celebration of Ozark English. University Press.
Huntsville, AR: OCLS Press. ———2015b. Why, what … for, how come
Elliot, D. E. 1974. Toward a grammar and why the hell. Papers from the
of exclamations. Foundations of International Spring Forum of the
Language 11: 231–46. English Linguistic Society of Japan,
Embick, D. 1998. Voice systems and 7: 10–16.
the syntax-morphology interface. ———2017. Inter-speaker variation and
MIT Working Papers in Linguistics subject-drop in how come questions
32: 41–72. in English. In The Syntax-
Emonds, J. 1970. Root and structure- Morphology Interface in Generative
preserving transformations. PhD Grammar, C. Yim (ed.) 99–114.
diss. MIT. Seoul National University.
———1976. A Transformational ———2018 Variation in wh‑expressions
Approach to English Syntax. New asking for a reason. Linguistic
York: Academic Press. Variation 18: 299–314.
 References 487

Engdahl, E. 1985. Parasitic gaps, Erteschik-Shir, N. 1986. Wh‑questions

resumptive pronouns, and subject and focus. Linguistics & Philosophy
extraction. Linguistics 23: 3–44. 9: 117–49.
Epstein, S. D., Kitahara, H. & Seely, ———1992. Resumptive pronouns in
T. D. 2015. From Aspects’ islands. In Island Constraints:
‘daughterless mothers’ (aka delta Theory, Acquisition and Processing,
nodes) to POP’s ‘motherless sets’ H. Goodluck & M. Rochemont (eds.)
(aka non-projection): A selective 89–108. Dordrecht: Kluwer.
history of the evolution of simplest Erteschik-Shir, N., Ibnbari, L., & Taube,
Merge. MIT Working Papers in S. 2013. Missing objects as topic
Linguistics 77: 99–112. drop. Lingua 136: 145–69.
Epstein, S. D., Pires, A. & Seely, T. D. Fabb, N. 1990. The difference between
2005. EPP in T: More controversial English restrictive and non-
subjects. Syntax 8: 65–80. restrictive clauses. Journal of
Epstein, S. D & Seely, T. D. 2006. Linguistics 26: 57–78.
Derivations in Minimalism. Fanselow, G. & Lenertová, D. 2010.
Cambridge: Cambridge University Left peripheral focus: Mismatches
Press. between syntax and information
Erdmann, P. 1980. On the history of structure. Natural Language and
subject contact relatives in English. Linguistic Theory 29: 169–209.
Folia Linguistica Historica 1: 139–70. Farkas, D. 1987. DO pro in Hungarian.
Erlewine, M. Y. 2014. Why the null In Approaches to Hungarian,
complementizer is special in Vol. 2: Theories and analyses, I.
the English that-trace effect. Kenesei (ed.) 191–211. Szeged:
Ms. MIT. [Link] JATE.
lingbuzz/002029 Farrell, P. 1990. Null objects in
———2017a. Why the null Brazilian Portuguese. Natural
complementizer is special in Language and Linguistic Theory 8:
complementiser-trace effects. MIT 325–46.
Working Papers in Linguistics 80: Fassi-Fehri, A. 1980. Some complement
371–80. phenomena in Arabic, the
———2017b. Vietnamese focus particles complementizer phrase hypothesis,
and derivation by phase. Journal of and the Non-Accessibility
East Asian Linguistics 26: 325–49. Condition. Analyse/Théorie
Ernst, T. 1991. On the scope principle. 54–114. Paris: Université de Paris,
Linguistic Inquiry 22: 750–56. Vincennes.
———2002. The Syntax of Adjuncts. Felser, C. & Rupp, L. 2001. Expletives
Cambridge: Cambridge University as arguments: Germanic existential
Press. sentences revisited. Linguistische
———2004. Principles of adverbial Berichte, 187: 289–324.
distribution in the lower clause. Ferguson, C. A. (1982) Simplified
Lingua: 114: 755–77. registers and linguistic theory.
488 References

In Exceptional Language and Fillmore, C. J. 1968. The case for

Linguistics, L. K. Obler & L. Menn case. In Universals in Linguistic
(eds.) 49–66. New York: Academic Theory, E. Bach & R. T. Harms (eds.)
Press. 1–88. New York: Holt Rinehart &
Fernández-Salgueiro, G. 2020. On Winston.
EPP effects and the properties of ———1970. Subjects, speakers and roles.
core functional categories. Glossa: Semantics of Natural Language 21:
A Journal of General Linguistics 251–74.
5(1): 36. ———1986. Pragmatically controlled
Ferreira, F. & Swets, B. 2005. The zero anaphora. Berkeley Linguistics
production and comprehension of Society Proceedings 12: 95–107.
resumptive pronouns in relative Fiorentino, G. 2007. European relative
clause ‘island’ contexts. In Twenty- clauses and the uniqueness of the
First Century Psycholinguistics: relative pronoun type. Rivista di
Four cornerstones, A. Cutler (ed.) Linguistica 19: 263–91.
263–78. Mahway, NJ: Lawrence Fitzpatrick, E., Bachenko, J. & Hindle,
Erlbaum Associates. D. 1986. The status of telegraphic
Ferreira, M. 2004. Hyperraising sublanguages. In Analyzing
and null subjects in Brazilian Language in Restricted Domains:
Portuguese. MIT Working Papers in Sublanguage description and
Linguistics 47: 57–85. processing, R. Grishman & R.
———2009. Null subjects and finite Kittredge (eds.) 39–51. Hillsdale,
control in Brazilian Portuguese. NJ: Erlbaum.
Linguistics Today 142: 17–49. Fitzpatrick, J. 2005a. The whys and
Ferreiro, S. M. 2003. Verbal inflectional how comes of presupposition
morphology in Broca’s aphasia. MA and NPI licensing in questions.
thesis. Universitat Autònoma de Proceedings of the West Coast
Barcelona. Conference on Formal Linguistics
Féry, C. 2007. The prosody of 24: 138–45.
topicalization. In On Information ———2005b. Anybody seen the
Structure: Meaning and form, K. auxiliary? MIT Working Papers in
Schwabe & S. Winkler (eds.) 69–86. Linguistics 50: 71–91.
Amsterdam: Benjamins. ———2006. The syntactic and semantic
Féry, C. & Samek-Lodovici, V. 2006. roots of floating quantification.
Focus projection and prosodic PhD diss. MIT.
prominence in nested foci. Flagg, E. 2002. Interface issues in the
Language 82: 131–50. English imperative. PhD diss. MIT.
Fielding, H. (1996). Bridget Jones’s Fong, S. 2018. Featural definition of
Diary. Oxford: Picador. syntactic positions: Evidence from
Fiengo, R. & Higginbotham, J. 1981. hyper‑raising. Proceedings of the
Opacity in NP. Linguistic Analysis Linguistic Society of America 3(2):
7: 347–73. 1–15.
 References 489

Forsmo, M. G. 2020. The order of Frank, R. & Vijay-Shanker, K. 2001.

adverbs: Comparing Cinque Primitive c‑command. Syntax 4:
and Ernst. Bachelor’s project, 164–204.
Norwegian University of Science Frascarelli, M. 2000. The Syntax-
and Technology. Phonology Interface in Focus and
Fox, D. 2000. Economy and Semantic Topic Constructions in Italian.
Interpretation. Cambridge, MA: MIT Dordrecht: Springer.
Press. Frascarelli, M. & Hinterhölzl, R.
Fox, D. & Grodzinsky, Y. 1998. 2007. Types of topics in German
Children’s passive: A view from the and Italian. In On Information
by‑phrase. Linguistic Inquiry 29: Structure, Meaning and Form,
311–32. S. Winkler & K. Schwabe (eds.)
Fox, D. & Nissenbaum, J. 1999. 86–116. Amsterdam: Benjamins.
Extraposition and scope: A case Frascarelli, M. & Puglielli, A. 2007.
for overt QR. Proceedings of the Focus in the Force-Fin system:
West Coast Conference on Formal Information structure in Cushitic
Linguistics 18: 132–44. Languages. In Focus Strategies:
———2004. Condition A and scope Evidence from African Languages,
reconstruction. Linguistic Inquiry E. Aboh, K. Hartmann & M.
35: 475–85. Zimmermann (eds.) 161–84. Berlin:
Francis, E. J., Lam, C., Zheng, C. C., Mouton de Gruyter.
Hitz, J. & Matthews, S. 2015. Fraser, J. B. 1965. An examination of
Resumptive pronouns, structural the verb-particle construction in
complexity, and the elusive English. PhD diss. MIT.
distinction between grammar Frazier, L. & Clifton, C. 2015. Without
and performance: Evidence from his shirt off he saved the child
Cantonese. Lingua 162: 56–81. from almost drowning: Interpreting
Francis, N. 2017. Modal scope in an uncertain input. Language,
Negative Inversion constructions. Cognition & Neuroscience 30(6):
Proceedings of the West Coast 635–47.
Conference on Formal Linguistics Freidin, R. & Vergnaud, J. R. 2001.
34: 214–21. Exquisite connections: Some
Franco, I. 2009. Verbs, subjects and remarks on the evolution of
stylistic fronting: A comparative linguistic theory. Lingua 111:
analysis of the interaction of CP 639–66.
properties with verb movement and Frey, W. 2004. Notes on the syntax
subject positions in Icelandic and and pragmatics of German left
Old Italian. PhD diss. University of dislocation. In The Syntax and
Siena. Semantics of the Left Periphery,
Frank, R. 2006. Phase theory and tree- H. Lohnstein & S. Trissler (eds.)
adjoining grammar. Lingua 116: 163–209. Berlin: Mouton de
145–202. Gruyter.
490 References

———2005. Pragmatic properties of Corver & J. Nunes (eds.) 291–326.

certain German and English Amsterdam: Benjamins.
left peripheral constructions. Fukuda, S. & Polinsky, M. 2014.
Linguistics 43: 89–129. Licensing of floating nominal
Friedemann, N. 1998. Functional modifiers and unaccusativity
categories in agrammatic in Japanese. Proceedings of the
production. PhD diss. Tel-Aviv West Coast Conference on Formal
University. Linguistics 31: 189–98.
Friedmann, N., Belletti, A. & Rizzi, L. Fukui, N. 1986. A theory of category
2009. Relativized minimality: Types projection and its application. PhD
of intervention in the acquisition diss. MIT.
of A‑bar dependencies. Lingua 119: ———2011. Merge and Bare Phrase
67–88. Structure. In The Oxford Handbook
Friedmann, N. & Grodzinsky, Y. of Linguistic Minimalism, C.
1997. Tense and agreement in Boeckx (ed.) 73–95. Oxford: Oxford
agrammatic production: Pruning University Press.
the syntactic tree. Brain and Fukui, N. & Speas, M. 1986. Specifiers
Language 56: 397–425. and projection. MIT Working
———2000. Split inflection in Papers in Linguistics 8: 128–72.
neurolinguistics. In The Acquisition Fukushima, K. 1991a. Phrase structure
of Syntax: Studies in comparative grammar, Montague semantics, and
developmental linguistics, M.-A. floating quantifiers in Japanese.
Friedmann & L. Rizzi (eds.) 84–101. Linguistics and Philosophy 14:
London: Longman. 581–628.
Friedmann, N., Novogrodsky, R., ———1991b. Generalized floating
Szterman, R. & Preminger, O. 2009. quantifiers. PhD diss. University of
Resumptive pronouns as a last Arizona.
resort when movement is impaired: Funakoshi, K. 2017. Disjunction and
Relative clauses in hearing Object Drop in Japanese. Florida
impairment. In Current Issues in Linguistics Papers 4: 11–20.
Generative Hebrew Linguistics 7: Fuß, E. 2004. Diachronic clues to
267–90. pro-drop and Complementiser
Fujii, T. 2005. Cycle, linearization Agreement in Bavarian. In
of chains, and multiple case Diachronic Clues to Synchronic
checking. In Proceedings of Grammar, E. Fuß & C. Trips (eds.)
Console XIII, S. Blaho, L. Vicente 59–100. Amsterdam: Benjamins.
& E. Schoorlemmer (eds.) 39–65. ———2005. The Rise of Agreement: A
University of Leiden: Student formal approach to the syntax
Organization of Linguistics in and grammaticalization of verbal
Europe. inflection. Amsterdam: Benjamins.
———2007. Cyclic chain reduction. In ———2008. Multiple agreement and the
The Copy Theory of Movement, N. representation of inflection in the
 References 491

c‑domain. Linguistische Berichte Gérard, J. 1980. L’Exclamation en

213: 77–106. Français. Tübingen: Niemeyer.
———2014. Complementizer Agreement Gerken, L. 1991. The metrical basis for
(in Bavarian): Feature inheritance children’s subjectless sentences.
or feature insertion? In Bavarian Journal of Memory and Language
Syntax: Contributions to the Theory 30: 431–51.
of Syntax, G. Grewendorf & H. ———1994. Young children’s
Weiß (eds.) 51–82. Amsterdam: representation of prosodic
Benjamins. phonology: Evidence from
Futagi, Y. 2004. Japanese focus particles English-speakers’ weak syllable
at the syntax-semantics interface. productions. Journal of Memory
PhD diss. Rutgers University. and Language 33: 19–38.
García Velasco, D. & Portero Muñoz, Ghosh, P. 2022. 12 news writing
C. 2002. Understood objects in rules used by subeditors
Functional Grammar. In Working for compelling headlines.
Papers in Functional Grammar 76. [Link].
University of Amsterdam. Ginzburg, J. & Sag, I. 2000.
Garzonio, J. 2005. Struttura Interrogative Investigations: The
informazionale e soggetti nulli in form, meaning, and use of English
russo: Un approccio cartografico. interrogatives. Stanford, CA: CSLI
PhD diss. University of Padua. Publications.
Gasper, S. 2014. An editing framework Giorgi, A. & Pianesi, T. F. 1997. Tense
guiding students towards self- and Aspect: From semantics to
correcting errors in writing. MA morphosyntax. Oxford: Oxford
thesis. Rowan University. University Press.
Gavarrò, A. 1991. A note on Catalan Gisbourne, N. 1996. English perception
clitics. Catalan Working Papers in verbs. PhD diss. University College
Linguistics 1: 65–73. London.
Geisler, C. 1998. Infinitival relative ———2010. The Event Structure of
clauses in spoken discourse. Perception Verbs. Oxford: Oxford
Language Variation and Change 10: University Press.
23–41. Giusti, G. 1990. Floating quantifiers,
Gelderen, E. van 2009. Renewal in the scrambling and configurationality.
left periphery: Economy and the Linguistic Inquiry 21: 633–41.
complementiser layer. Transactions ———1991a. The categorial status of
of the Philological Society 107: quantified nominals. Linguistische
131–95. Berichte 136: 438–452.
Geluykens, R. (1992) From Discourse ———1991b. The syntax of floating alles
Process to Grammatical in German. In Issues in German
Construction: On left dislocation Syntax, W. Abraham, W. Kosmeijer
in English. Amsterdam: & E. Reuland (eds.) 327–50. The
Benjamins. Hague: Mouton.
492 References

———1996. Is there a FocusP and Grant, H. (2015). Tumblinguistics:

a TopicP in the Noun Phrase? Innovation and variation in new
University of Venice Working forms of written CMC. MA thesis.
Papers in Linguistics 6: 105–28. University of Glasgow.
———1997. The categorial status Greco, C., Haegeman, L. & Phan, T.
of determiners. In The New 2017. Expletives and speaker-
Comparative Syntax, L. Haegeman related meaning. In Order and
(ed.) 94–113. Cambridge: Structure in Syntax II: Subjecthood
Cambridge University Press. and argument structure, M. Sheehan
———2012. On force and case, fin and & L. R. Bailey (eds.) 69–93. Berlin:
num. In Enjoy Linguistics: Papers Language Science Press.
Presented to Luigi Rizzi on the Green, L. J. 2002. African American
Occasion of his 60th Birthday, V. English: A linguistic introduction.
Bianchi & C. Chesi (eds.) 205–17. Cambridge: Cambridge University
Siena: CISCL Press. Press.
Goedegebuure, P. 2009. Focus structure Green, M. 2007. Focus in Hausa.
and Q‑word questions in Hittite. Oxford: Blackwell.
Linguistics 47: 945–69. Green, M. & Jaggar, P. 2003. Ex-
Goldshlag, E. 2005. The EPP, situ and in-situ focus in Hausa:
nominative case and expletives. Syntax, semantics and discourse. In
MA thesis. Tel Aviv University. Research in Afroasiatic Grammar
Gonçalves, A., Santos, A. L. & Duarte, II, J. Lecarme (ed.) 187–213.
I. 2014. (Pseudo-) inflected Amsterdam: Benjamins.
infinitives and control as Agree. Greenberg, G. 1984. Left dislocation,
In Selected Papers from Going topicalization, and interjections.
Romance, Leuven 2012, K. Natural Language and Linguistic
Lahousse & S. Marzo (eds.) 161–80. Theory 2: 283–87.
Amsterdam: Benjamins. Gregory, M. & Michaelis, L. 2001.
Goodall, G. 1997. Theta-alignment Topicalization and left dislocation:
and the by‑phrase. Proceedings of A functional opposition revisited.
the Chicago Linguistics Society 33: Journal of Pragmatics 33: 1665–
129–39. 1706.
———2001. The EPP in Spanish. In Grewendorf, G. 2002. Left dislocation
Objects and Other Subjects: as movement. Georgetown
Grammatical functions, functional University Working Papers in
categories, and configurationality, Theoretical Linguistics 2: 31–81.
W. D. Davies & S. Dubinsky (eds.) Grewendorf, G. & Poletto, C. 2009. The
193–223. Dordrecht: Kluwer. hybrid complementizer system of
———2017. Finiteness contrasts without Cimbrian. Studies in Linguistics 3:
tense? A view from Mandarin 181–94.
Chinese. Journal of East Asian Grice, H. P. 1975. Logic and
Linguistics 26: 259–99. conversation. In Syntax and
 References 493

Semantics 3: Speech Acts, P. Cole (eds.) 249–88. Cambridge, MA: MIT

& J. Morgan (eds.) 41–58. New Press.
York: Academic Press. Grohmann, K. K., Drury J. & Castillo, J.
Griffin, W. E. 2003. The split-INFL C. 2000. No more EPP. Proceedings
hypothesis and AgrsP in Universal of the West Coast Conference on
Grammar. In The Role of Agreement Formal Linguistics 19: 153–66.
in Natural Languages: TLS 5 Grohmann, K. K. & Haegeman, L. 2002.
Proceedings, W. E. Griffin (ed.) Resuming reflexives. [Link]
25–36. Munich, Lincom Europa. .[Link]/lingbuzz/000140
Grimshaw, J. 1979. Complement Grolla, E. 2005. Resumptive pronouns
selection and the lexicon. Linguistic as last resort: Implications for
Inquiry 10: 279–326. language acquisition. Penn Working
———1993. Minimal projection, heads, Papers in Linguistics 11: 71–84.
and optimality. Ms. Rutgers Groos, A. & van Riemsdijk, H. 1981.
University. Matching effects in free relatives:
Groat, E. 1995. English expletives: A A parameter of core grammar.
minimalist approach. Linguistic In Theory of Markedness in
Inquiry 26: 354–65. Generative Grammar, A. Belletti,
———1997. A derivational program for L. Brandi & L. Rizzi (eds.)
syntactic theory. PhD diss. Harvard 171–216. Pisa: Scuola Normale
University. Superiore.
———1999. Raising the case Grosu, A. 1996. The proper analysis
of expletives. In Working of missing-P free relative
Minimalism, S. D. Epstein & N. constructions. Linguistic Inquiry
Hornstein (eds.) 27–43. Cambridge, 27: 257–93.
MA: MIT Press. ———2003. A unified theory of standard
Groefsema, M. 1995. Understood and transparent free relatives.
arguments: A semantic/pragmatic Natural Language and Linguistic
approach. Lingua 96: 139–61. Theory 21: 247–331.
Grohmann, K. K. 1997. On left Grosu, A. & Horvath, J. 1984. The GB
dislocation. Groninger Arbeiten theory and Raising in Rumanian.
zur Germanistischen Linguistik 40: Linguistic Inquiry 15: 348–53.
1–33. Grosu, A. & Landman, F. 1998. Strange
———2000. Prolific peripheries: A relatives of the third kind. Natural
radical view from the left. PhD diss. Language Semantics 6: 125–70.
University of Maryland. Gruber, B. 2008. Complementizer
———2003. Prolific Domains. Agreement: New evidence from
Amsterdam: Benjamins. the Upper Austrian variant of
———2006. Top issues in questions: Gmunden. MA thesis. University of
Topics – Topicalization – Vienna.
Topicalizability. In Wh‑Movement: Gruber, J. S. 1965. Studies in lexical
Moving On, L. Cheng & N. Corver relations. PhD diss. MIT.
494 References

———1967. Topicalization in child Guy, G. R. & Bayley, R. 1995. On

language. Foundations of Language the choice of relative pronouns
3: 37–65. in English. American Speech 70:
———1976. Lexical Structures in Syntax 148–62.
and Semantics. Amsterdam: North Guz, W. 2017. Resumptive pronouns in
Holland. Polish co relative clauses. Journal
Guasti, M. T. & Rizzi, L. 2002. of Slavic Linguistics 25: 95–130.
Agreement and tense as distinct Gwynne, N. M. 2013. Gwynne’s
syntactic positions: Evidence from Grammar: The ultimate
acquisition. In Functional Structure introduction to grammar and the
in DP and IP, G. Cinque (ed.) writing of good English. London:
167–94. Oxford: Oxford University Ebury press.
Press. Ha, S. 2010. A prosody analysis of
Guilfoyle, E., Hung, H. & Travis, L. the comp‑trace effect. English
1992. Spec of IP and spec of VP: Language and Linguistics 16:
Two subjects in Austronesian 109–33.
languages. Natural Language and Haaften, T. van, Smits, R. & Vat,
Linguistic Theory 10: 375–414. J. 1983. Left dislocation,
Gundel, J. K. 1975. Left dislocation and connectedness and reconstruction.
the role of topic-comment structure Linguistics Today 3: 133–54.
in linguistic theory. Ohio State Hackl, M. & Nissenbaum, J. 2012. A
Working Papers in Linguistics 18: modal ambiguity in for-infinitival
72–131. relative clauses. Natural Language
———1985. Shared knowledge and Semantics 20: 59–81.
topicality. Journal of Pragmatics 9: Haeberli, E. 2003. Categorial
83–107. features as the source of EPP
———1988. Universals of topic- and abstract case phenomena. In
comment structure. In Studies in New Perspectives on Case Theory,
Syntactic Typology, M. Hammond, E. Brandner & H. Zinsmeister
E. A. Moravcsik & J. Werth (eds.) (eds.) 89–126. Stanford, CA: CSLI
209–42. Amsterdam: Benjamins. publications.
Gutiérrez-Rexach, J. 2001. Spanish Haegeman, L. 1985. The get passive
exclamatives and the interpretation and Burzio’s generalization. Lingua
of the left periphery. In Romance 66: 53–77.
Languages and Linguistic Theory ———1987a. Complement ellipsis in
1999, Y. D’hulst, J. Rooryck, English: Or, how to cook without
& J. Schroten (eds.) 167–94. objects. In Studies in Honour
Amsterdam: Benjamins. of René Derolez, A. M. Simon-
———2008. Spanish root exclamatives Vandenbergen (ed.) 248–61. Ghent:
at the syntax/semantics interface. University of Ghent.
Catalan Journal of Linguistics 7: ———1987b. Register variation
117–33. in English: Some theoretical
 References 495

observations. Journal of English ———2002b. Non-overt subject

Linguistics 20: 230–48. pronouns in written English. In
———1990a. Non-overt subjects in diary Language, Context and Cognition:
contexts. In Grammar in Progress, Papers in honour of Wolf Dietrich
J. Mascarò & M. Nespor (eds.) Bald’s 60th birthday, S. Sybil,
167–74. Dordrecht: Foris. M. Klages, E. Hantson, & U.
———1990b. Understood subjects in Römer (eds.) 135–49. Munich:
English diaries. Multilingua 9: Langenscheidt-Longman.
157–99. ———2003. Notes on long adverbial
———1992. Theory and Description in fronting in English and the left
Generative Syntax. Cambridge: periphery. Linguistic Inquiry 34:
Cambridge University Press. 640–49.
———1994. Introduction to Government ———2004. Topicalization, CLLD and
and Binding Theory, 2nd edition. the left periphery. In Proceedings of
Oxford: Blackwell. the Dislocated Elements Workshop,
———1995. The Syntax of Negation. B. Shaer, W. Frey & C. Maienborn
Cambridge: Cambridge University (eds.) 157–92. Berlin: ZAS.
Press. ———2006a. Argument fronting in
———1997. Register variation, English, Romance CLLD and the
truncation, and subject omission left periphery. In Crosslinguistic
in English and in French. English Research in Syntax and Semantics:
Language and Linguistics 1: Negation, tense, and clausal
233–70. architecture, R. Zanuttini, H.
———2000a. Adult null subjects in Campos, E. Herburger & P. H.
non pro-drop languages. In Portner (eds.) 27–52. Washington,
The Acquisition of Syntax, M.- DC: Georgetown University Press.
A. Friedemann & L. Rizzi (eds.) ———2006b. Conditionals, factives and
129–69. London: Addison, Wesley the left periphery. Lingua 116:
and Longman. 1651–69.
———2000b. Inversion, non-adjacent ———2006c. Register variation: Core
inversion and adjuncts in CP. grammar and periphery. In
Transactions of the Philological Encyclopedia of Language and
Society 98: 121–60. Linguistics: 2nd edition, K. Brown
———2000c. Negative preposing, (ed.) 468–74. Elsevier Science.
Negative Inversion and the split CP. ———2007a. Operator movement and
In Negation and Polarity: Syntactic topicalization in adverbial clauses.
and semantic perspectives, L. Horn Folia Linguistica 18: 485–502.
& Y. Kato (eds.), 21–62. Oxford: ———2007b. Subject omission in
Oxford University Press. present-day written English: On the
———2002a. Sentence-medial NP- theoretical relevance of peripheral
adjuncts in English. Nordic Journal data. Rivista di Grammatica
of Linguistics 25: 79–108. Generativa 32: 91–124.
496 References

———2009. The movement analysis of case study in West Flemish. In

temporal adverbial clauses. English On Peripheries: Exploring clause
Language and Linguistics 13: initial and clause final positions, C.
385–408. Cardinaletti, G. Cinque & Y. Endo
———2010. The internal syntax of (eds.) 209–36. Tokyo: Hituzi Syobe
adverbial clauses. Lingua 120: Publishing.
628–48. Haegeman, L. & Ihsane, T. 1999.
———2012. Adverbial Clauses, Subject ellipsis in embedded
Main Clause Phenomena, and clauses in English. Journal of
Composition of the Left Periphery. English Language and Linguistics 3:
Oxford: Oxford University Press. 117–45.
———2013. The syntax of registers: ———2001 Adult null subjects in the
Diary subject omission and the non-pro drop languages: Two diary
privilege of the root. Lingua 130: dialects. Language Acquisition 9:
88–110. 329–46.
———2017. Unspeakable sentences: Haegeman, L., Jiménez-Fernández,
Subject omission in written Á. & Radford, A. 2014.
registers – A Cartographic analysis. Deconstructing the Subject
Linguistic Variation 17: 229–50. Condition in terms of cumulative
———2019. Register-based subject constraint violation. The Linguistic
omission in English and its Review 31: 73–150.
implication for the syntax of Haegeman, L. & van Koppen, M. 2012.
adjuncts. Anglophonia 28. https:// Complementizer agreement and the
[Link]/10.4000/anglophonia.2873. relation between T and C. Linguistic
———2021. The decomposition of Inquiry 43: 441–54.
subjects and the role of SubjP. Ms. Haegeman, L., & Starke, E. 2021.
Ghent University. Register-specific subject omission
Haegeman, L. & Danckaert, L. 2017. in English and French and
Variation in English subject the syntax of coordination. In
extraction: The case of hyperactive Continuity and Variation in
subjects. In Studies on Syntactic Germanic and Romance, S. Wolfe
Cartography, F. Si (ed.) 302–35. & C. Meklenborg (eds.) 15–43.
Beijing: China Social Sciences Oxford: Oxford University Press.
Press. Haegeman, L., Weir, A, Danckaert, L.,
Haegeman, L. & Greco, C. 2017. Initial D’Hulster, T. & Buelens, L. 2015.
adverbial clauses and West Flemish Against the root analysis of subject
V3. Ms. University of Ghent. contact relatives in English. Lingua
Haegeman, L. & Guéron, J. 1999. 163: 61–74.
English Grammar: A Generative Haider, H. 2000. Adverb placement:
perspective. Oxford: Blackwell. Convergence of structure and
Haegeman, L. & Hill, V. 2014. Vocatives licensing. Theoretical Linguistics
and speech-act projections: A 26: 95–134.
 References 497

———2004. Pre- and post-verbal and force in Universal Grammar.

adverbials in OV and VO. Lingua London: Psychology Press.
114: 779–807. ———2001. Force, negation and
———2005. How to turn German into imperatives. The Linguistic Review
Icelandic – and derive the OV-VO 18: 289–325.
contrasts. Journal of Comparative Han, C., Elouazizi, N., Galeano,
Germanic Linguistics 8: 1–53. C., Görgülü, E., Hedberg, N.,
———2017. In the absence of a subject. Jeffrey, M., Kim, K. & Kirby, S.
Wiener Linguistische Gazette 82: 2012. Processing strategies and
87–98. resumptive pronouns in English.
———2019. On absent, expletive Proceedings of the West Coast
and non-referential subjects. In Conference on Formal Linguistics
Semantic and Syntactic Aspects of 30: 153–61.
Impersonality, P. Herbeck, B. Pöll, Han, H.-S. 2004. There as an existential
& A. Wolfsgruber (eds.) 11–46. operator. Language Research 40:
Hamburg: Helmut Buske Verlag. 451–64.
Haig, J. 1980. Some observations on Hankamer, J. 1971. Constraints on
quantifier floating in Japanese. deletion in syntax. PhD diss. Yale
Linguistics 18: 1065–83. University.
Hallman, P. 1997. Reiterative syntax. Hankamer, J. & Sag, I. 1976. Deep
In Clitics, Pronouns and Movement, and surface anaphora. Linguistic
J. R. Black & V. Motapayane (eds.) Inquiry 7: 391–428.
87–131. Amsterdam: Benjamins. Hanson, R. 1999. Split INFL and
Hamamo, S. 1997. On Japanese the acquisition of Neg and Aux.
quantifier floating. In Directions Calgary Working Papers in
in Functional Linguistics, A. Linguistics 21: 2–13.
Kamio (ed.) 173–97. Amsterdam: Hardt, D. 1993. Verb phrase ellipsis:
Benjamins. Form, meaning and processing.
Han, C.-H. 1998. The structure and PhD diss. University of
interpretation of imperatives: Pennsylvania.
Mood and force in Universal Harford Perez, C. 1985. Aspects of
Grammar. PhD diss. University of complementation in three Bantu
Pennsylvania. languages. PhD diss. University of
———2000a. The evolution of do- Wisconsin–Madison.
support in English imperatives. Harris, M. & Vincent, N. 1980. On zero
In Diachronic Syntax: Models relatives. Linguistic Inquiry 11:
and mechanisms, S. Pintzuk, G. 805–807.
Tsoulas & A. Warner (eds.) 275– Hartmann, J. M. 2008. Expletives in
95. Oxford: Oxford University existentials: English there and
Press. German da. Utrecht: LOT.
———2000b. The Structure and Hartman, J. & Ai, R. R. 2009. A focus
Interpretation of Imperatives: Mood account of Swiping. In Selected
498 References

Papers from the 2006 Cyprus Evidence from the definiteness of

Syntaxfest, K. K. Grohmann & variables. In The Representation of
P. Panagiotidis (eds.) 92–122. (In)definiteness, E. J. Reuland & A.
Newcastle: Cambridge Scholars ter Meulen (eds.) 21–42. Cambridge,
Publishing. MA: MIT Press.
Harvie, D. 1998. Null subject in Heim, I. & Kratzer, A. 1998. Semantics
English: Wonder if it exists. Cahiers in Generative Grammar. Oxford:
Linguistiques d’Ottowa 26: 15–25. Blackwell.
Harwood, W. 2012. There are several Hellan, L. 1986. The headedness of
positions available: English NPs in Norwegian. In Features and
intermediate subject positions. Projections, P. Muysken & H. van
Proceedings of ConSOLE 19: Riemsdijk (eds.) 89–122. Dordrecht:
215–39. Foris.
Hasegawa, N. 1993. Floating Henderson, B. 2006. Multiple
quantifiers and bare NP agreement and inversion in Bantu.
expressions. In Japanese Syntax Syntax 9: 275–89.
in Comparative Grammar, N. Hendrick, R. 1982. Reduced questions
Hasegawa (ed.) 115–45. Tokyo: and their theoretical implications.
Kuroshio. Language 58: 800–19.
Hawkins, J. A. 2001. Why are Henry, A. 1992. Infinitives in a for-to
categories adjacent? Journal of dialect. Natural Language and
Linguistics 37: 1–34. Linguistic Theory 10: 279–301.
Hazout, I. 2004a. Long-distance ———1995 Belfast English and Standard
agreement and the syntax of for-to English: Dialect variation and
infinitives. Linguistic Inquiry 3: parameter-setting. Oxford: Oxford
338–43. University Press.
———2004b. The syntax of existential ———2012. Phase edges, quantifier
constructions. Linguistic Inquiry float and the nature of (micro-)
35: 393–430. variation. Iberia 4: 23–39.
Heejeong Ko, E. O. 2010. A hybrid Henry, A. & Cottell, S. 2007. A new
approach to floating quantifiers: approach to transitive expletives:
Experimental evidence. Linguistic Evidence from Belfast English.
Research 29: 69–106. English Language and Linguistics
Heffer, S. 2011. Strictly English: The 11: 279–99.
correct way to write … and why it Herdan, S. 2008. Degrees and Amounts
matters. London: Windmill Books. in Relative Clauses. Proquest:
Heestand, D., Xiang, M. & Polinsky, Google Books.
M. 2011 Resumption still does not Herrmann, T. 2003. Relative
rescue islands. Linguistic Inquiry clauses in dialects of English:
42: 138–52. A typological approach. PhD
Heim, I. 1987. Where does the diss. Albert-Ludwigs-Universität,
definiteness restriction apply? Freiburg.
 References 499

———2005. Relative clauses in English Ruwet (eds.) vol. 1, 9–17, Tübingen:

dialects of the British Isles. In A Max Niemeyer Verlag.
Comparative Grammar of British ———1975. On the source of lefthand
English Dialects, B. Kortmann, T. NPs in French. Linguistic Inquiry 6:
Herrmann, L. Pietsch & S. Wagner 155–65.
(eds.) 21–124. Berlin: Mouton. ———1976. Two analyses of free relatives
Heycock, C. 1994. Layers of in French. Proceedings of the North
Predication. New York: Garland. East Linguistic Society 6: 137–52.
———2005. On the interaction of Hirschbühler, P. & Rivero, M. L. 1983.
adjectival modifiers and relative Remarks on free relatives and
clauses. Natural Language matching phenomena. Linguistic
Semantics 13: 359–82. Inquiry 14: 505–19.
Higgins, F. R. 1973. On J. Emonds’ Hladnik, M. 2015. Mind the gap:
analysis of Extraposition. In J. Resumption in Slavic relative
Kimball (ed.) Syntax and Semantics clauses. Utrecht: LOT.
2: 149–95. New York: Academic Hoeksema, J. 1996. Floating
Press. quantifiers, partitives and
Hill, J. A. C. 1983. A computational distributivity. In Partitives: Studies
model of language acquisition on the syntax and semantics of
in the two-year-old. Indiana partitive and related constructions,
University Linguistics Club, J. Hoeksema (ed.) 57–106. Berlin:
Bloomington, IN. Mouton de Gruyter.
Hinrichs, L., Smzrecsanyl, B. & Hoekstra, E. & Smits, C. 1999.
Bohmann, A. 2015. Which- Everything you always wanted
hunting and the Standard English to know about Complementizer
relative clause. Language 91: Agreement. In Proceedings of
806–36. WECOL 10, E. van Gelderen & V.
Hiraira, K. 2001. Multiple agree and the Samiian (eds.) 189–200.
Defective Intervention Constraint. Hoffmann, T. 2011. Preposition
MIT Working Papers in Linguistics Placement in English. Cambridge:
40: 67–80. Cambridge University Press.
———2005. Dimensions of symmetry Hofmeister, P. & Norcliffe, E. 2013.
in syntax: Agreement and clausal Does resumption facilitate sentence
architecture. PhD diss. MIT. comprehension? In The Core
Hirschbühler, P. 1973. La dislocation à and the Periphery: Data-driven
gauche en français. Le Langage et perspectives on syntax inspired by
l’Homme 23: 19–125. Ivan A. Sag, P. Hofmeister & E.
———1974. La dislocation à gauche Norcliffe (eds.) 225–46. Stanford,
comme construction basique en CA: CSLI Publications.
français. In Actes du Colloque Hoji, H. 1998. Null object and sloppy
Franco-Allemande de Grammaire identity in Japanese. Linguistic
Transformationelle, C. Rohrer & N. Inquiry 29: 127–52.
500 References

Holmberg, A. 2000. Scandinavian Movement, N. Corver & J. Nunes

stylistic fronting: How any (eds.) 351–85. Amsterdam:
category can become an expletive. Benjamins.
Linguistic Inquiry 31: 445–83. ———2009. A Theory of Syntax:
Holmer, A. 2013. Evidence from Minimal operations and Universal
Formosan for a unified theory Grammar. Cambridge: Cambridge
of adverb ordering. Lingua 122: University Press.
902–21. Hornstein, N. & Nunes, J. 2008.
Honda, M. 2011. Wh‑exclamatives, Adjunction, labeling, and bare
factivity, and topicalization phrase structure. Biolinguistics 2:
revisited. Studies in Language 57–86.
Sciences 17: 87–117. Horsey, R. 1998. Null arguments in
Hooper, J. B. & Thompson, S. A. English registers: A minimalist
1973. On the applicability of root account. BA thesis. La Trobe
transformations. Linguistic Inquiry University.
4: 465–97. Horvath, J. 1986. Focus in the Theory
Hopper, R. 1992. Telephone of Grammar and the Syntax of
Conversation. Bloomington, IN: Hungarian. Dordrecht: Foris.
Indiana University Press. ———1995. Structural focus, structural
Horch, E. & Reich, I. 2016. On ‘Article case and the notion of feature
Omission’ in German and the assignment. In Discourse-
‘Uniform Information Density Configurational Languages, K. É.
Hypothesis’. In Proceedings of Kiss (ed.) 28–64. Oxford: Oxford
the 13th Conference on Natural University Press.
Language Processing, S. Dipper, F. ———2000. Interfaces vs. the
Neubarth & H. Zinsmeister (eds.) computational system in the syntax
125–27. Bochum. of focus. In Interface Strategies, H.
Horn, G. 1981. A pragmatic approach Bennis, M. Everaert & E. Reuland
to certain ambiguities. Linguistics (eds.) 183–206. Amsterdam: KNAW
and Philosophy 4: 321–58. publications.
Hornstein, N. 1995. Logical Form: ———2007. Separating ‘focus movement’
From GB to minimalism. Oxford: from focus. In Clausal and Phrasal
Blackwell. Architecture: Syntactic derivation
———1999. Movement and Control. and interpretation, S. Karimi, V.
Linguistic Inquiry 30: 69–96. Samiian & W. K. Wilkins (eds.)
———2001. Move: A minimalist theory 108–45. Amsterdam: Benjamins.
of construal. Oxford: Blackwell. Householder, F. W. 1987. Some facts
———2003. On control. In Minimalist about me and I. Language Research
Syntax, R. Hendrick (ed.) 6–81. 23: 163–84.
Oxford: Blackwell. Huang, C.-T. J. 1982. Logical relations
———2007. Pronouns in a minimalist in Chinese and the theory of
setting. In The Copy Theory of grammar. PhD diss. MIT.
 References 501

———1984. On the distribution and Ihsane, T. 1998. The syntax of diaries:

reference of empty pronouns. Grammar and register variation.
Linguistic Inquiry 15: 531–74. Mémoire de licence, University of
———1991. Remarks on the status Geneva, Department of English.
of the null object. In Principles Ikarashi, K. 2015. A functional
and Parameters in Comparative approach to English constructions
Grammar, R. Freidin (ed.) 56–76. related to evidentiality. PhD diss.
Cambridge, MA: MIT Press. University of Tsukuba.
———1993. Reconstruction and the Inada, S. 2007. Towards a syntax
structure of VP: Some theoretical of two types of relative clauses.
consequences. Linguistic Inquiry Linguistic Research 23: 1–41.
24: 103–38. Ingham, R. 1998. Tense without
Huang, N. & Mendes, G. 2019. On agreement in early clause structure.
pronominalisation and clausal Language Acquisition 7: 51–81.
idioms. Ms. University of Maryland. Inoue, I. 1991. On the genesis of there-
Huddleston, R. 1993. On exclamatory- constructions. English Linguistics
inversion sentences. Lingua 90: 8: 34–51.
259–69. Irani, A. 2014. Focusing in Hindi
Huddleston, R. & Pullum, G. K. 2002. Syntax. MA thesis. Georgetown
The Cambridge Grammar of the University.
English Language. Cambridge: Isac, D. 2006. In defense of a
Cambridge University Press. quantificational account of definite
Hulsey, S. & Sauerland, U. 2006. DPs. Linguistic Inquiry 37: 275–88.
Sorting out relative clauses. Natural ———2015. The Morphosyntax of
Language Semantics 14: 111–37. Imperatives. Oxford: Oxford
Iatridou, S. 1993. On nominative case University Press.
assignment and a few related Ishii, Y. 1999. A note on floating
things. MIT Working Papers in quantifiers in Japanese. In
Linguistics 19: 175–96. Linguistics: In search of the human
Ihalainen, O. 1980. Relative clauses mind, M. Muraki & E. Iwamoto
in the dialect of Somerset. (eds.) 236–67. Tokyo: Kaitakusha.
Neuphilologische Mitteilungen 81: Ito, T. 1993. Object ellipsis in
187–96. subjectless nonfinite clauses in
———1991. The grammatical subject in English. English Linguistics 10:
educated and dialectal English: 75–94.
Comparing the London-Lund Ito, T. & Kashihara, Y. 2010. Studies on
corpus and the Helsinki corpus of English literature and linguistics
modern English dialects. In English with the aid of personal computers.
Computer Corpora: Selected papers Kawasaki Journal of Medical
and research guide, S. Johansson Welfare 16: 19–33.
& A.-B. Stenstrøm (eds.) 201–14. Izvorski, R. 2001. Free adjunct free
Berlin: Mouton de Gruyter. relatives. Proceedings of the West
502 References

Coast Conference on Formal Jayaseelan, K. A. 1996. Question-word

Linguistics 19: 232–45. movement to focus and scrambling
Jackendoff, R. S. 1972. Semantic in Malayam. Linguistic Analysis
Interpretation in Generative 26: 63–83.
Grammar. Cambridge, MA: MIT ———2003. Question words in focus
Press. positions. Linguistic Variation
———1974. Introduction to the X-bar Yearbook 3: 69–99.
convention. Indiana University ———2008. Topic, focus and adverb
Linguistics Club. positions in clause structure.
———1977a. X-bar Syntax: A study of Nanzan Linguistics 4: 43–68.
phrase structure. Cambridge, MA: Jenkins, L. 1975. The English
MIT Press. Existential. Tübingen: Narr.
———1977b. Constraints on phrase Jenks, P. 2013. Quantifier float,
structure rules. In Formal Syntax, focus and scope in Thai. Berkeley
P. W. Culicover, T. Wasow & A. Linguistic Society Proceedings 39:
Akmajian (eds.) 249–83. New York: 90–107.
Academic Press. Jensen, B. 2003. Syntax and semantics
Jackendoff, R. S. & Culicover, P. W. of imperative subjects. Nordlyd 31:
2003. The semantic basis of control 150–64.
in English. Language 79: 517–56. ———2007. In favour of a truncated
Jacobs, R. A & Rosenbaum, P. S. imperative clause structure:
1968. English Transformational Evidence from adverbs. Working
Grammar. Waltham, MA: Blaisdell Papers in Scandinavian Syntax 80:
Publishing Co. 163–85.
Jacobsson, B. 1994. Nonrestrictive Jiménez-Fernández, Á. 2011. On
relative that-clauses revisited. the order of multiple topics and
Studia Neophilologica 66: 181–95. discourse-feature inheritance.
———2007. A new look at Negative Dilbilim Araştırmaları 2011: 5–32.
and Correlative Subject–Auxiliary ———2015. Towards a typology of focus:
Inversion in English. Studia Subject position and microvariation
Neophilologica 79: 35–44. at the discourse-syntax interface.
Jaeggli, O. 1982. Topics in Romance Ampersand 2: 49–60.
Syntax. Dordrecht: Foris. ———2018. Negative Preposing:
———1986. Passive. Linguistic Inquiry Intervention and parametric
17: 587–622. variation in complement clauses.
Janda, R. J. 1985. Note-taking English Atlantis 40.1.01.
as a simplified register. Discourse Johnson, K. 1991. Object positions.
Processes 8: 437–54. Natural Language and Linguistic
Janke, V. & Neeleman, A. 2005. Theory 9: 577−636.
Floating quantifiers and English VP ———2001. What VP-ellipsis can do,
structure. Ms. University College and what it can’t, but not why. In
London. The Handbook of Contemporary
 References 503

Syntactic Theory, M. Baltin & C. in Teotitlán del Valle Zapotec.

Collins (eds.) 439–79. Oxford: Berkeley Linguistic Society
Blackwell. Proceedings 41: 219–43.
Jones, M. A. 1994. Sardinian Syntax. Kallulli, D. 2008. There is secondary
London: Routledge. predication in there-existentials.
Jónsson, J. G. 2010. Icelandic Proceedings of the West Coast
exclamatives and the structure of Conference on Formal Linguistics
the CP layer. Studia Linguistica 64: 26: 279–87.
37–54. Kameshima, N. 1989. The syntax
Joseph, B. D. 1976. Raising in modern of restrictive and nonrestrictive
Greek: A copy process? Harvard relative clauses in Japanese. PhD
Studies in Syntax 2: 241–81. diss. University of Wisconsin–
Junker, M.-O. 1990. Floating quantifiers Madison.
and distributivity. Cahiers Kanda, Y. & Honda, M. 2018. A
Linguistiques d’Ottowa 18: 13–42. predication approach to Copy
Jurka, J. 2010 The importance of Raising from the perspective of
being a complement: CED effects evidentiality. Tsukuba English
revisited. PhD diss. University of Studies 37: 141–55.
Maryland. Kandybowicz, J. 2006. Comp‑trace
Jurka, J., Nakao, C. & Omaki, A. 2011. effects explained away. Proceedings
It’s not the end of the CED as of the West Coast Conference on
we know it: Revisiting German Formal Linguistics 25: 220–28.
and Japanese Subject Islands. Kang, B. M. 2002. Categories and
Proceedings of the West Coast meanings of Korean floated
Conference on Formal Linguistics quantifiers – with some reference
28: 124–32. to Japanese. Journal of East Asian
Kalinina, E. 2011. Exclamative Linguistics 11: 375–98.
clauses in the languages of the Kato, M. A. 1993. The distribution of
North Caucasus and the problem pronouns & null elements in object
of finiteness. In Tense, Aspect, position in Brazilian Portuguese. In
Modality and Finiteness in East- Linguistic Perspectives on Romance
Caucasian Languages, G. Authier & Languages, W. J. Ashby, M. Mithun
T. Maisak (eds.) 161–201. Bochum: & G. Perissinotto (eds.) 225–35.
Brockmeyer. Amsterdam: Benjamins.
Kalin, L. 2020. Opacity in agreement. Katz, J. J. & Postal, P.M. 1964. An
In Agree to Agree: Agreement in Integrated Theory of Linguistic
the minimalist program, P. W. Descriptions. Cambridge, MA: MIT
Smith, J. Mursell & K. Hartmann Press.
(eds.) 149–77. Berlin: Language Kauf, C. & Zeijlstra, H. 2018. Towards
Science Press. a new explanation of sequence of
Kalivoda, N. & Zyman, E. 2015. On tense. Semantics and Linguistic
the derivation of relative clauses Theory 28: 59–77.
504 References

Kaufmann, M. 2011. Interpreting Keenan, E. & Schieffelin, B. 1976a.

Imperatives. Berlin: Springer. Foregrounding referents: A
Kay, P. 2002. English subjectless tagged reconsideration of left dislocation
sentences. Language 78: 453–81. in discourse. Berkeley Linguistic
Kayne, R. S. 1981. Unambiguous Society Proceedings 2: 240–57.
paths. In Levels of Syntactic ———1976b. Topic as a discourse
Representation, R. May & J. Koster notion: A study of topic in the
(eds.) 143–83. Dordrecht: Foris. conversations of children and
———1984. Connectedness and Binary adults. In Subject and Topic, C.
Branching. Dordrecht: Foris. N. Li (ed.) 335–84. New York:
———1989. Null subjects and clitic Academic Press.
climbing. In The Null Subject Keffala, B. 2011. Resumption and gaps
Parameter, O. Jaeggli & K. Safir in English relative clauses: Relative
(eds.) 239–59. Dordrecht: Kluwer. acceptability creates an illusion of
———1991. Romance clitics, verb saving. Berkeley Linguistics Society
movement, and PRO. Linguistic Proceedings 37: 140–54.
Inquiry 22: 647–86. Keffala, B. & Goodall, C. 2011. Do
———1994. The Antisymmetry of Syntax. resumptive pronouns ever rescue
Cambridge, MA: MIT Press. illicit gaps? Poster presented at
———1998. Overt vs covert movement. 24th annual CUNY conference
Syntax 1: 128–91. on Human Sentence Processing,
———2007. Some thoughts on Stanford, CA.
grammaticalization: The case Kekalainen, K. 1985. Relative
of that. Talk delivered at XVIII clauses in the dialect of Suffolk.
Conférence Internationale de Neuphilologische Mitteilungen 86:
Linguistique Historique, UQAM, 353–57.
Montreal. Keller, F. & Lapata, M. 1998. Object
———2015. A note on some even more Drop and discourse accessibility.
unusual relative clauses. Ms. New Proceedings of the West Coast
York University. [Link] Conference on Formal Linguistics
.net/lingbuzz/002835 17: 362–74.
———2016. The unicity of there and the Keshev, M. & Meltzer-Asscher, A.
definiteness effect. Ms. New York 2017. Active dependency formation
University. in islands: How grammatical
———2019. Notes on expletive there. The resumption affects sentence
Linguistic Review 37: 209–30. processing. Language 93: 549–68.
Keenan, E. 1985. Relative clauses. In Khalifa, J.-C. 1999. A propos des
Language Typology and Syntactic relatives appositives: Syntaxe,
Description. Vol. II: Complex sémantique, pragmatique.
constructions, T. Shopen (ed.) Anglophonia 6: 7–29.
141–70. Cambridge: Cambridge Kibsgaard, M. K. M. 2019. Murder
University Press. suspect seen on video. BA thesis.
 References 505

Norwegian University of Science Svenonius (ed.) 107–24. Oxford:

and Technology. Oxford University Press.
Kidwai, A. 1999. Word order and focus Kissock, M. J. 2013. Evidence for
positions in Universal Grammar. ‘finiteness’ in Telugu. Natural
In The Grammar of Focus, G. Language & Linguistic Theory 32:
Rebuschi & L. Tuller (eds.) 213–44. 1–30.
Amsterdam: Benjamins. Kitagawa, C. 1982. Topic construction
Kim, J.-B. 2014. English copy in Japanese. Lingua 57: 175–214.
raising constructions: Argument ———1986. Subjects in English and
realization and characterization Japanese. PhD diss. University of
condition. Linguistics 52: 167–203. Massachusetts.
Kim, K.-S. 2020. On the distribution of ———1994. Subjects in Japanese and
floating quantifiers. Ms. Hankuk English. New York: Garland.
University of Foreign Studies. ———2018. Floating quantifiers in
Kim, S.-S. 2002. Intervention effects Japanese passives and beyond.
are focus effects. Japanese/Korean Journal of Japanese Linguistics 34:
Linguistics 10: 615–28. 245–79.
———2006. Intervention effects are Kjellmer, G. 1988. Conjunctional/
focus effects. Harvard Studies in adverbial which in substandard
Korean Linguistics 11: 520–33. English. Studia Anglica
Kim, S. 1999. Sloppy/strict identity, Posnaniensia 21: 125–37.
empty objects, and NP ellipsis. ———2010. And which: A note on (more
Journal of East Asian Linguistics 8: or less) coordinated relatives.
255–84. English Studies 91: 457–66.
Kim, J.-B. & Kim, J. S. 2009. English Klein, S. 1977. A base analysis of
floating quantifier constructions: A the floating quantifier in French.
non‑movement approach. Language Proceedings of the North East
and Information 13: 57–75. Linguistic Society 7: 147–63.
Kim, J.-B. & Kim, O. 2011. English how Klein, W. 1998. Assertion and
come construction: A double life. finiteness. In Issues in the Theory
Paper presented to the Arizona of Language Acquisition: Essays in
Linguistics Circle 5, 28–30 October. honor of Jürgen Weissenborn, N.
Kiss, K. É. 1995. Introduction. Dittmar & Z. Penner (eds.), 225–45.
In Discourse-Configurational Bern: Peter Lang.
Languages, K. É. Kiss (ed.) 3–27. ———2006. On finiteness. In Semantics
Oxford: Oxford University Press. in Acquisition, V. van Geenhoven
———1998. Identificational focus versus (ed.) 245–72. New York: Springer.
informational focus. Language 74: Klima, E. S. 1964. Negation in English.
245–73. In The Structure of Language, J. A.
———2001. The EPP in a topic- Fodor & J. J. Katz (eds.) 246–323:
prominent language, In Subjects, Englewood Cliffs, NJ: Prentice-
Expletives and the EPP, P. Hall.
506 References

Ko, H. & Oh, E. 2012. A hybrid Koppen, M. van 2005. One probe – two
approach to floating quantifiers. goals: Aspects of agreement in
Linguistic Research 29: 69–106. Dutch dialects, PhD diss. University
Kobayashi, K. 2018. A note on copy of Leiden.
raising constructions in English. ———2017. Complementizer Agreement.
Ms. Kanto Gakuin University. In The Wiley-Blackwell Companion
Kobuchi-Philip, M. 2003. Distributivity to Syntax, 2nd edition, M. Everaert
and the Japanese floating numeral & H. van Riemsdijk (eds.) 923–62.
quantifier. PhD diss. City University London: Wiley-Blackwell.
of New York. Kostopoulos, E. E. 2017. Revising
———2007. Floating numerals and floating floating quantifiers: The syntax
quantifiers. Lingua 117: 814–31. of Modern Greek ola. MA thesis.
König, E. 1991. The Meaning of Focus University of York.
Particles. London: Routledge. Koster, J. & Zwart, J.-W. 2000.
———2017. Syntax and semantics of Transitive expletive constructions
additive focus markers from a and the object shift parameter.
cross-linguistic perspective. In Focus Linguistics in the Netherlands 17:
on Additivity: Adverbial modifiers 159–70.
in Romance, Germanic and Krapova, I. 2002. On the left periphery
Slavic languages, A. M. de Cesare of the Bulgarian sentence.
Greenwald & C. Andorno (eds.) University of Venice Working
23–44. Amsterdam: Benjamins. Papers in Linguistics 12: 107–28.
Koopman, H. 2000. The Syntax of Krapova, I. & Cinque, G. 2008. On the
Specifiers and Heads. London: order of wh‑phrases in Bulgarian
Routledge. multiple wh‑fronting. In Formal
———2007. Topics in imperatives. In Description of Slavic Languages,
Imperative Clauses in Generative G. Zybatow, L. Szucsich, U.
Grammar: Studies in honour Junghanns, & R. Meyer (eds.)
of Frits Beukema, W. van der 318–36. Frankfurt am Main: Peter
Wurff (ed.) 153–80. Amsterdam: Lang.
Benjamins. Kroch, A. 1981. On the role of
———2010. On Dutch allemaal and West resumptive pronouns in amnestying
Ulster English all. In Structure island constraint violations.
Preserved, W. Zwart & J. van de Chicago Linguistic Society
Vries (eds.) 267–76. New York: Proceedings 17: 125–35.
Academic Press. Kuno, S. 1981. Functional Syntax.
Koopman, H. & Sportiche, D. 1991. The Syntax and Semantics 13: 117–35.
position of subjects. Lingua 85: Kuroda, S.-Y. 1968. English
211–58. relativization and certain related
Koopman, H. & Szabolsci, A. 2000. problems. Language 44: 244–66.
Verbal Complexes. Cambridge, MA: ———1988. Whether we agree or not.
MIT Press. Lingvisticae Investigationes 12: 1–47.
 References 507

Labov, W. 1969. Contraction, deletion Lakoff, G. 1969. On derivational

and the inherent variability of the constraints. Chicago Linguistic
English copula. Language 45: 715–62. Society Proceedings 5: 117–39.
———1995. The case of the missing Lambrecht, K. 1994 Information
copula: The interpretation of zeroes Structure and Sentence Form.
in African-American English. In Cambridge: Cambridge University
An Invitation to Cognitive Science, Press.
L. Gleitman & M. Liberman (eds.) Lambrecht, K. & Lemoine, K. 1996. Vers
25–54. Cambridge, MA: MIT Press. une grammaire des compléments
Lacerda, R. 2016. Rebel without a case: zéro en français parlé. In Absence
Quantifier floating in Brazilian de Marques et Représentation de
Portuguese and Spanish. In The l’Absence, J. Chuquet & M. Frid
Morphosyntax of Portuguese and (eds.) 279–309. Rennes: Presses
Spanish in Latin America, M. A. Universitaires de Rennes.
Kato & F. Ordóñez (eds.) 78–106. ———2005. Definite null objects in
Oxford: Oxford University Press. (spoken) French: A construction-
Laenzlinger, C. 1993. Principles for a grammar account. In Grammatical
formal and computational account Constructions: Back to the roots
of adverbial syntax. MA thesis. M. Fried & H. C. Boas (eds.) 13–55.
University of Geneva. Amsterdam: Benjamins.
———1996. Adverb syntax and phrase Landa, M. A. 1995. Conditions on null
structure. In Configurations: Essays objects in Basque Spanish and
on structure and interpretation, their relation to leísmo and clitic
A. M. Di Sciullo (ed.) 99–127. doubling. PhD diss. University of
Somerville, MA: Cascadilla Press. Southern California, Los Angeles.
———1999. Comparative Studies in Landau, I. 2004. The scale of finiteness
Word Order Variation: Adverbs, and the calculus of Control. Natural
pronouns, and clause structure Language and Linguistic Theory 22:
in Romance and Germanic. 811–77.
Amsterdam: Benjamins. ———2006. Severing the distribution of
———2000. More on adverb syntax PRO from case. Syntax 9: 153–70.
and phrase structure. In Adverbs ———2007. EPP extensions. Linguistic
and Adjunction, A. Alexiadou & P. Inquiry 38: 485–523.
Svenonius (eds.) 103–22. Potsdam: ———2009. This construction looks
Institut für Linguistik, Universität like a copy is optional. Linguistic
Potsdam. Inquiry 40: 343–46.
Lahiri, U. 1991. Embedded ———2011. Predication vs. aboutness in
interrogatives and predicates that copy raising. Natural Language and
embed them. PhD diss. MIT. Linguistic Theory 29: 779–813.
———1998. Focus and negative polarity Lansens, E. 2019. Subject ellipsis
in Hindi. Natural Language on social media: Analyzing
Semantics 6: 57–123. extralinguistic variables ‘age’ and
508 References

‘motive’ in relation to first person Washington, DC: Georgetown

subject ellipsis in finite clauses on University Press.
Twitter. Ms, Ghent University. ———1998. Some reconstruction
Lappin, S. 1983. Θ‑roles and riddles. in Penn Working Papers in
NP‑movement. Proceedings of the Linguistics 5: 83–98.
North East Linguistics Society 13: ———1999. Chains of arguments. In
121–28. Working Minimalism, S. D. Epstein
———1984. Predication and raising. & N. Hornstein (eds.) 189–215.
Proceedings of the North East Cambridge, MA: MIT Press.
Linguistics Society 14: 236–52. ———2001. A note on the EPP. Linguistic
Larjavaara, M. 2000. Présence ou Inquiry 32: 356–62.
Absence de l’Objet: Limites du ———2003. Minimalist Investigations
possible en français contemporain. in Linguistic Theory. London:
Helsinki: Academia Scientiarum Routledge.
Fennica. ———2006. Conceptions of the cycle.
Larsen, D. 2014. Particles and particle- In Wh‑Movement: Moving On, L.
verb constructions in English and Cheng & N. Corver (eds.) 197–216.
other Germanic Languages. PhD Cambridge, MA: MIT Press.
diss. University of Delaware. Lasnik, H. & Saito, M. 1992. Move α:
Larson, B. 2014. Russian comitatives Conditions on its application and
and the ambiguity of adjunction. output. Cambridge, MA: MIT Press.
Journal of Slavic Linguistics 22: Lasnik, H. & Sobin, N. 2000. The who/
11–49. whom puzzle: On the preservation
Larson, R. K. 1985. On the syntax of an archaic feature. Natural
of disjunction scope, Natural Language and Linguistic Theory 18:
Language and Linguistic Theory 3: 343–71.
217–64. Lasser, I. 1997. Finiteness in adult
Lasnik, H. 1981. Restricting the theory and child German. PhD. diss. City
of transformations. In Explanation University of New York.
in Linguistics, N. Hornstein & D. Law, P. 1996. Remarks on the verb
Lightfoot (eds.) 152–73. London: be and the expletive there in
Longman. English. Linguistische Berichte 166:
———1992. Case and expletives 492–529.
revisited: Notes towards a Lebeaux, D. 1991. Relative clauses,
parametric account. Linguistic licensing and the nature of
Inquiry 23: 381–405. derivation. In Syntax and
———1995. Verbal Morphology: Semantics 25: Perspectives on
Syntactic Structures meets phrase structure, S. Rothstein
the minimalist program. In (ed.) 209–39. New York: Academic
Evolution and Revolution in Press.
Linguistic Theory, H. Campos & ———1995. Where does Binding Theory
P. Kempchinsky (eds.) 251–75. apply? University of Maryland
 References 509

Working Papers in Linguistics 3: 17–35. Oxford: Oxford University

63–88. Press.
Lee, P. C.-W. 2017. Anaphoric object Lin, T.-H. J. 2011. Finiteness of
drop in Chinese. In Order and clauses and raising of arguments
Structure in Syntax II: Subjecthood in Mandarin Chinese. Syntax 14:
and argument structure, M. Sheehan 48–73.
& L. R. Bailey (eds.) 329–38. Berlin: Lipták, A. 2006. Word order in
Language Science Press. Hungarian exclamatives. Acta
Lee, Y. 2004. The syntax and semantics Linguistica Hungarica 53: 343–91.
of focus particles. PhD diss. MIT. Lindstrom, A. M. 2020. Unexpressed
Legate, J. A. 2002. Warlpiri: Theoretical Subjects in English: An
implications. PhD diss. MIT. empirical analysis of narrative
———2005. Phases and cyclic agreement. and conversational discourse.
MIT Working Papers in Linguistics Washington, DC: Rowman &
49: 234–53. Littlefield.
———2014. Voice and v: Lessons from Llinàs-Grau, M. & Fernández-Sánchez,
Achenese. Cambridge, MA: MIT J. F. 2013. Reflexiones en torno a la
Press. supresión del complementante en
Lemke, R., Horch, E. & Reich, inglés, español y catalán. Revista
I. 2017. Optimal encoding: Española de Lingüística 43: 55–88.
Information Theory constrains Lobeck, A. 1995. Ellipsis: Functional
article omission in newspaper heads, licensing and identification.
headlines. In Proceedings of the Oxford: Oxford University Press.
15th Conference of the European Löbel, E. 1989. Q as a functional
Chapter of the Association for category. In Syntactic Phrase
Computational Linguistics: Volume Structure Phenomena, C. Bhatt, E.
2, Short papers, 131–35. Valencia: Löbel, & C. Schmidt (eds.) 133–58.
Association for Computational Amsterdam: Benjamins.
Linguistics. Lohndal, T. 2009. Comp‑t effects:
Lewis, R. E. 2013. Complementiser Variation in the position and
agreement in Najdi Arabic. MA features of C. Studia Linguistica 63:
thesis. University of Kansas. 204–32.
Li, G. 1998. Null object and VP Ellipsis Longobardi, G. 1994. Reference and
in Chinese? Proceedings of the proper names. Linguistic Inquiry
North American Conference on 25: 609–66.
Chinese Linguistics 9: 155–72. Los ———1996. The syntax of N‑raising: A
Angeles: GSIL. minimalist theory. OTS Working
Light, C. 2015. Expletive there in Papers no 5. Utrecht: Research
West Germanic. In Syntax Over Institute for Language and Speech.
Time: Lexical, morphological, and ———2001. The structure of DPs:
information-structural interactions, Some principles, parameters and
T. Biberauer & T. Walkden (eds.) problems. In The Handbook of
510 References

Contemporary Syntactic Theory, M. Mack, J. E. 2010. Information structure

Baltin & C. Collins (eds.) 562–603. and the licensing of English
Oxford: Blackwell. subjects. PhD diss. Yale University.
Loock, R. 2003. Les fonctions des Mack, J. E., Clifton, C. Jr., Frazier, L.
propositions subordonnées & Taylor, P. V. 2012. (Not) hearing
relatives appositives en discours. optional subjects: The effects
Anglophonia 12: 113–31. of pragmatic usage preferences.
———2005. Appositive relative clauses Journal of Memory and Language
in contemporary written and 67: 211–23.
spoken English: Discourse functions Mack, J. E. & Fuerst, Y. 2009. English
and competitive structures. PhD optional expletives and the
diss. University of Lille III. pragmatics of judgments. Paper
———2007a. Appositive relative clauses presented to the 45th annual
and their functions in discourse. meeting of the Chicago Linguistic
Journal of Pragmatics 39: 336–62. Society, 17 April.
———2007b. Are you a good which or a Mackenzie, J. L. 1998 The basis of
bad which? The relative pronoun as syntax in the holophrase. In
a plain connective. In Connectives Functional Grammar and Verbal
as Discourse Landmarks, A. Celle & Interaction, M. Hannay & A.
R. Huart (eds.) 71–87. Amsterdam: M. Bolkestein (eds.) 267–96.
Benjamins. Amsterdam: Benjamins.
———2010. Appositive Relative Clauses Maekawa, T. 2007. The English left
in English: Discourse functions and periphery in linearisation-based
competing structures. Amsterdam: HPSG. PhD diss. University of
Benjamins. Essex.
López-Couso, M. J. & Méndez-Naya, B. Mahajan, A. 1994. Active passives.
2012. On the use of as if, though, Proceedings of the West Coast
and like in present-day English Conference on Formal Linguistics
complementation structures. 23: 286–301.
Journal of English Linguistics 40: Maienborn, C. 2001. On the position
172–95. and interpretation of locative
Loss, S. S. & Wicklund, M. 2020. modifiers. Natural Language
Is English resumption different Semantics 9: 191–240.
in appositive relative clauses? Maki, H., Kaiser, L. & Ochi, M. 1999.
Canadian Journal of Linguistics 65: Embedded topicalization in English
25–51. and Japanese. Lingua 109: 1–14.
Lowe, J. J. 2019. The syntax and Malamud-Makowski, M. 1994. The
semantics of nonfinite forms. structure of IP: Evidence from
Annual Review of Linguistics 5: acquisition data. Paper presented
309–28. at the 18th Annual Boston
Lyons, C. 1999. Definiteness. Cambridge: University Conference on Language
Cambridge University Press. Development, Boston, MA.
 References 511

Maling, J. 1976. Notes on quantifier ———1999. Case, the extended projection

postposing. Linguistic Inquiry 7: principle, and minimalism. In
708–18. Working Minimalism, S. D.
Marandin, J.-M. 2008. The exclamative Epstein & N. Hornstein (eds.) 1–26.
clause type in French. In Cambridge, MA: MIT Press.
Proceedings of the HPSG08 ———2001. Null case and the
Conference, S. Müller (ed.) 436–56. distribution of PRO. Linguistic
Stanford, CA: CSLI Publications. Inquiry 32: 141–66.
Marantz, A. 1984. On the Nature of Martins, A. M. & Nunes, J. 2005.
Grammatical Relations. Cambridge, Raising issues in European and
MA: MIT Press. Brazilian Portuguese. Journal of
Manetta, E. 2007. Unexpected left Portuguese Linguistics 4: 53–77.
dislocation: An English corpus ———2009. Syntactic change as
study. Journal of Pragmatics 39: chain reaction: The emergence
1029–35. of hyper‑raising in Brazilian
Manninen, S. 2005. Review of Cinque Portuguese. In Historical Syntax
(1999). Journal of Linguistics 41: and Linguistic Theory, P. Crisma
452–57. & G. Longobardi (eds.) 144–57.
Manzini, M. R. & Roussou, A. 2000. A Oxford: Oxford University Press.
minimalist approach to A‑movement Massam, D. 1987. Middles, tough, and
and control. Lingua 110: 409–47. recipe context constructions in
———2019. Morphological and syntactic English. Proceedings of the North
(non)finiteness: A comparison East Linguistics Society 18: 315–32.
between English and Balkan ———1989. Null objects and non-
languages. Working Papers in thematic subjects in middles and
Linguistics and Oriental Studies 5: tough constructions. Toronto
195–229. Working Papers in Linguistics, 10:
Manzini, M. & Savoia, L. M. 2018. Finite 95–118.
and nonfinite complementation, ———1992. Null objects and non-
particles and control in Aromanian, thematic subjects. Journal of
compared to other Romance Linguistics 28: 115–37.
varieties and Albanian. Linguistic Massam, D., Bamba, K. & Murphy, P.
Variation 18: 215–64. 2017. Obligatorily null pronouns
Mårdh, I. 1980. Headlinese: On the in the instructional register and
grammar of English front page beyond. Linguistic Variation 17:
headlines. Malmö: CWK Gleerup. 272–91.
Martín, L. M. 2019. Negative Inversion Massam, D. & Roberge, Y. 1989. Recipe
in standard English. Odisea 20: context null objects in English.
91–116. Linguistic Inquiry, 20: 134–39.
Martin, R. 1996. A minimalist theory Massam, D. & Stowell, T. (eds.) 2017.
of PRO and control. PhD diss. Register Variation and Syntactic
University of Connecticut, Storrs. Theory. Amsterdam: Benjamins.
512 References

Matushansky, O. 1995. Le sujet nul & P. Portner (eds.) 86–126.

dans les propositions à temps Washington, DC: Georgetown
fini en anglais. Maîtrise paper, University Press.
University of Paris VIII. ———2017a. Observations and
———2002. Tipping the scales: The syntax speculations on resumption (in
of scalarity in the complement of Irish). Ms. University of California,
seem. Syntax 5: 219–76. Santa Cruz.
Mauck, S., Pak, M., Portner, P. & ———2017b. Resumption. In The Wiley-
Zanuttini, R. 2005. Imperative Blackwell Companion to Syntax,
subjects: A cross-linguistic 2nd edition, M. Everaert & H. van
perspective. Georgetown University Riemsdijk (eds.) 3809–38. Hoboken,
Working Papers in Theoretical NJ: John Wiley & Sons.
Linguistics 4: 135–52. McDaniel, D. & Cowart, W. 1999.
Mayr, C. 2010. Ç. In The Experimental evidence of a
Complementiser Phase: Subjects minimalist account of English
and wh‑dependencies, P. resumptive pronouns. Cognition 70:
Panagiotidis (ed.) 117–42. Oxford: B15–B24.
Oxford University Press. McKee, C. & McDaniel, D. 2001.
McCawley, J. D. 1981. The syntax Resumptive pronouns in English
and semantics of English relative relative clauses. Language
clauses. Lingua 53: 99–149. Acquisition 9: 113–56.
———1993. Gapping with shared McNally, L. 1992. VP-coordination and
operators. Berkeley Linguistic the VP‑internal subject hypothesis.
Society Proceedings 19: 245–54. Linguistic Inquiry 23: 336–41.
McCloskey, J. 1997. Subjecthood and Medeiros, D. J. 2013. Approaches
subject positions. In Elements to the syntax and semantics of
of Grammar, L. Haegeman (ed.) imperatives. PhD diss. University of
197–235. Dordrecht: Kluwer. Michigan.
———2000. Quantifier Float and Meira, B. J. 2004. Review of Cinque
Wh‑Movement in an Irish English. (1999). Cadernos de Estudos
Linguistic Inquiry 31: 57–84. Lingüisticos 46: 283–90.
———2006a. Resumption, successive Meltzer-Asscher, A. 2012. Verbal
cyclicity, and the locality of passives in English and Hebrew:
operations. In Derivations in A comparative study. In The Theta
Minimalism, S. D. Epstein & T. D. System: Argument structure at the
Seely (eds.) 184–226. Cambridge: interface, M. Everaert, M. Marelj
Cambridge University Press. & T. Siloni (eds.) 280–307. Oxford:
———2006b. Questions and questioning Oxford University Press.
in a local English. In Negation, Merchant, J. 1996. Object scrambling
Tense and Clausal Architecture: and quantifier float in German.
Cross-linguistic Investigations, R. Proceedings of the North East
Zanuttini, H. Campos, E. Herburger Linguistic Society 26: 179–93.
 References 513

———1999. The syntax of silence: PPs may vary by category. The

Sluicing, islands, and identity in Linguistic Review 36: 325–41.
ellipsis. PhD diss. University of Michaelis, L. 2001. Exclamative
California, Santa Cruz. constructions. Language Typology
———2001. The Syntax of Silence: and Language Universals 2:
Sluicing, Islands and Identity in 1038–50.
Ellipsis. Oxford: Oxford University Michaelis, L. & Lambrecht, K. 1996.
Press. The exclamative sentence type in
———2002. Swiping in Germanic. In English. In Conceptual Structure,
Studies in Comparative Germanic Discourse and Language, A.
Syntax, W. Abraham & J.-W. Goldberg (ed.) 375–89. Stanford,
Zwart (eds.) 295–321. Amsterdam: CA: CSLI.
Benjamins. Milambiling, L. 2011. Null object
———2003. Sluicing. Ms. University of constructions in Tagalog.
Chicago. Proceedings of the Canadian
———2004a. Fragments and ellipsis. Linguistics Association 2011.
Linguistics and Philosophy 27: Miller, J. 1988. That: A relative
661–738. pronoun? In Edinburgh Studies
———2004b. Resumptivity and in the English Language, J. M.
non‑movement. Studies in Greek Anderson & M. Macleod (eds.)
Linguistics 24: 471–81. 113–19. Edinburgh: John Donald.
———2008a. Variable island repair Milner, J.-C. 1978. De la Syntaxe à
under ellipsis. In Topics in l’Interprétation: Quantités, insultes,
Ellipsis, K. Johnson (ed.) 132–53. exclamations. Paris: Éditions du
Cambridge: Cambridge University Seuil.
Press. Milsark, G. 1977. Toward an
———2008b. An asymmetry in voice explanation of certain peculiarities
mismatches in VP ellipsis and of the existential construction in
pseudogapping. Linguistic Inquiry English. Linguistic Analysis 3:
39: 169–79. 1–29.
———2013a. Diagnosing ellipsis. In Miró, E. C. 2006. ‘Wh’-exclamatives in
Diagnosing Syntax, L. Cheng & Catalan. PhD diss. Universitat de
N. Corver (eds.) 537–42. Oxford: Barcelona.
Oxford University Press. ———2008. Deconstructing
———2013b. Polarity items under exclamations. Catalan Journal of
ellipsis. In Diagnosing Syntax, L. Linguistics 7: 41–90.
Cheng & N. Corver (eds.) 441–62. Miyagawa, S. 2005. On the EPP. MIT
Oxford: Oxford University Press. Working Papers in Linguistics 49:
———2014. ‘Some definitions’. Ms. 201–36.
University of Chicago. ———2006a. Moving to the edge. In
———2019. ‘Roots don’t select, categorial Proceedings of the 2006 KALS-
heads do: Lexical-selection of KASELL International Conference
514 References

on English and Linguistics, 3–18. Morgan, A. M., & Wagers, M. 2018.

Busan: Pusan National University. English resumptive pronouns are
———2006b. Locality in syntax and more common where gaps are less
floated numeral quantifiers in acceptable. Linguistic Inquiry 49:
Japanese and Korean. Japanese/ 1–24.
Korean Linguistics 14: 270–82. Morgan, J. 1972. Some aspects of
———2010. Why Agree? Why Move? relative clauses in English and
Unifying agreement-based and Albanian. In The Chicago Which
discourse-configurational languages. Hunt: Papers from the Relative
Cambridge, MA: MIT Press. Clause Festival, P. M. Peranteau,
———2017a. Topicalization. Ms. MIT/ J. N. Levi & G. C. Phares (eds.)
University of Tokyo. 63–72. Chicago: Chicago Linguistic
———2017b. Numeral quantifiers. In Society.
Handbook of Japanese Syntax, M. ———1973. Sentence fragments and
Shibatani, S. Miyagawa & H. Noda the notion ‘sentence’. In Issues in
(eds.) 581–610. Berlin: De Gruyter Linguistics: Papers in honor of
Mouton. Henry and Renee Kahane, B. B.
Miyagawa, S. & Arikawa, K. 2007. Kachru (ed.) 719–51. Urbana, IL:
Locality in syntax and floating University of Illinois Press.
numeral quantifiers. Linguistic Moro, A. 1997. The Raising of
Inquiry 38: 645–70. Predicates. Cambridge: Cambridge
Moncomble, F. 2018. The deviant University Press.
syntax of headlines and its role in ———2006. Existential sentences and
the pragmatics of headlines. E-rea expletive there. In The Blackwell
15(2). [Link] Companion to Syntax: vol.2, M.
erea.6124 Everaert & H. van Riemsdijk (eds.)
Montgomery, M. & Hall, J. S. 2004. 210–36. Oxford: Blackwell.
Dictionary of Smoky Mountain Müller, G. 2010. On deriving CED
English. Knoxville, TN: University effects from the PIC. Linguistic
of Tennessee Press. Inquiry 41: 35–82.
Moon, G. G.-S. 2001. Grammatical ———2014. Resumption by buffers:
and discourse properties of the German relative clauses.
imperative subject in English. PhD Linguistische Berichte 239: 267–95.
diss. Harvard University. ———2015. Algorithms, representations,
Moore, J. 1998. Turkish copy‑raising and ATR ratios. Ms. Universität
and A‑chain locality. Natural Leipzig. [Link]
Language and Linguistic Theory 16: lingbuzz/002551
149–89. ———2017. Structure removal: An
de Moraes Sieiro, P. L. 2020. Sentential argument for feature-driven Merge.
wh‑exclamatives in Brazilian Glossa 2(1): 28.
Portuguese. MA diss. University of Munaro, N. 2001. Free relatives as
Brazil, Brasilia. defective wh elements: Evidence
 References 515

from the NorthWestern Italian Linguistics, S. Miyagawa & M.

Dialects. In Romance Languages Saito (eds.) 287–319. Oxford:
and Linguistic Theory 1999, Y. Oxford University Press.
D’Hulst, J. Rooryk & J. Schroten Namai, K. 2000. Gender features in
(eds.) 281–306. Amsterdam: English. Linguistics 38: 771–79.
Benjamins. Nanyan, V. 2013. Subject omission in
———2003. On some differences between English diaries. MA thesis. Ghent
exclamative and interrogative University.
wh‑phrases in Bellunese: Further Napoli, D. J. 1982. Initial material
evidence for a split-CP hypothesis. deletion in English. Glossa 16:
In The Syntax of Italian Dialects, 85–111.
C. Tortora (ed.) 137–51. Oxford: Nariyama, S. 2004. Subject ellipsis in
Oxford University Press. English. Journal of Pragmatics 36:
Muñoz Pérez, C., Verdecchia, M. 237–64.
& Carranza, F. 2021. Prosodic ———2006. Pragmatic information
constraints on wh‑extraction from extraction from subject ellipsis
infinitival clauses. Talk presented in informal English. Proceedings
at the Linguistic Symposium on of the 3rd Workshop on Scalable
Romance Languages (LSRL 51), Natural Language Understanding,
University of Illinois at Urbana- 1–8. New York: Association for
Champaign. Slides available online Computational Linguistics.
at [Link] Neal, W. F. 1985. Bill Neal’s Southern
Nakajima, H. 2006. Adverbial cognate Cooking. Chapel Hill, NC:
objects. Linguistic Inquiry 37: University of North Carolina Press.
674–84. Neeleman, A. 1994. Complex
Nakamura, M. 1983. A predicates. PhD diss. University of
nontransformational approach to Utrecht.
quantifier floating phenomena. Nemes, M. 2019. The syntax of English
Studies in English Linguistics 11: news headlines: A truncation
1–10. approach. BA thesis. Eötvös Loránd
Nakamura, T. 2009. Headed relatives, University.
free relatives, and determiner- Netkachev, I. 2020. Are there floating
headed free relatives. English quantifiers in Indonesian? Research
Linguistics 26: 329–55. paper no. WP BRP 105, Moscow:
Nakanishi, K. 2004. Domains of HSE University.
measurement: Formal properties Neubauer, P. 1970. On the notion
of non-split/split quantifier chopping rule. Chicago Linguistic
constructions. PhD diss. University Society Proceedings 6: 400–07.
of Pennsylvania. Nevins, A. 2005. Derivations without
———2008. Syntax and semantics of the Activity Condition. MIT
floating numeral quantifiers. In Working Papers in Linguistics 49:
Oxford Handbook of Japanese 287–310.
516 References

Newbrook, M. 1998. Which way? chains in the minimalist program.

That way? Variation and ongoing PhD diss. University of Maryland.
changes in the English relative ———1999. Linearization of chains
clause. World Englishes 17: 43–59. and phonetic realization of chain
Newman, J & Teddiman, L. 2010. First links. In Working Minimalism, S.
person pronouns in online diary D. Epstein & N. Hornstein (eds.)
writing. In Handbook of Research 217–49. Cambridge, MA: MIT Press.
on Discourse Behavior and Digital ———2001. Sideward movement.
Communication: Language structures Linguistic Inquiry 32: 303–44.
and social interaction, vol. 1, R. ———2004. Linearization of Chains and
Taiwo (ed.) 281–95. Hershey, PA: Sideward Movement. Cambridge,
Information Science Reference. MA: MIT Press.
Newmeyer, F. J. 2003. Grammar is ———2008. Inherent case as a licensing
grammar and usage is usage. condition for A‑movement: The
Language 79: 682–707. case of Hyper‑Raising constructions
———2005. A reply to the critiques of in Brazilian Portuguese. Journal of
‘Grammar is grammar and usage is Portuguese Linguistics 7: 83–108.
usage’. Language 81: 229–36. Nunes, J. & Uriagereka, J. 2000.
———2006a. On Gahl and Garnsey on Cyclicity and extraction domains.
grammar and usage. Language 82: Syntax 3: 20–43.
399–404. Nye, R. 2013. How complement clauses
———2006b. Grammar and usage: distribute: Complementiser-how
A response to Gregory R. Guy. and the case against clause-type.
Language 82: 705–6. PhD diss. University of Ghent.
Nikolaeva, I. 2007 (ed). Finiteness: Obenauer, H.-G. 1994. Aspects
Theoretical and empirical de la syntaxe A‑barre: Effets
foundations. Oxford: Oxford d’intervention et mouvement des
University Press. quantifieurs. PhD diss. University
Nilsen, Ø. 2003. Eliminating positions: of Paris VIII.
Syntax and semantics of sentence Ochi, M. 2004. How come and other
modification. Utrecht: LOT. adjunct wh‑phrases: A cross-
Nomura, T. 2006. ModalP and linguistic perspective. Language
Subjunctive Present. Tokyo: Hituzi and Linguistics 5: 29–57.
Syobo. Oh, S.-Y. 2005. English zero anaphora
Norris, M. 2014. A theory of nominal as an interactional resource.
concord. PhD diss. University of Research on Language and Social
California, Santa Cruz. Interaction 38: 267–302.
Nouwen, R. & Chernilovskaya, A. 2015. ———2006. English zero anaphora as an
Two types of wh‑exclamatives. interactional resource II. Discourse
Linguistic Variation 15: 201–24. Studies 8: 817–46.
Nunes, J. 1995. The copy theory of Ohara, M. 2007. Object drop in English
movement and linearization of and Japanese. Annals of Foreign
 References 517

Studies 67: 69–80. Kobe City the Second International Conference

University of Foreign Studies. on Dependency Linguistics, 272–81.
Olsen, S. 1989. AGR(eement) in the Prague: Matfyzpress.
German noun phrase. In Syntactic Ostrove, J. 2018. When φ-agreement
Phrase Structure Phenomena, C. targets topics: The view from San
Bhatt, E. Löbel, & C. Schmidt (eds.) Martín Peras Mixtec. PhD diss.
39–49. Amsterdam: Benjamins. University of California, Santa
Omaki, A. & Nakao, C. 2010. Does Cruz.
English resumption really help to Otoguro, R. & Snijders, L. 2016.
repair island violations? Snippets Syntactic, semantic and
21: 11–12. information structures of floating
O’Neil, J. 1995. Out of control. quantifiers. In Proceedings of the
Proceedings of the North East Joint 2016 Conference on Head-
Linguistics Society 25: 361–71. Driven Phrase Structure Grammar
O’Neill, T. E. 2015. The domain of and Lexical-Functional Grammar,
finiteness: Anchoring without tense D. Arnold, M. Butt, B. Crysmann,
in copular amalgam sentences. PhD T.H. King & S. Müller (eds.) 478–98.
diss. City University of New York. Stanford, CA: CSLI Publications.
Ono, H. 2006. An investigation of Ott, D. 2012a. Movement and ellipsis
exclamatives in English and in contrastive left-dislocation.
Japanese: Syntax and sentence Proceedings of the West Coast
processing. PhD diss. University of Conference on Formal Linguistics
Maryland. 30: 281–91.
Ono, H. & Fujii, T. 2006. English ———2012b. Local Instability: Split
wh‑exclamatives and the role of topicalization and quantifier float
T‑to‑C in wh‑clauses. University in German. Berlin: de Gruyter.
of Maryland Working Papers in ———2014. An ellipsis approach to
Linguistics 14: 163–87. contrastive left-dislocation.
Oomen, U. 1979. Structural properties Linguistic Inquiry 45: 269–303.
of English exclamatory sentences. ———2015. Connectivity in left-
Folia Linguistica 13: 159–74. dislocation and the composition
Oosterhof, A. & Rawoens, G. of the left periphery. Linguistic
2017. Register variation and Variation 15: 225–90.
distributional patterns in article Ozaki, S. 2010. Subjectless sentences in
omission in Dutch. Linguistic English. Nagoya Bunri University
Variation 17: 205–28. Working Papers in Linguistics 10:
Ortega-Santos, I. 2013. Corrective focus 35–46.
at the right edge in Spanish. Lingua Öztürk, B. 2005. Case, Referentiality,
131: 112–35. and Phrase Structure. Amsterdam:
Osborne, T. 2013. The distribution of Benjamins.
floating quantifiers: A dependency Paesani, K. 2006. Extending the non-
grammar analysis. Proceedings of sentential analysis: The case of
518 References

special registers. In The Syntax Recherches Linguistiques de

of Non-sententials, L. Progovac, Vincennes 33: 111–34.
K. Paesani, E. Casielles & E. ———2014. New Perspectives on Chinese
Barton (eds.) 147–82. Amsterdam: Syntax. Berlin: De Gruyter.
Benjamins. Paul, I. & Massam, D. 2020. Recipes
Palacios Roman, M. 2018. English in Malagasy and other languages.
and Spanish exclamatives: A Proceedings of Austronesian Formal
comparative study of the left Linguistics Association 27: 98–112.
periphery. Bachelor thesis. Leiden National University of Singapore.
University. Paunovic, Z. 1997. Some notes on
Panitz, E. J. 2015. Null objects in floating quantifiers. MA thesis.
Brazilian Portuguese, revisited. University of Essex.
Caderno de Squibs 1: 25–34. Pearce, E. 1999. Topic and focus in
Paoli, S. 2003. COMP and the left a head-initial language: Maori.
periphery: Comparative evidence Toronto Working Papers in
from Romance. PhD diss. University Linguistics 16: 249–63.
of Manchester. Pérez-Sabatar, C. 2012. The
———2007. The fine structure of the left linguistics of social networking:
periphery: COMPs and subjects: A study of writing conventions
Evidence from Romance. Lingua on Facebook. Linguistik Online
117: 1057–79. 56: 81–93.
———2014. Defective object clitic Perlmutter, D. M. 1968. Deep and
paradigms and the relation between surface structure constraints in
language development and loss. syntax. PhD diss. MIT.
Journal of Linguistics 50: 143–83. ———1970. The two verbs begin.
Park, B.-S. 2005. Focus, parallelism, In Readings in English
and identity in ellipsis. PhD diss. Transformational Grammar, R. A.
University of Connecticut, Storrs. Jacobs & P. S. Rosenbaum (eds.)
Parry, M. 2007. La frase relativa (con 107–19. Waltham, MA: Ginn.
antecedente) negli antichi volgari ———1971. Deep and Surface Structure
dell’Italia nord-occidentale. Constraints in Syntax. New York:
LabRomAn 1: 9–32. Holt, Rinehart & Winston.
Patterson, G. & Caponigro, I. 2016. The Perlmutter, D. & Soames, S. 1979.
puzzling degraded status of who Syntactic Argumentation and the
free relative clauses in English. Structure of English. Berkeley, CA:
English Language and Linguistics University of California Press.
20: 341–52. Pesetsky, D. 1982a. Paths and
Paul, I. (2017). Reduced structure in categories. PhD diss. MIT.
Malagasy headlines. Linguistic ———1982b. Complementiser-trace
Variation 17: 292–308. phenomena and the Nominative
Paul, W. 2005. Low IP area and left Island Condition. The Linguistic
periphery in Mandarin Chinese. Review 1: 297–343.
 References 519

———1989. Language-particular Phillips, C. 1996. Order and structure.

processes and the Earliness PhD diss. MIT.
Principle. Ms. MIT. ———2003. Linear order and constituency.
———2016. Complementizer-trace Linguistic Inquiry 34: 37–90.
effects. Ms. MIT. [Link] Pierce, A. E. 1992. Language
.net/lingbuzz/002385 Acquisition and Syntactic Theory.
———2020. Lack of ambition as Dordrecht: Kluwer.
explanation when a clause is Platzack, C. 2001. Multiple interfaces.
reduced. Ms. MIT. Text of paper In Cognitive Interfaces: Constraints
presented to the 13th Brussels on linking cognitive information,
Conference on Generative U. Nikanne & E. van der Zee (eds.)
Linguistics, December. 21–53. Oxford: Oxford University
———2021. Exfoliation: Towards a Press.
derivational theory of clause size. ———2004. Cross-linguistic word order
Ms. MIT. [Link] variation at the left periphery: The
lingbuzz/004440 case of object-first main clauses.
Pesetsky, D. & Torrego, E. 2001. Studies in Natural Language and
T‑to‑C movement: Causes and Linguistic Theory 59: 191–210.
consequences. In Ken Hale: A life Platzack, C. & Rosengren, I. 1997.
in language, M. Kenstowicz (ed.) On the subject of imperatives:
355–426. Cambridge MA: MIT A minimalist account of the
Press. imperative clause. Journal of
———2004. Tense, case, and the nature Comparative Germanic Linguistics
of syntactic categories. In The 1: 177–224.
Syntax of Time, J. Guéron & J. Poletto, C. 1993. La Sintassi del
Lecarme (eds.) 495–537. Cambridge, Soggetto nei Dialetti Italiani
MA: MIT Press. Settentrionali. Padua: Unipress.
———2007. The syntax of valuation ———2000. The Higher Functional Field:
and the interpretability of features. Evidence from Northern Italian
Linguistics Today 101: 262–94. Dialects. Oxford: Oxford University
Petersen, C. 2011. The licensing of null Press.
subjects in Brazilian Portuguese. Polinsky, M. 2003. Non-canonical
MA thesis. University of Sao Paolo. agreement is canonical.
Petersen, C. & Terzi, A. 2015. Transactions of the Philological
Hyper‑Raising and locality: A view Society 101: 279–312.
from Brazilian Portuguese and Polinsky, M., Clemens, L. E., Morgan,
Greek. Chicago Linguistic Society A. M., Xiang, M. & Heestand, D.
Proceedings 50: 365–79. 2013. Resumption in English. In
Petronio, K. & Lillo-Martin, D. 1997. Experimental Syntax and Island
Wh Movement and the position of Effects, J. Sprouse & N. Hornstein
Spec‑CP: Evidence from American (eds.) 341–59. Cambridge:
Sign Language. Language 73: 18–57. Cambridge University Press.
520 References

Polinsky, M. & Potsdam, E. 2001. Language Studies, F. Dinneen

Long-distance agreement and topic (ed.) 177–206. Washington, DC:
in Tsez. Natural Language and Georgetown University Press.
Linguistic Theory 19: 583–646. ———1971. Cross-Over Phenomena. New
———2006. Expanding the scope of York: Holt, Rinehart and Winston.
control and raising. Syntax 9: ———1972. On some rules that are not
171–92. successive cyclic. Linguistic Inquiry
Pollock, J.-Y. 1989. Verb movement, 3: 211–22.
universal grammar, and the ———1974. On Raising. Cambridge, MA:
structure of IP. Linguistic Inquiry MIT Press.
20: 365–424. Postal, P. M., & Pullum, G. K. 1988.
Poortvliet, M. 2016. Copy Raising Expletive noun phrases in
in English, German and Dutch: subcategorised positions. Linguistic
Synchrony and Diachrony. Journal Inquiry 19: 425–49.
of Germanic Linguistics 28: Potsdam, E. 1998. Syntactic Issues in
370–402. the English Imperative. New York:
Portner, P. & Zanuttini, R. 2000. Garland.
The force of negation in wh ———2018. Syntactic Issues in the English
exclamatives and interrogatives. Imperative. London: Routledge.
In Studies in Negation and Potsdam, E. & Runner, J. T. 2001.
Polarity: Syntactic and semantic Richard returns: Copy‑raising and
perspectives, L. R. Horn & Y. Kato its implications. Chicago Linguistic
(eds.) 201–39. Oxford: Oxford Society Proceedings 37: 206–22.
University Press. Preminger, O. 2012. Agreement as a
Portner, P. 2004. The syntax of fallible operation. PhD diss. MIT.
imperatives within a theory of ———2014. Agreement and its Failures.
clause types. In Proceedings of the Cambridge, MA: MIT Press.
14th Semantics and Linguistic Prince, E. 1981a. Topicalization,
Theory Conference, K. Watanabe & Focus‑Movement, and
R. Young (eds.) 235–52. Ithaca, NY: Yiddish‑Movement: A pragmatic
Cornell University. differentiation. Berkeley Linguistic
———2007. Imperatives and modals. Society Proceedings 7: 249–64.
Natural Language Semantics 15: ———1981b. Toward a taxonomy of
351–83. given-new information. In Radical
Portner, P. & Yabushita, K. 1998. The Pragmatics, P. Cole (ed.) 222–55.
semantics and pragmatics of topic New York: Academic Press.
phrases. Linguistics and Philosophy ———1984. Topicalization and left
21: 117–57. dislocation: A functional analysis.
Postal, P. M. 1966. On so-called Annals of the New York Academy
pronouns in English. In Report on of Sciences 433: 213–25.
the Seventeenth Annual Round ———1990. Syntax and discourse: A
Table Meeting on Linguistics and look at resumptive pronouns.
 References 521

Berkeley Linguistics Society Studies in the Romance Verb, N.

Proceedings 16: 482–97. Vincent & M. Harris (eds.) 185–204.
———1995. On kind-sentences, London: Croom Helm.
resumptive pronouns, and relative ———1988. Transformational Grammar:
clauses. In Towards a Social A First Course. Cambridge:
Science of Language: A Festschrift Cambridge University Press.
for William Labov, G. Guy, J. ———1989. The Status of exclamative
Baugh & D. Schiffrin (eds.) 223–35. particles in French. In Essays
Cambridge: Cambridge University on Grammatical Theory and
Press. Universal Grammar, D. J. Arnold,
———1997 On the functions of left- M. Atkinson, J. Durand, C. Grover
dislocation in English discourse. & L. Sadler (eds.) 223–84. Oxford:
In Directions in Functional Oxford University Press.
Linguistics, A. Kamio (ed.) 117–43. ———1997a. Syntactic Theory and the
Amsterdam: Benjamins. Structure of English. Cambridge:
Puskás, G. 2000. Word Order in Cambridge University Press.
Hungarian: The Syntax of ———1997b. Syntax: A minimalist
Ā-positions. Amsterdam: Benjamins. introduction. Cambridge:
———2003. Floating quantifiers: What Cambridge University Press.
they can tell us about the syntax ———1997c. Verso un’analisi delle frasi
and semantics of quantifiers. esclamative in italiano (= ‘Towards
Generative Grammar in Geneva 3: an analysis of clausal exclamatives
105–28. in Italian’). In La Linguistica
Quirk, R. S. 1957. Relative clauses in Italiana Fuori d’Italia, L. Renzi
educated spoken English. English & M. Cortelazzo (eds.) 93–123.
Studies 38: 97–109. Rome: Bulzoni, Società Linguistica
Quirk, R., Greenbaum, S., Leech, G. & Italiana.
Svartik, J. 1985. A Comprehensive ———2004a. Minimalist Syntax:
Grammar of the English Language. Exploring the structure of English.
London: Longman. Cambridge: Cambridge University
Rackowski, A. & Richards, N. 2005. Press.
Phase edge and extraction: A ———2004b. English Syntax: An
Tagalog case study. Linguistic introduction. Cambridge:
Inquiry 36: 565–99. Cambridge University Press.
Radford, A. 1980. On English ———2008. Feature correlations in
exclamatives. Ms. University nominative case-marking by L1
College of North Wales, Bangor. and L2 learners of English. In The
———1981. Transformational Syntax. Role of Formal Features in Second
Cambridge: Cambridge University Language Acquisition, J. M. Liceras,
Press. H. Zobl & H. Goodluck (eds.)
———1982. The syntax of verbal 81–102. Mahwah, NJ: Laurence
wh‑exclamatives in Italian. In Erlbaum Associates.
522 References

———2009a. Analysing English in Romance Linguistics, O. Jaeggli

Sentences. Cambridge: Cambridge & C. Silva-Corvalan (eds.) 373–90.
University Press. Dordrecht: Foris.
———2009b. An Introduction to English ———1987. Case theory and Infl-to-
Sentence Structure. Cambridge: Comp: The inflected infinitive in
Cambridge University Press. European Portuguese. Linguistic
———2013. The complementiser system Inquiry 18: 85–109.
in spoken English: Evidence from Rawlins, K. 2008. (Un)conditionals:
broadcast media. In Agreement, An investigation in the syntax and
Information Structure and the CP, semantics of conditional structures.
V. Camacho-Taboada, Á Jiménez- PhD diss. University of California,
Fernández, J. Martín-González, Santa Cruz.
& M. Reyes-Tejedor (eds.) 11–54. Reed, A. M. 2010. The discourse
Amsterdam: Benjamins. function of floated quantifiers.
———2016. Analysing English Sentences, Journal of Pragmatics 42: 1737–61.
2nd edition. Cambridge: Cambridge Reed, L. 2011. Get-passives. The
University Press. Linguistic Review 28: 41–78.
———2018. Colloquial English: Structure Reich, I. 2017. On the omission of
and variation. Cambridge: articles and copulae in German
Cambridge University Press. newspaper headlines. Linguistic
———2019. Relative Clauses: Structure Variation 17: 186–206.
and variation in everyday English. Reid, J. 1997. Relatives and their
Cambridge: Cambridge University relatives: Relative clauses in
Press. conversational Australian English.
———2020. An Introduction to English PhD diss. La Trobe University.
Sentence Structure, 2nd edition. Reiman, P. W. 1994. Subjectless
Cambridge: Cambridge University sentences in English. UTA Working
Press. Papers in Linguistics 1: 141–52.
Radford, A. & Iwasaki, E. 2015. On Reinhart, T. 1981. Pragmatics and
swiping in English. Natural Language linguistics: An analysis of sentence
and Linguistic Theory 33: 703–44. topics. Philosophica 27: 53–94.
Radford, A., Felser, C & Boxell, O. 2012. Rett, J. 2012. Exclamatives, degrees
Preposition copying and pruning and speech acts. Linguistics and
in present-day English. English Philosophy 34: 411–42.
Language and Linguistics 16: Rezak, M. 2004. Elements of cyclic
403–26. syntax: Agree and Merge. PhD diss.
Ramat, P. 2014. Categories, features University of Toronto.
and values in the definition of ———2006. The interaction of Th/Ex
a word class. Italian Linguistic and Locative Inversion. Linguistic
Studies 52, 2. Inquiry 37: 685–97.
Raposo, E. 1986. On the null object in ———2010. φ-agree versus φ-feature
European Portuguese. In Studies movement: Evidence from floating
 References 523

quantifiers. Linguistic Inquiry 41: exploration of tense, tenseless

496–508. languages and tenseless
———2011. Building and interpreting constructions. Natural Language &
nonthematic A‑positions: Linguistic Theory 32: 1331–86.
A‑resumption in English and Rivero, M.-L. 1978. Topicalization
Breton. In Resumptive Pronouns and wh‑movement in Spanish.
at the Interfaces, A. Rouveret (ed.) Linguistic Inquiry 9: 513–7.
241–86. Amsterdam: Benjamins. ———1980. On left-dislocation and
Richards, N. 1999. Dependency topicalization in Spanish. Linguistic
formation and directionality of tree Inquiry 11: 363–93.
construction. MIT Working Papers ———1989. Barriers and Rumanian. In
in Linguistics 34: 67–105. Studies in Romance Linguistics. C.
———2003. Why is there an EPP? Gengo Kirschner & J. A. DeCesaris (eds.)
Kenkyu 123: 221–56. 289–312. Amsterdam: Benjamins.
———2010. Uttering Trees. Cambridge, Rizzi, L. 1982. Issues in Italian Syntax.
MA: MIT Press. Dordrecht: Foris.
———2016. Continuity Theory. ———1986. Null objects in Italian and
Cambridge, MA: MIT Press. the theory of pro. Linguistic Inquiry
Riemsdijk, H. van 1985. On pied-piped 17: 501–57.
infinitives in German relative ———1990. Relativised Minimality.
clauses. In Studies in German Cambridge, MA: MIT Press.
Grammar, J. Toman (ed.) 165–92. ———1994. Early null subjects and
Dordrecht: Foris. root null subjects. In Language
———2006. Free relatives: A syntactic Acquisition Studies in Generative
case study. In The Blackwell Grammar, T. Hoekstra &
Companion to Syntax, M. Everaert B. Schwartz (eds.) 151–76.
& H. van Riemsdijk (eds.) 336–78. Amsterdam: Benjamins.
Oxford: Blackwell. ———1996. Residual verb-second and
Riemsdijk, H. van & F. Zwarts. 1974. the wh‑criterion. In Parameters and
Left dislocation in Dutch and the Functional Heads, A. Belletti & L.
status of copying rules. Ms. MIT Rizzi (eds.) 63–90. Oxford: Oxford
and University of Amsterdam. University Press.
Reprinted as van Riemsdijk & ———1997. The fine structure of the left
Zwarts (1997). periphery. In Elements of Grammar,
———1997. Left dislocation in Dutch L. Haegeman (ed.) 281–337.
and the status of copying rules. Dordrecht: Kluwer.
In Materials on Left Dislocation, ———1998. Remarks on early null
E. Anagnastopoulou, H. van subjects. Proceedings of the
Riemsdijk & F. Zwarts (eds.) 13–30. Annual Boston University
Amsterdam: Benjamins. Conference on Language
Ritter, E. & Wiltschko, M. 2014. Development 23: 14–38. Reprinted
The composition of INFL: An as Rizzi 2000a.
524 References

———2000a. Remarks on early null Handout for talk presented to

subjects. In The Acquisition of EALing 2006. ENS, Paris.
Syntax, M.-A. Freidemann & ———2006c. Grammatically-based
L. Rizzi (eds.) 269–92. London: target-inconsistencies in child
Longman. language. University of Connecticut
———2000b. Comparative Syntax and Occasional Papers in Linguistics 4:
Language Acquisition. London: 19–49.
Routledge. ———2007. On some properties of
———2001. On the position criterial freezing. CISCL Working
‘Int(errogative)’ in the left periphery Papers on Language and Cognition
of the clause. In Current Issues 1: 145–58.
in Italian Syntax, G. Cinque & G. ———2010. On some properties
Salvi (eds.) 287–96. Amsterdam: of criterial freezing. In The
Elsevier. Complementizer Phase: Subjects
———2003. Relativised minimality and Operators, E. P. Panagiotidis
effects. In The Handbook of (ed.) 17–32. Oxford: Oxford
Contemporary Syntactic Theory, M. University Press.
Baltin & C. Collins (eds.) 89–110. ———2011. Syntactic cartography
Oxford: Blackwell. and the syntacticisation of
———2004a. Locality and Left Periphery. scope-discourse semantics. Ms.
In The Cartography of Syntactic Universities of Siena and Geneva.
Structures, Vol. 3: Structures and ———2012. Delimitation effects and the
beyond, A. Belletti (ed.) 223–51. cartography of the left periphery.
Oxford: Oxford University Press. In Discourse and Grammar: From
———2004b (ed). The Cartography of sentence types to lexical categories,
Syntactic Structures, Vol. 2: The G. Grewendorf & T. E. Zimmermann
structure of IP and CP, Oxford: (eds.) 115–45. Berlin: De Gruyter
Oxford University Press. Mouton.
———2005. On some properties of ———2013a. Notes on cartography and
subjects and topics. In Proceedings further explanation. Probus 25:
of the XXX Incontro di Grammatica 197–226.
Generativa, L. Brugé, G. Giusti, ———2013b. A Note on locality and
N. Munaro, W. Schweikert & G. selection. In Deep Insights, Broad
Turano (eds.) 203–24. Venice: Perspectives: Essays in honor of
Cafoscarina. Mamoru Saito, Y. Miyamoto, D.
———2006a. On the form of chains: Takahashi & H. Maki (eds.) 325–41.
Criterial positions and ECP effects. Tokyo: Kaitakusha.
In Wh‑Movement: Moving On, L. ———2013c. Topic, focus and the
Cheng & N. Corver (eds.) 97–133. cartography of the left periphery.
Cambridge, MA: MIT Press. In The Bloomsbury Companion
———2006b. On intermediate positions: to Syntax, S. Luraghi & C. Parodi
Intervention and impenetrability. (eds.) 436–51. London: Bloomsbury.
 References 525

———2014a. Some consequences Sound to Structures: Beyond the

of criterial freezing. In The veil of Maya, R. Petrosino & P.
Cartography of Syntactic Cerrone (eds.) 510–529. Berlin:
Structures, Vol. 9: Functional Mouton de Gruyter.
structure from top to toe, P. ———2020. Notes on the map of the
Svenonius (ed.) 19–45. Oxford: left periphery in Danish. In The
Oxford University Press. Sign of the V: Papers in honour
———2014b. The Cartography of of Sten Vikner, K. R. Christensen,
syntactic structures: Locality and H. Jørgensen & J. L. Wood (eds.)
freezing effects on movement. In 489–501. Aarhus: Department of
On Peripheries: Exploring clause English, School of Communication
initial and clause final positions C. and Culture, Aarhus University.
Cardinaletti, G. Cinque, & Y. Endo Rizzi, L & Bocci, G. 2017. Left
(eds.) 29–59. Tokyo: Hituzi Syobe periphery of the clause: Primarily
Publishing. illustrated for Italian. In The Wiley-
———2015a. Cartography, criteria and Blackwell Companion to Syntax,
labelling. In The Cartography of 2nd edition, M. Everaert & H. van
Syntactic Structures, Vol. 10: Riemsdijk (eds.) 1–30. Hoboken, NJ:
Beyond functional sequence, U. John Wiley & Sons.
Shlonsky (ed.) 314–38. Oxford: Rizzi, L. & Cinque, G. 2016. Functional
Oxford University Press. categories and syntactic theory.
———2015b. Notes on labelling and Annual Review of Linguistics 2:
subject positions. In Structures, 139–63.
Strategies and Beyond: Studies Rizzi, L. & Shlonsky, U. 2006.
in honour of Adriana Belletti, E. Satisfying the subject criterion by
Di Domenico, C. Hamann & S. a non-subject: Locative inversion
Matterini (eds.) 17–46. Amsterdam: and heavy NP shift. In Phases
Benjamins. of Interpretation, M. Frascarelli
———2016a. Uniqueness of left (ed.) 341–61. Berlin: Mouton de
peripheral focus: Further Gruyter.
explanation and introduction. In ———2007. Strategies of
Order and Structure in Syntax, L. R. subject extraction. In
Bailey & M. Sheehan (eds.) 333–43. Interfaces+Recursion = Language?
Language Science Press. U. Sauerland & H.-M. Gärtner (eds.)
———2016b. Labeling, maximality and 115–60. Berlin: Mouton de Gruyter.
the head–phrase distinction. The Roberge, Y. 1991. The recoverability
Linguistic Review 33: 103–27. of null objects. In New Analyses
———2017. Types of criterial freezing. in Romance Linguistics: Selected
Rivista di Grammatica Generativa papers from the XVIIIth linguistic
39: 1–19. symposium on Romance languages.
———2018. Subjects, topics and the D. Wanner & D. A. Kibbee (eds.)
interpretation of pro. In From 299–312. Amsterdam: Benjamins.
526 References

Roberts, I. 1987. The Representation of Riemsdijk & F. Zwarts (eds.) 31–55.

Implicit and Dethematized Subjects. Amsterdam: Benjamins.
Dordrecht: Foris. Rodriguez, C. 2004. Impoverished
———1993a. Verbs and Diachronic morphology and A‑movement
Syntax. Dordrecht: Kluwer. out of case domains. PhD diss.
———1993b. The nature of subject clitics University of Maryland.
in Franco-Provençal Valdotain. In Roehrs, D. 2008. High floating
Syntactic Theory and the Dialects quantifiers: Syntactic or ‘delayed’
of Italy, A. Belletti (ed.) 319–53. V2? Snippets 17: 7–8.
Turin: Rosenberg & Sellier. Rogers, A. D. 1971. Three kinds of
———2004. The C‑system in Brythionic physical perception verbs. Chicago
Celtic languages and the EPP. Linguistic Society Proceedings 7:
In The Cartography of Syntactic 206–23.
Structures, Vol. 2: The structure ———1972. Another look at flip
of IP and CP, L. Rizzi (ed.) 297– perception verbs. Chicago
328. Oxford: Oxford University Linguistic Society Proceedings 8:
Press. 303–15.
———2010. Agreement and Head ———1974. Physical perception verbs
Movement: Clitics, incorporation in English: A study in lexical
and defective goals. Cambridge, relatedness. PhD diss. UCLA.
MA: MIT Press. Rohrbacher, B. 1999. Morphology-
Roberts, I. & Roussou, A. 2002. The driven syntax: A theory of V-to-I
extended projection principle as a raising and pro-drop. Amsterdam:
condition on the tense dependency. Benjamins.
In Subjects, Expletives and the EPP, Romero, M. 1997. The correlation
P. Svenonius (ed.) 125–55. Oxford: between scope reconstruction and
Oxford University Press. connectivity effects. Proceedings
Rochemont, M. 1986. Focus in of the West Coast Conference in
Generative Grammar. Amsterdam: Formal Linguistics 16: 351–65.
Benjamins. ———1998. Focus and reconstruction
Rochman, L. 2010. Why float? Floating effects in wh‑phrases. PhD diss.
quantifiers and focus marking. In University of Massachusetts,
The Sound Patterns of Syntax, N. Amherst.
Erteschik-Shir & L. Rochman (eds.) Rooryck, J. 1994. Generalized
72–92. Oxford: Oxford University transformations and the wh‑cycle:
Press. Free relatives as bare wh‑CPs.
Rodman, R. 1974. On left dislocation. In Minimalism and Kayne’s
Papers in Linguistics 7: 437–66. Asymmetry Hypothesis. C. J.-W.
Reprinted as Rodman 1997. Zwart (ed.) 195–208. Groningen:
———1997. On left dislocation. In Groningen University.
Materials on Left Dislocation, ———2000. Configurations of Sentential
E. Anagnastopoulou, H. van Complementation: Perspectives
 References 527

from Romance Languages. London: Linguistic Variation Yearbook 2:

Routledge. 123–84.
Rooth, M. 1985. Association with ———2011 (ed.) Resumptive Pronouns
focus. PhD diss. University of at the Interfaces. Amsterdam:
Massachusetts, Amherst. Benjamins.
———1992. A theory of focus Ruda, M. 2014. Missing objects in
interpretation. Natural Language special registers: The syntax of
Semantics 1: 75–116. null objects in English. Canadian
———1996. Focus. In The Handbook of Journal of Linguistics 59: 339–72.
Contemporary Semantic Theory, ———2017. On the syntax of missing
S. Lappin (ed.) 272–97. Oxford: objects: A study with special
Blackwell. reference to English, Polish and
Rosen, S. T. 1990. Argument Structure Hungarian. Amsterdam: Benjamins.
and Complex Predicates. New York: Rudolph, R. E. 2019. A closer look at the
Garland. perceptual source in copy raising
Rosengren, I. 2000. Rethinking the constructions. Proceedings of Sinn
adjunct. ZAS Papers in Linguistics Und Bedeutung 23: 287–304.
17: 217–40. Rühlemann, C. 2012. Conversational
———2002. A syntactic device in Grammar. Wiley Online
the service of semantics. Studia Encyclopedia of Applied
Linguistica 56: 145–90. Linguistics. [Link]
Ross, J. R. 1967. Constraints on .org/10.1002/9781405198431
Variables in Syntax. PhD diss. MIT. .wbeal0220 .
Published as Ross 1986. Rupp, L. 1999. Aspects of the syntax
———1969. Guess who. Chicago Linguistic of English imperatives. PhD diss.
Society Proceedings 5: 252–86. University of Essex.
———1973. Ross, J. R. The same side ———2003. The Syntax of Imperatives
filter. Chicago Linguistic Society in English and Germanic: Word
Proceedings 9: 549–67. order variation in the minimalist
———1986 Infinite Syntax! Norwood, framework. New York: Palgrave
NJ: Ablex Publishing Corporation. Macmillan.
Rothstein, S. D. 1983. The syntactic ———2005. Grammatical constraints on
form of predication. PhD diss. MIT. nonstandard -s in expletive there
———1995. Pleonastics and the sentences. English Language and
interpretation of pronouns. Linguistics 9: 225–88.
Linguistic Inquiry 26: 499–529. Ruppenhofer, J. & Michaelis, L. A.
Roussou, A. 2000. On the left 2010. A constructional account of
periphery: Modal particles and genre-based argument omissions.
complementizers. Journal of Greek Constructions and Frames 2:
Linguistics 1: 65–94. 158–84.
Rouveret, A. 2002. How are resumptive Sabel, J. 1996. Restrukturierung
pronouns linked to the periphery? und Lokalität: Universelle
528 References

Beschränkungen für From sentence types to lexical

Wortstellungsvarianten. Berlin: categories, G. Grewendorf & T. E.
Akademie Verlag. Zimmermann (eds.) 147–76. Berlin:
———2000. Expletives as features. de Gruyter.
Proceedings of the West Coast Sakamoto, A. & Ikarashi, K.
Conference on Formal Linguistics 2017. Subjectless sentences in
19: 411–24. conversation and the defectiveness
———2002. A minimalist analysis of of their syntactic structures. In
syntactic islands. The Linguistic Studies on Syntactic Cartography,
Review 19: 271–315. Fuzhen Si (ed.) 336–51. Beijing:
Sadock, J. M. 1974. Read at your China Social Sciences Press.
own risk: Syntactic and semantic Salvi, G. 2005. Some firm points
horrors you can find in your own on Latin word order: The left
medicine chest. Chicago Linguistic periphery. In Universal Grammar
Society Proceedings 10: 599–607. and the Reconstruction of Ancient
Sæbø, K. 2006. Explaining clausal Languages, K. É. Kiss (ed.) 429–56.
exclamatives. 3ième Journée de Berlin: Mouton de Gruyter.
Sémantique et Modélisation. Paris: Salzmann, M. 2006. Resumptive
ENS. pronouns and matching effects in
———2010. On the semantics of Zurich German relative clauses as
‘embedded exclamatives’. Studia distributed deletion. Leiden Papers
Linguistica 64: 116–40. in Linguistics 3: 17–50.
Safir, K. 1986a. Syntactic Chains. ———2011. Reconstruction and
Cambridge: Cambridge University Resumption in Indirect A′-
Press. dependencies: On the syntax of
———1986b. Relative clauses in a theory prolepsis and relativization in
of binding and levels. Linguistic (Swiss) German and beyond. Berlin:
Inquiry 17: 663–90. Mouton de Gruyter.
———1999. Vehicle change and Samek-Lodovici, V. [Link]
reconstruction in A‑bar chains. right dislocation meets the left-
Linguistic Inquiry 30: 587–620. periphery: A unified analysis of
Sag, I. 1978. Floating quantifiers, Italian non-final focus. Lingua 116:
adverbs and extraction sites. 836–73.
Linguistic Inquiry 9: 146–50. ———2009. Topic, focus and background
———1980. Deletion and Logical Form. in Italian clauses. In Focus and
New York: Garland. Background in Romance Languages,
———1997. English relative clause A. Dufter & D. Jacob (eds.) 333–57.
constructions. Journal of Amsterdam: Benjamins.
Linguistics 33: 431–83. Sato, Y. & Dobashi, Y. 2013. Functional
Saito, M. 2012. Sentence types and categories and prosodic phrasing
the Japanese right periphery. in English: Evidence from that-
In Discourse and Grammar: trace effects and pronominal object
 References 529

shift. Ms. National University of Schneider-Zioga, P. 2007. Anti-

Singapore and Niigata University. agreement, anti-locality, and
Sato, Y. & Kim, C. 2012. Radical pro minimality: The syntax of dislocated
drop and the role of syntactic subjects. Natural Language and
agreement in Colloquial Singapore Linguistic Theory 25: 403–46.
English. Lingua 122: 858–73. Schönenberger, M. 2010. Optional
Sauerland, U. 1998. The meaning of doubly-filled COMPs (DFCs) in
chains. PhD diss. MIT. wh‑complements in child and adult
———1999. Two structures for English Swiss German. In Variation in the
restrictive relative clauses. In Input, M. Anderssen, K. Bentzen
Proceedings of the Nanzan GLOW, & M. Westergaard (eds.) 33–64.
M. Saito (ed.) 351−66. Nagoya: Dordrecht: Springer.
Nanzan University. Schütze, C. 1996 The Empirical Basis
———2003. Unpronounced heads in of Linguistics: Grammaticality
relative clauses. In The Interfaces: judgments and linguistic
Deriving and interpreting omitted methodology. Chicago: University
structures, K. Schwabe & S. of Chicago Press.
Winkler (eds.) 205–26. Amsterdam: ———1999. English expletive
Benjamins. constructions are not infected.
Sauerland, U. & Elbourne, P. 2002. Linguistic Inquiry 30: 467–84.
Total reconstruction, PF movement ———2001. On the nature of default
and derivational order. Linguistic case. Syntax 4: 205–38.
Inquiry 33: 283–319. ———2009. Web searches should
Sawada, H. 1995. Studies in English supplement judgements, not
and Japanese Auxiliaries: A multi- supplant them. Zeitschrift für
stratal approach. Tokyo: Hituzi Sprachwissenschaft 28: 151–56.
Syobo. Schütze, C. & Sprouse, J. 2014.
Schachter, P. 1973. Focus and Judgment data. In Research Methods
relativization. Language 49: 19–46. in Linguistics, R. J. Podesva & D.
Schaeffer, J. 1994. The split INFL Sharma (eds.) 27–50. Cambridge:
hypothesis: Evidence from impaired Cambridge University Press.
language. First Language 14: Schwarz, B. 1999. On the syntax of
334–35. either … or. Natural Language and
Schirer, K. C. 2008. Have you heard the Linguistic Theory 17: 339–70.
new theory? A syntactic analysis ———2000. Topics in Ellipsis Amherst,
of null subjects and null auxiliaries MA: GLSA Publications.
in English. MA thesis. University of Schwarz, F. 2012. Situation pronouns
Kansas. in determiner phrases. Natural
Schmerling, S. 1973. Subjectless Language Semantics 20: 431–75.
sentences and the notion of surface Schwenter, S. A. 2006. Null objects
structure. Chicago Linguistic across South America. In Selected
Society Proceedings 9: 577–86. Proceedings of the 8th Hispanic
530 References

Linguistics Symposium, T. L. Face & Shan, C.-C. & Barker, C. 2006.

C. A. Klee (eds.) 23–36. Somerville, Explaining crossover and
MA: Cascadilla Proceedings Project. superiority as left-to-right
Schwenter, S.A. & Silva, G. 2002. Overt evaluation. Linguistics and
vs. null direct objects in spoken Philosophy 29: 91–134.
Brazilian Portuguese: A semantic/ Sharvit, Y. 1999. Resumptive pronouns
pragmatic account. Hispania 85: in relative clauses. Natural
577–86. Language and Linguistic Theory 17:
Scott, K. J. 2004. Diary drop in English: 587–612.
A syntactic and pragmatic analysis. Sheehan, M. 2006. The EPP and null
MA thesis. University College subjects in Romance. PhD diss.
London. University of Newcastle-upon-
———2010. The relevance of referring Tyne.
expressions: The case of diary drop ———2013a. The resuscitation of CED.
in English. PhD diss. University Proceedings of the North East
College London. Linguistic Society 40: 135–50.
Sells, P. 1984. Syntax and semantics ———2013b. Some implications of a
of resumptive pronouns. PhD copy theory of labelling. Syntax 16:
diss. University of Massachusetts, 362–96.
Amherst. Shibata, K. 2017. Subject ellipsis in
———1987. Binding resumptive English: The perception verbs feel,
pronouns. Linguistics and look, sound, smell, and taste. PhD
Philosophy 10: 261–98. diss. Kyoto Prefectural University.
Seppänen, A. 1997: Relative that and Shlonsky, U. 1991. Quantifiers as
prepositional complementation. functional heads: A study of
English Language and Linguistics quantifier float in Hebrew. Lingua
1: 111–33. 84: 159–80.
Seppänen, A. & Kjellmer, G. 1995. The ———1992. Resumptive pronouns as a
dog that’s leg was broken: On the last resort. Linguistic Inquiry 23:
genitive of the relative pronoun. 443–68.
English Studies 76: 389–400. ———1993. Agreement in COMP. The
Seppänen, A. & Trotta, J. 2000. The Linguistic Review 11: 351–75.
wh+that pattern in present- ———1997. Clause Structure and Word
day English. In Corpora Galore: Order in Hebrew and Arabic: An
Analyses and Techniques in essay in comparative Semitic
Describing English, J. M. Kirk (ed.) syntax. Oxford: Oxford University
161–75. Amsterdam: Rodopi. Press.
Shaer, B. & Frey, W. 2004. Integrated ———2010. The Cartographic enterprise
and non-integrated left-peripheral in syntax. Language and Linguistics
elements in German and English. Compass 4: 417–39.
ZAS Papers in Linguistics 35: ———2014. Subject positions, subject
465–502. extraction, EPP, and the Subject
 References 531

Criterion. In Locality, E. Aboh, M. Aristotle University, vol 1, 188–98.

T. Guasti & I. Roberts (eds.) 58–85. Thessaloniki: Monochromia.
Oxford: Oxford University Press. Sigley, R. 1997. The influence of
Shlonsky, U. & Soare, G. 2011. Where’s formality and channel on relative
‘why’? Linguistic Inquiry 42: pronoun choice in New Zealand
651–69. English. English Language and
Shlonsky, U. & Bocci, G. 2019. Linguistics 1: 207–32.
Syntactic cartography. Oxford Sigurðsson, H. Á. 2010. On EPP effects.
Research Encyclopedia of Studia Linguistica 64: 159–89.
Linguistics. [Link] Sigurðsson, H. Á. & Maling, J.
acrefore/9780199384655.013.310 2007. On null arguments. In
Shnukal, A. 1981. There’s a lot Proceedings of the XXXII Incontro
mightn’t believe this … Variable di Grammatica Generativa, M.
subject relative pronoun absence C. Picci & A. Pona (eds.) 167–80.
in Australian English. In Variation Alessandria: Edizioni dell’Orso.
Omnibus, D. Sankoff & H. ———2010. The Empty Left Edge
Cedergren (eds.) 321–8. Carbondale, Condition. In Exploring Crash-
IL: Linguistic Research. Proof Grammars, M. Putnam (ed.)
Shormani, M. 2021. Imperatives in 57–84. Amsterdam: Benjamins.
Arabic: Syntax, discourse and Simonin, O. 2012. Adverbial and
interface. International Journal of relative to‑infinitives. Journal of
Arabic Linguistics 7: 19–50. English Linguistics 41: 4–32.
Shorrocks, G. 1982. Relative pronouns Simpson, A. 2011. Floating Quantifiers
and relative clauses in the dialect in Burmese and Thai. Journal of
of Farnworth and district (Greater the Southeast Asian Linguistics
Manchester County, formerly Society, 4: 115–46.
Lancashire). Zeltschrift für Simon-Vandenbergen, A. M. 1981.
Dialektologie und Linguistik 49: The Grammar of Headlines in The
334–43. Times 1870–1970. Brussels: Paleis
Sichel, I. 2014. Resumptive pronouns der Academiën.
and competition. Linguistic Inquiry Šímová, P. 2005. The transition
45: 655–93. between restrictive and
Siemund, P. 2015. Exclamative clauses nonrestrictive adnominal relative
in English and their relevance for clauses. PhD diss. Charles
theories of clause types. Studies in University, Prague.
Language 39: 698–728. Sistrunk, W. 2012. The syntax of
Sifaki, E. & Sitaridou, I. 2007. EPP zero in African American relative
revisited: Evidence from null clauses. PhD diss. Michigan State
subject languages. In Selected University.
Papers in Theoretical and Applied Smit, J. 2014. An investigation into the
Linguistics: 17th International adequacy of Cinque’s functional
Symposium, English Department, theory as a framework for the
532 References

analysis of adverbs in Afrikaans. ———2016. The halting problem.

MA thesis. Stellenbosch University. Proceedings of the Linguistic
Smith, P. W. 2015. Feature mismatches: Society of America 1.3: 1–10.
Consequences for syntax, Sorace, A. 2000. Gradients in auxiliary
morphology and semantics. PhD selection with intransitive verbs.
diss. University of Connecticut. Language 76: 859–90.
———2017. The syntax of semantic Speas, M. 1986. Adjunction and
agreement in English. Journal of projections in syntax. PhD diss.
Linguistics 53: 823–63. MIT.
———2020. Object agreement and ———1994. Null arguments in a
grammatical functions: A re- theory of economy of projection.
evaluation. In Agree to Agree: University of Massachusetts
Agreement in the minimalist Occasional Papers in Linguistics
program, P. W. Smith, J. Mursell & 17: 179–208.
K. Hartmann (eds.) 117–47. Berlin: Speas, P. & Tenny, C. 2003.
Language Science Press. Configurational properties of point
Smith, P. W., Mursell, J. & Hartmann, of view roles. In Asymmetry in
K. (2020). Some remarks on Grammar, A. M. di Sciullo (ed.)
agreement within the minimalist 315–45. Amsterdam: Benjamins.
program. In Agree to Agree: Spector, I. 2008. Hebrew floating
Agreement in the minimalist quantifiers: A non-derivational
program, P. W. Smith, J. Mursell & approach. MA thesis. Hebrew
K. Hartmann (eds.) 1–29. Berlin: University of Jerusalem.
Language Science Press. ———2009. Hebrew floating quantifiers.
Sobin, N. 1987. The variable status of In Proceedings of the LFG09
Comp‑trace phenomena. Natural Conference, M. Butt & T. H. King
Language and Linguistic Theory 5: (eds.) 520–40. Stanford, CA: CSLI
33–60. Publications.
———1997. Agreement, default rules Spencer, A. 1991. Morphological
and grammatical viruses. Linguistic Theory. Oxford: Blackwell.
Inquiry 28: 318–43. Spinillo, M. G. 2004. Reconceptualising
———2002. The Comp‑trace effect, the the English determiner class. PhD
adverb effect and minimal CP. diss. University College London.
Journal of Linguistics 38: 527–60. Sportiche, D. 1986. A theory of floating
———2003. Negative Inversion as quantifiers. Proceedings of the North
non‑movement. Syntax 6: 183–212. East Linguistics Society 17: 581–94.
———2004. Expletive constructions are ———1988. A theory of floating
not ‘lower right corner’ movement quantifiers and its corollaries for
constructions. Linguistic Inquiry constituent structure. Linguistic
35: 503–08. Inquiry 19: 425–49.
———2009. Prestige case forms and the ———2008. Inward bound: Splitting
Comp‑trace effect. Syntax 12: 32–59. the wh‑paradigm and French
 References 533

Relative qui. [Link] using a random sample from

lingbuzz/000623 Linguistic Inquiry 2001–2010.
———2015. Relative clauses: Head Lingua 134: 219–48.
raising only. [Link] Starke, M. 2001. Move dissolves into
.it/ad60/doc/sportiche_abs.pdf Merge: A theory of locality. PhD
———2017. Relative clauses. Ms. diss. University of Geneva.
UCLA. [Link] Stassen, L. 2013. Zero copula for
lingbuzz/003444 predicate nominals. In The World
Sprouse, J. 2011. A test of the cognitive Atlas of Language Structures
assumptions of magnitude Online, M. Dryer & M. Haspelmath
estimation: Commutativity does not (eds.) feature 120A. Leipzig: Max
hold for acceptability judgments. Plank Institute for Evolutionary
Language 87: 274–88. Anthropology.
Sprouse, J. & Almeida, D. 2011a. Stepanov, A. 2001. Late adjunction and
Power in acceptability judgment minimalist phrase structure. Syntax
experiments and the reliability of 4: 94–125.
data in syntax. Ms. University of ———2007. The end of CED? Minimalism
California, Irvine and Michigan and extraction domains. Linguistic
State University. Inquiry 10: 80–126.
———2011b. A formal experimental Stockwell, R., Schachter, P. & Partee,
investigation of the empirical B. 1973. The Major Syntactic
foundation of generative syntactic Structures of English. New York:
theory. Ms. University of California, Holt Rinehart and Winston.
Irvine and Michigan State University. Stowell, T. 1981. Origins of phrase
———2012. Assessing the reliability of structure. PhD diss. MIT.
textbook data in syntax: Adger’s ———1990. C‑command effects found in
Core Syntax. Journal of Linguistics newspaper headlines. Ms. UCLA.
48: 609–52. ———1991. Abbreviated English. Glow
———2013. The empirical status of data Newsletter, March.
in syntax: A reply to Gibson and ———1996. Empty heads in abbreviated
Fedorenko. Language and Cognitive English. Ms. UCLA.
Processes. 28: 229–40. ———1999. Words lost and syntax
Sprouse, J., Caponigro, I., Greco, C. & found in Headlinese: The hidden
Cecchetto, C. 2013. Experimental structure of abbreviated English in
syntax and the crosslinguistic headlines, instructions and diaries.
variation of island effects in Paper presented at York University,
English and Italian. Ms. Universities Toronto.
of California, Irvine and San Diego, Stradmann, M. 2013. Reviewing the
and Milan-Bicocca. Binary Branching Hypothesis. BA
Sprouse, J. Schütze, C. & Almeida, D. thesis. Utrecht University.
2013. A comparison of informal Straumann, H. 1935. Newspaper
and formal acceptability judgments Headlines: A study in linguistic
534 References

method. London: George Allen & Littoral, P. Poussa (ed.) 147–65.

Unwin Ltd. Munich: Lincom Europa.
Suárez-Gómez, C. 2014. Relative Taglicht, J. 1984. Message and
clauses in Southeast Asian Emphasis: On focus and scope in
Englishes. Journal of English English. London: Longman.
Linguistics 42: 245–68. Takahashi, H. 2004. The English
Suñer, M. 1998. Resumptive imperative: A cognitive and
restrictive relatives: A functional analysis. PhD diss.
crosslinguistic perspective. Hokkaido University.
Language 74: 335–64. Takano, Y. 2013. Movement of
Suñer, M. & Yépez, M. 1988. Null antecedents and minimality.
definite objects in Quiteño. Nanzan Linguistics 9: 193–214.
Linguistic Inquiry 19: 511–19. Tanaka, T. 2008. Aspectual restriction
Svenonius, P. 2002a (ed.) Subjects, for floating quantifiers. Nanzan
Expletives and the EPP. Oxford: Linguistics 3: 199–228.
Oxford University Press. Tateishi, K. 1995. The Syntax of
———2002b. Subject positions and ‘Subjects’. Cambridge: Cambridge
the placement of adverbials. In University Press.
Subjects, Expletives and the EPP, Teddiman, L & Newman, J. 2010.
P. Svenonius (ed.) 199–40. Oxford: Subject ellipsis in English:
Oxford University Press. Construction of and findings from
———2017. Syntactic features. Ms. a diary corpus. In Handbook of
CASTL, University of Tromsø. Research on Discourse Behaviour
Szczegielniak, A. 1999. That-trace and Digital Communications:
effects cross-linguistically and Language structures and social
successive-cyclic movement. MIT interaction, R. Taiwo (ed.) 281–95.
Working Papers in Linguistics 33: Hershey, PA: IGI Global.
1–26. Tel, A. 2020. Article omission in
———2005. Two types of resumptive headlines: A comparison between
pronouns in Polish relative clauses. the physical and digital version
Linguistic Variation Yearbook 5: of a newspaper. BA thesis. Leiden
165–85. University.
Szendröi, K. 2001. Focus and the Terada, H. 2000. Multiple expletive
syntax-phonology interface. PhD construction. English Linguistics
diss. University College London. 17: 1–25.
Tagg, C. 2009. A corpus linguistics ———2003. Floating quantifiers as
study of SMS text messaging. PhD probes. English Linguistics 20:
diss., University of Birmingham. 467–92.
Tagliamonte, S. 2002. Variation and Thornton, R. 1995. Referentiality and
change in the British relative Wh‑Movement in Child English:
marker system. In Dialect Contact Juvenile D‑Linkuency. Language
and History on the North Sea Acquisition 4: 139–75.
 References 535

Thrasher, R. 1973 A conspiracy on the 45–52. [Link]

far left. University of Michigan elsjregional.9.0_45
Papers in Linguistics l: 169–79. Tottie, G. 1997a. Relatively speaking:
———1974. Shouldn’t ignore these Relative marker usage in the British
strings: A study of conversational National Corpus. In To Explain the
deletion. PhD diss. University of Present: Studies in the changing
Michigan. English language in honour of
———1977 One Way to Say More by Matti Rissanen, T. Nevalainen &
Saying Less: A study of so-called L. Kahlas-Tarkka (eds.) 465–81.
subjectless sentences. Kwansei Helsinki: Société Néophilologique.
Gakuin University Monograph ———1997b. Overseas relatives: British-
Series Vol. 11. Tokyo: Eihosha Ltd. American differences in relative
Ticio, M. E. 2003. On the structure marker usage. In Studies in English
of DPs. PhD diss. University of Language Research and Teaching,
Connecticut. J. Aarts & H. Wekker (eds.) 153–65.
———2005. Locality and anti-locality in Amsterdam: Rodopi.
Spanish DPs. Syntax 8: 229–86. Trotzke, A. & Giannakidou, A. 2019.
Tonoike, S. 1996. A note on floating Exclamatives as emotive assertions
quantifiers. Metropolitan of intensity. [Link]
Linguistics 16: 1–9. lingbuzz/004838
Torrego, E. 1996. On quantifier float in Trudgill, P., Nevalainen, T. & Wischer,
control clauses. Linguistic Inquiry I. 2002. Dynamic have in North
27: 111–26. American and British Isles English.
Torrence, H. 2013. The Clause English Language and Linguistics
Structure of Wolof: Insights into 6: 1–15.
the left periphery. Amsterdam: Truswell, R. 2007. Extraction from
Benjamins. adjuncts and the structure of
Torres Cacoullos, R. & Travis, C. events. Lingua 117: 1355–77.
E. 2014. Prosody, priming and ———2009. Preposition stranding,
particular constructions: The passivisation, and extraction from
patterning of English first-person adjuncts in Germanic. Linguistic
singular subject expression in Variation Yearbook 8: 131–77.
conversation. Journal of Pragmatics ———2011. Events, Phrases, and
63: 19–34. Questions. Oxford: Oxford
Totsuka, M. 2015. On phasehood of University Press.
the functional categories in the Tsai, W.-T. D. 2008. Left periphery and
left periphery. PhD diss. Tohoku why-how alternations. Journal of
University, Sendai. East Asian Linguistics 17: 83–115.
———2017. On ellipsis in the left Tsoulas, G. 2003. Floating quantifiers
periphery: A view from transfer. as overt scope markers. Korean
English Literary Society of Journal of English Language and
Japan Regional Proceedings 9: Linguistics 3: 157–80.
536 References

Uchibori, A. 2000. The syntax of Villa-García, J. 2010.

subjunctive complements: Evidence Recomplementation and locality
from Japanese. PhD diss. University of movement in Spanish. Second
of Connecticut. general examination paper,
Ura, H. 1994. Varieties of Raising and University of Connecticut, Storrs.
the feature-based phrase structure Published as Villa-García 2012c.
theory. MIT Occasional Papers in ———2011a. On COMP‑t effects in
Linguistics, no. 7. Spanish: A new argument for
———1995. Multiple feature checking. rescue by PF deletion. Paper
PhD diss. MIT. presented to GLOW 34, University
Ussery, C. 2009. Optionality and of Vienna.
variability: Syntactic licensing ———2011b. On the Spanish clausal
meets morphological spellout. PhD leftedge: In defence of a TopicP
diss. University of Massachusetts, account of recomplementation.
Amherst, MA. Paper presented to 41st Linguistic
———2012. Case and phi-features as Symposium on Romance
probes. Proceedings of the West Languages, University of Ottawa.
Coast Conference on Formal ———2012a Characterizing medial and
Linguistics 29: 127–35. low complementizers in Spanish:
Valmara, V. 2008. Topic, focus and Recomplementation que and
quantifier float. In Gramatika jussive/optative que. In Current
Jaietan, X. Artigoitia & J. A. Formal Aspects of Spanish Syntax
Lakarra (eds.) 837–57. Bilbao: and Semantics, M. González-Rivera
EHU. & S. Sessarego (eds.) 198–228.
Varlokosta, S. 1994. Issues on Modern Newcastle upon Tyne: Cambridge
Greek sentential complementation. Scholars.
PhD diss. University of Maryland. ———2012b The Spanish complementizer
Vat, J. 1981. Left dislocation, system: Consequences for the
connectedness and reconstruction. syntax of dislocations and subjects,
Groninger Arbeiten zur locality of movement, and clausal
Germanistischen Linguistik 20: structure. PhD diss. University of
80–103. Connecticut, Storrs.
Vater, H. 1980. Quantifier float in ———2012c Recomplementation and
German. In The Semantics of locality of movement in Spanish.
Determiners, J. van der Auwera Probus 24: 257–314.
(ed.) 232–49. London: Croom Helm. ———2015 The Syntax of Multiple-
Vergnaud, J. R. 1974. French relative que Sentences in Spanish: Along
clauses. PhD diss. MIT. the left periphery. Amsterdam:
Vikner, S. 1995. Verb Movement and Benjamins.
Expletive Subjects in Germanic Villalba, X. 2000. The Syntax of
Languages. Oxford: Oxford Sentence Periphery. Barcelona:
University Press. Servei de Publicacions.
 References 537

———2001. The right edge of ———1994. The role of triggers in the

exclamative sentences in Catalan. extended split INFL hypothesis:
Catalan Working Papers in Unlearnable parameter settings.
Linguistics 9: 119–35. Studia Linguistica 48: 156–78.
———2003. An exceptional exclamative Weir, A. 2008. Subject pronoun drop
sentence type in Romance. Lingua in informal English. MA thesis.
113: 713–45. University of Edinburgh.
———2008. Exclamatives: A thematic ———2009a. Subject pronoun drop in
guide with many questions and informal English. Ms. University
few answers. Catalan Journal of College London.
Linguistics 7: 9–40. ———2009b. Article drop in English
Vishenkova, A., & Zevakhina, N. 2019. headlinese. MA thesis. University
Wh‑exclamatives with and without College London.
predicates in Russian. Russian ———2012. Left-edge deletion in English
Linguistics 43: 107–25. and subject omission in diaries.
Visser, J. 1996. Object Drop in Dutch English Language and Linguistics
imperatives. In Linguistics in the 16: 105–09.
Netherlands 1996, C. Cremers ———2013. Article drop in headlines
& M. den Dikken (eds.) 257–68. and truncation of CP. Text of paper
Amsterdam: Benjamins. presented to LSA Annual Meeting,
Wagner, M. 2012. Focus and givenness: Boston, 3–6 January.
A unified approach. In Contrasts ———2014. Fragments and clausal
and Positions in Information ellipsis. PhD diss. University of
Structure, I. Kučerová & A. Massachusetts, Amherst.
Neeleman (eds.) 102–47. Cambridge: ———2016. The prosodic licensing of
Cambridge University Press. left-edge ellipsis and implications
Wagner, S. 2012. Null subjects in for clausal ellipsis. Presentation to
English: Variable rules, variable the Ellipsis and Prosody Workshop,
language? Post-doctoral diss. Leiden University.
Chemnitz University of Technology: ———2017. Object drop and article
———2016. Never saw one: First-person drop in reduced written register.
null subjects in spoken English. Linguistic Variation 17: 157–85.
English Language and Linguistics ———2018. Left-edge ellipsis and clausal
22: 1–34. ellipsis: The division of labor
Wal, J. van der 2016. Diagnosing focus. between syntax, semantics and
Studies in Language 40: 259–301. prosody. Talk at workshop on The
Ward, G. & E. Prince. 1991. On the Timing of Ellipsis, at the annual
topicalization of indefinite NPs. meeting of the Societas Linguistica
Journal of Pragmatics 15: 167–77. Europaea, Tallinn University.
Watanabe, A. 1993. Agr-based case ———2022. Fragments and left-edge
theory and its interaction with the ellipsis: The division of labour
A‑bar system. PhD diss. MIT. between syntax, semantics, and
538 References

prosody. In The Derivational ———1994b. Some properties of ellipsis

Timing of Ellipsis, G. Güneş & in coordination. Geneva Generative
A. Lipták (eds.) 253–90. Oxford: Papers 2: 23–61.
Oxford University Press. Wiljergård, E. 2011. The use of articles
Weisler, S. 1980. The syntax of that- in newspaper headlines: Omission
less relatives. Linguistic Inquiry 11: and retention in a British quality
624–31. paper and tabloid. BA thesis.
Weiß, H. 2005. Inflected complementisers Linnaeus University.
in Continental West Germanic Williams, E. 1984. There-insertion.
dialects. Zeitschrift für Dialektologie Linguistic Inquiry 15: 131–53.
und Linguistik 72: 148–66. ———2002. Adjunct modification. Italian
Wen, H. 2020. Relative clauses in Journal of Linguistics 12: 129–54.
Mandarin Chinese. PhD diss. Queen ———2006. The subject-predicate theory
Mary, University of London. of there. Linguistic Inquiry 37:
Weskott, T. & Fanselow, G. 2011. On the 648–51.
informativity of different measures Williams, P. 2013. 8 grammar rules
of linguistic acceptability. Language for writing newspaper headlines.
87: 249–73. [Link]
Wexler, K. & Culicover, P.W. 1980. Willim, E. 2012. On the feature
Formal Principles of Language valuation/interpretability
Acquisition. Cambridge, MA: MIT biconditional in minimalist theory.
Press. [Link]
Wharton, T. 2010. Recipes: Beyond the publication/298215508
words. Gastrononomica 10: 67–73. Wiltschko, M. 1998. On the syntax and
Wierzba, M. 2014. What is special semantics of (relative) pronouns
about fronted focused objects in and determiners. Journal of
German: A study on the relation Comparative Germanic Linguistics
between syntax, intonation and 2: 143–81.
emphasis. MA thesis. University of ———2001. The syntax of pronouns:
Potsdam. Evidence from Halkomelem Salish.
———2019. Focus projection: Extending Natural Language and Linguistic
the empirical data base. Talk Theory 20: 157–95.
presented to the Phonology Woolford, E. 1991. VP‑internal subjects
Colloquium, Goethe Universität in VSO and nonconfigurational
Frankfurt, 6 November. languages. Linguistic Inquiry 22:
Wilder, C. 1989. The syntax of German 503–40.
infinitives. PhD diss. University ———2006. Case-agreement mismatches.
College London. In Agreement Systems, C. Boeckx
———1994a. Coordination, ATB, and (ed.) 317–39. Amsterdam:
ellipsis. Groningen Arbeiten zur Benjamins.
Germanistischen Linguistik 37: Wurff, W. van der 2007 (ed.)
291–329. Imperative Clauses in Generative
 References 539

Grammar: Studies in honour In Materials on Left Dislocation, E.

of Frits Beukema. Amsterdam: Anagnastopoulou, H. van Riemsdijk
Benjamins. & F. Zwarts (eds.) 119–48.
Wurmbrandt, S. 2006. Licensing case. Amsterdam: Benjamins.
Journal of Germanic Linguistics 18: Zagona, K. 1987. Verb Phrase Syntax.
175–236. Dordrecht: Kluwer.
———2019. Cross-clausal A‑dependencies. Zamparelli, R. 2000. Layers in the
Chicago Linguistic Society Determiner Phrase. New York:
Proceedings 54: 585–604. Garland.
Xu, L. 2003. Choice between the overt Zanuttini, R. 2008. Encoding the
and the covert. Transactions of the addressee in the syntax: Evidence
Philological Society 101: 81–107. from English imperative subjects.
Yamakoshi, K. 2002. Wh‑drop in Natural Language and Linguistic
child languages and adult ASL. Theory 26: 185–218.
In Proceedings of the GALA Zanuttini, R. & Portner, P. 2000. The
2001 Conference on Language characterization of exclamative
Acquisition, M. J. Freitas (ed.) clauses in Paduan. Language 76:
217–31. University of Groningen. 123–32.
Yamato, N. 2010. The left periphery ———2003. Exclamative clauses: At
of Japanese exclamatives. Studia the syntax-semantics interface.
Linguistica: 64: 55–80. Language 79: 39–81.
Yang, Y. 2009. Infinitival relative Zeijlstra, H. 2004. Sentential negation
clauses in English. English Language and negative concord. PhD diss.
and Linguistics 28: 137–55. University of Amsterdam.
Yatabe, S. 1990. Quantifier floating ———2012. There is only one way to
in Japanese and the θ‑hierarchy. agree. The Linguistic Review 29:
Chicago Linguistics Society 491–539.
Proceedings 26: 437–51. Zeller, J. 2006. Raising out of finite CP
Yoo, E.-J. 2001. English floating in Nguni: The case of fanele. South
quantifiers and lexical specification African Linguistics and Applied
of quantifier retrieval. Language Language Studies 24: 255–75.
and Information 1: 1–15. Zevakhina, N. 2013. Syntactic
———2002. A lexical approach to strategies of exclamatives. The
English floating quantifiers. Journal of Estonian and Finno-
Berkeley Linguistic Society Ugric Linguistics 4: 157–78.
Proceedings 28: 351–62. Zhang, S. 1990. The status of
Yuan, M. 2016. Subordinate clause types imperatives in theories of grammar.
and the left periphery in Kikuyu. PhD diss. University of Arizona.
Handout of paper presented to LSA Ziegler, J. 2014. Copy raising and phase
meeting in Washington, DC. theory: Finite complementisers look
Zaenen, A. 1997. Contrastive like they’re defective too. Poster
dislocation in Dutch and Icelandic. presented at the Linguistic Society
540 References

of America’s 88th Annual Meeting, Vennemann (eds.) 437–50. London:

Minneapolis. Routledge.
Ziv, Y. 1994. Left and right ———2002. I wonder what kind of
dislocations: Discourse functions construction that this example
and anaphora. Journal of illustrates. In The Construction
Pragmatics 22: 629–45. of Meaning, D. Beaver, L. D.
Zribi-Hertz, A. 1984. Prépositions Casillas Martínez, B. Z. Clark & S.
orphélines et pronoms nuls. Kaufmann (eds.) 219–48. Stanford,
Recherches Linguistiques 12: CA: CSLI Publications.
46–91. ———2005. Saying more with less.
Zubiaguirre Sebastián, Z. 2017. The Language Log post, 19 March.
syntax of English and Basque [Link]
wh‑exclamatives. BA diss. languagelog/archives/[Link]
University of the Basque Country. Zwicky, A. & Pullum, G. K. 1983.
Zubizaretta, M.-L. 1998. Prosody, Focus Deleting named morphemes.
and Word Order. Cambridge, MA: Lingua, 59: 155–75.
MIT Press. Zwicky, A. M. & Zwicky, A. D. 1971.
———2009. The syntax and prosody How come and what for. Ohio
of focus: The Bantu-Italian State University Working Papers in
connection. Iberia 2: 1–39. Linguistics 8: 173–85.
Zwart, J.-W. 1993. Dutch Syntax: A ———1981. Telegraphic registers in
minimalist approach. PhD diss. written English. In Variation
University of Groningen. Omnibus, D. Sankoff & H.
———1997. A Minimalist Approach to J. Cedergren (eds.) 535–44.
the Syntax of Dutch. Dordrecht: Edmonton: Linguistic Research lnc.
Kluwer. ———1982. Register as a dimension
———2000. A head raising analysis of of linguistic variation. In
relative clauses in Dutch. In The Sublanguage: Studies of language
Syntax of Relative Clauses, A. in restricted semantic domains, J.
Alexiadou, P. Law, A. Meinunger Lehrberger & R. Kittredge (eds.)
& C. Wilder (eds.) 349–86. 213–18. Berlin: De Gruyter.
Amsterdam: Benjamins. Zyman, E. 2018a. On the driving force
———2006. Complementiser agreement for syntactic movement. PhD diss.
and dependency marking typology. University of California at Santa
Leiden Working Papers in Cruz.
Linguistics 3: 53–72. ———2018b. Quantifier float as
Zwicky, A. M. 1988. On the subject of stranding: Evidence from Janitzio
bare imperatives in English. In On P’urhepecha. Natural Language and
Language: Rhetorica, phonologica, Linguistic Theory 36: 991–1034.
syntactica: A festschrift for Robert ———2022. On the definition of Merge.
P. Stockwell, C. Duncan Rose and T. Ms. University of Chicago.
 Index

A‑bar Movement, 32–36, 60 Be Drop, 375–90, 406 Island Constraint,

of negative XP, 258–62 Be Raising, 41–43 Freezing Principle
of topic, 222–24 Belfast English, 69–71, 87– constraint. See condition
of wh‑XP. See 88, 126, 209, 380–81 Constraint on Extraction
Wh‑Movement binary-branching structure, Domains, 45, 85–86,
A‑bar position, 35–36 17, 58 126
abbreviated registers, 34–35 blocking mechanisms, Control, 110–15
Activity Conditions, 141 44–52 agreement and case in,
adjunct(ion) 181–84
head adjunction, 18–19, 59 cartography contrasted with Raising,
Island Constraint, 44–45 of clause periphery, 115–19, 127
phrasal adjunction, 217–56 movement theory of, 127
20–22, 59, 306–9 of clause subperiphery, Coordinate Structure
phrasal adjuncts as 300–46 Constraint, 44–45
specifiers, 310–27, 340 of dropped constituents, copy
adverb(s) 347–419 Highest Copy Condition,
as adjuncts, 306–9 case-marking, 10, 28–31, 141
as specifiers, 310–24, 340 60, 144–49, 207–8, 214, Raising, 196–203, 209
Hierarchy, 316–17 216, 237–38 Theory of Movement,
order of, 315–24 default, 225 33–34, 60
Affix Hopping, 27–28, 60 exceptional. See ECM Criterial Freezing Condition,
agreement, 138–216 of topic, 238 219, 308
across finite clause, of who(m), 1–3, 216, 288 cross-clausal
184–98, 209 categories, 8–9, 58 A‑bar Movement, 46–47
deficit account of Be c‑command, 28–31, 60 agreement and case-
Drop, 378–82 Condition on Article marking, 184–203, 209
in A-Movement, 138–44 Drop, 374–75, 388–89 A-Movement/
in case-marking, 144–49 Clause Typing Conditions, Passivisation, 99–103,
in expletive clauses, 40–41 127
152–65 Complement Raising,
locus of, 209 321–24, 334–36 default
multiple, 159–65, 208 Complementiser Agreement, case, 225
A-Movement, 36–38, 146–47, 207 Spellout Rule, 33–34
69–137, 214, 216 complete clauses, 238–40 defective clause. See
Antilocality Constraint, 328 Completeness Condition, 141 truncated clause
argument structure, 77–79 Complex NP Constraint, Diary Drop, 356–57
Article Drop, 367–75, 44–45 discontinuous spellout. See
383–90, 395–97, 406 COMP-Trace Filter, 50–52, spellout
auxiliary 61, 105–6, 126 Distinctness Condition, 227
Drop, 354–66, 405 condition/constraint/ Doubly Filled COMP Filter,
Inversion, 38–43, 257–67, principle, 44–48 See 49–50, 61, 248
290 individual conditions, DP Hypothesis, 24
projections, 99–101, constraints and
300–5 principles such as ECM, 174–77
Raising, 300–5, 339 Impenetrability ellipsis, 23 See also
selection, 88, 126 Condition, Adjunct abbreviated registers
542 Index

empty categories. See null goal mixed structures, 119–23

EPP feature, 36–38, 60, 126, condition, 141 modifier projection/MODP,
141–43, 208 for a probe, 140–42 233–35
exclamatives, 32–34, theta role, 79–80 mood, 11
281–82, 290–91 grammar multiple
expletives components of, 13 agreement, 159–65, 208
agreement and case with, descriptive and topics, 225–28
152–65 prescriptive, 1–3 XPs of same type. See
conditions on the use of, sources of data for, 3–8, recursion
165–68 58
other analyses of, 208 negative
passive, 36–37, 95–97, head Fronting/Inversion,
99–101 Movement, 38–44 258–62
unaccusative, 85–86, 91 of a projection, 17 not in spec‑NEGP, 41–43,
External Argument how come, 270–76, 290 303–5
Condition, 165–66 n’t in NEG, 72–73
Impenetrability Condition, no TP no DP analysis,
feature(s), 9–12, 58 45–47, 61, 104–6 384–87
interpretable and imperatives, 87–88, 283–84, nonfinite. See finite
uninterpretable, 392–93, 406 null
149–52 Inactivity Condition, 165 article, 371–72
Valuation Conditions, 141 Indefiniteness Condition, case, 31
valuation of, 139–209 165 complementiser, 26–27
valued and unvalued, infinitive(s) constituents, 23–28,
149–52 agreement, A-Movement, 59–60
filters, 49–52 and case-marking in, determiner, 24
finite(ness) 168–85 object. See Object Drop
contrasted with nonfinite, Control, 181–83 preposition, 24–26
12–13 ECM, 174–77 quantifier, 23–24
projection/FINP, 235–38 for. See for-infinitives relative operator, 284–88
research on, 248 particle to, 168–84, 208–9 yes–no question operator,
floating quantifier. See passive, 177–79 39–41, 61, 277–81
quantifier, spellout Raising, 169–74
focus(ing), 35 interrogative. See also Object Drop, 390–401, 406
compared with topic, questions One Feature One Head
231–33 Generalisation, 270 Principle, 219
of interrogative wh‑XP, Inversion, 262–67 operator, null. See null
262–67 Intervention Condition, 48,
of negative XP, 258–62 61, 248 Passivisation, 36–38, 95–
of yes–no question Inversion. See Auxiliary 106, 126–27, 139–44
operator, 277–79 Inversion constraints on, 103–6
projection/FOCP, 228–31 in infinitives, 177–79
research on, 248 Left Dislocation, 222, 248 Path Containment
force projection/FORCEP, Left Edge Ellipsis, 362–64, Condition, 265
219–21 374, 378–82 periphery of clause, 217–56
in exclamatives, 281–82 Legibility Condition, 141 possessives, 369–70,
in imperatives, 283–85, Like-Adjoins-to-Like 372–73, 384–85
392–94 Constraint, 308–9 principle. See condition
in relatives, 283–87 locality Principle of Unambiguous
in wh‑questions, 262–70 of movement, 45–47 Reverse Engineering,
for-infinitives, 179–81, 359–60
235–36 merge, 13–18, 58 pro analysis of Object Drop,
Freezing Principle, 47–48 Hypothesis, 326–27 398–401
Functional Specifier Minimality Condition, probe, 140–42
Hypothesis, 326–27 30–31, 60, 103–4 Condition, 141
 Index 543

projection spellout in situ and ex situ,

types of, 17 Default Spellout Rule, 222–25
prosodic account of weak 34–36 multiple, 225–28
syllable drop, 364–66 discontinuous/split, projection/TOPP, 219–28
93–94, 97–98, 109–10, research on, 247–48
QP Hypothesis, 23–24 123 truncated clause, 241–43
quantifier Floating Quantifier Truncation account
floating/stranded, 73–74, Spellout Rule, 330 of Article Drop, 368–71
87, 98–99, 128–29, Low Spellout Rule, 94, of Auxiliary Drop, 358–62
327–36, 340 97–98 of e Drop, 378–83
scope, 72–73, 126 split of Subject (+Auxiliary)
questions INFL, 339 Drop, 352–58
how come, 270–76, 290 spellout. See spellout
wh‑questions, 262–70, 290 Structure Dependence unaccusative structures,
yes–no, 39–41, 61, Principle, 30 84–94, 126
277–81, 290 subject Uniform Theta Assignment
+Auxiliary Drop, 354–58 Hypothesis/UTAH,
Raising, 107–10 Drop, 347–58, 405 89–90, 99, 127
agreement and case in, Island Constraint, 44–45
169–74 originating within VP, variation
contrasted with Control, 69–137 between registers, 405
115–19, 127 projection/SUBJP, 305–10, between speakers, 6–8,
Recoverability Condition, 23 339 257–58
recursion subordinate clause, 261–62, Verb Raising (from V to
of complementiser, 236–37 324–27 T), 43
of CP, 220–21 subperiphery of clause, VP-Internal Subject
of FINP, 242–43 300–46 Hypothesis, 69–78,
of MODP, 234–35 successive-cyclic movement, 125
of TOPP, 225–28 45–47
registers, 405 Swiping, 63–64 weak syllable drop, 364–66
relative clause, 283–88, 291 Wh‑Movement
resumptive tense in exclamatives, 32–34,
relative clause, 287–88, affix, 27–28 281–82
292 deficit analysis of Be in interrogatives, 34–35,
topic structure, 222 Drop, 382–90 262–70
projection/TP, 16, 305–10 in relatives, 34–35,
scope theta 283–88
of auxiliary and negation, Criterion, 82, 126 wh‑questions. See
303–5 -marking/-roles, 78–84 questions
of quantifier and topic
negation, 72–73 case-marking of, 224–25 X-bar Syntax, 17–18
selection compared with focus,
Condition, 269–70 231–33 yes–no questions. See
restrictions, 99 Drop, 392–94 questions