English Phonological analysis – Chapter 6: Word Stress Page 1 of 32

Chapter 6

WORD STRESS
Miklós Törkenczy
Contents
6.1 Introduction: word stress
6.1.1 Weight sensitivity
6.1.2 The domain of metrification
6.2 Degrees of English stress: how many?
6.2.1 1ry stress vs. 2ry stress
6.2.2 3ry stress vs. other stresses
6.2.2.1 3ry stress vs. major stress(es)
6.2.2.2 3ry stress vs. zero stress
6.3 The predictability of stress in English words
6.3.1 Determining the place of primary stress within the word
6.3.1.1 Primary stress in words with a short-vowelled ult
6.3.1.2 Primary stress in words with a long-vowelled ult
6.3.1.3. Some complications
6.3.1.3.1 Word-medial .r. plus consonant clusters (sC)
6.3.1.3.2 Some ‘prefixes’ of Latin origin in verbs
6.3.1.3.3 Conversion and stress
6.3.1.4 Primary stress in derived words: suffixes
6.3.1.4.1 Stress-neutral suffixes
6.3.1.4.2 Stress-placing suffixes
6.3.1.4.3 Types of stress-placing suffixes
6.3.1.5 Primary stress patterns unaccounted for by the analysis
6.3.1.6 Summary of primary stress patterns
6.3.2 Determining the place of stresses preceding the primary stress
6.3.2.1 Determining the place of secondary stress within the word
6.3.2.1.1 Summary of secondary stress patterns
6.3.2.2 The place of 3ry stress before the 1ry stress
6.3.3 3ry stress after the primary stress
6.4 Checklist
English Phonological analysis – Chapter 6: Word Stress Page 2 of 32

6.1 Introduction: word stress

This chapter is about stress assignment in words, i.e. about the location of stress(es) in words
when they occur in isolation (what happens to these stresses when words are combined into
sentences is discussed in Chapter 7).
Stress is a suprasegmental feature. Unlike the features discussed in Chapter 3, it is not
realised on a single segment, but it extends over more than one segment: it is associated with a
syllable.
Stress is not an absolute property: it is the relative prominence of syllables. In contrast to
features like [voice] or [coronal], whose value is determinable independently of the environment
of the segment, it is not possible to tell whether a particular syllable is stressed or unstressed
without comparing it to other (neighbouring) syllables.
Notation: primary stress is indicated by an acute accent on top of a vowel letter in spelling,
e.g. átom, and the phonetic symbol . !. before the first segment of a syllable in transcription, e.g.
.!zs?l.; secondary stress is indicated by a grave accent on top of a vowel letter in spelling, e.g.
the first syllable of àtomístic, and the phonetic symbol .$. before the first segment of a syllable
in transcription, e.g. the first syllable of .$zs?!lHrsHj..
Metrification means determining where the stresses are in a word. For example, the word
cigarette is metrified as cìgarétte .$rHf?!qds..

6.1.1 Weight sensitivity

In some languages metrification is influenced by the weight of the syllables that make up
a word. These languages are weight-sensitive: they distinguish heavy (H) and light (L) syllables
(see 4.5.1). In these languages heavy syllables tend to attract stress. English is a weight-sensitive
language. Hungarian is weight-insensitive because in Hungarian, a word-initial syllable of any
weight is stressed regardless of the weight of any other syllable of the word.

6.1.2 The domain of metrification

In English, certain parts of the word are systematically excluded when the stresses are
determined. These portions of the word (a) do not get stress and (b) do not count for the
English Phonological analysis – Chapter 6: Word Stress Page 3 of 32

placement of stress. This means that metrification is restricted to a domain which is smaller than
the word. Thus, the domain of metrification is the portion of the word within which stress(es) can
occur and which may influence the placement of stress.
Some parts of the word may be outside the domain of metrification for morphological
reasons. English strong boundary affixes belong here (see 6.3.1.4.1); e.g. the suffix -ing is never
stressed and never changes the place of the stresses of the stem to which it is added (compare
èxcommúnicate with èxcommúnicating). Notation: strong boundary affixes are separated from
the stem by a number sign # in spelling and transcription, e.g. #excommunicat#ing# (the same
symbol # also appears at the beginning and the end of a word).
Some parts of the word may be outside the domain of metrification for phonological
reasons. These may be segments or syllables at the edges of words. Such phonological material
is called extrametrical (for extrametricality in English see 6.3.1.1). Extrametrical parts of the
word do not get stress and do not count for the placement of stress. Notation: extrametrical
material appears in angled brackets < > in spelling and transcription, e.g. ani<mal>.

In what follows we discuss the stress patterns of English words, focussing on two issues:
the degrees of stress and the predictability of stress(es). In the discussion we often have to refer
to specific syllables of a word. We will call

(i) the last syllable of the word (the ultimate syllable) the ult,
(ii) the second-last syllable (the penultimate syllable) the penult and
(iii) the third-last syllable (the antepenultimate syllable) the antepenult.

We will use the terms ult, penult and antepenult to refer to the actual syllables occurring in the
phonetic form of a word, i.e. regardless whether the final syllable is analysed as extrametrical or
not. For example, the underlined syllables of the words melon .!ldk?m. and balloon .a?!kt9m. are
the ultimate syllables of these words although the last syllable is extrametrical in the former, but
not in the latter me<lon> vs. balloon (see 6.3.1.1 and 6.3.1.2).
English Phonological analysis – Chapter 6: Word Stress Page 4 of 32

6.2 Degrees of English stress: how many?

As opposed to Hungarian, where there is only one stress in every word, a long(er) English word
may have more than one prominent syllable, e.g. CIgaRETTE, ExcoMUniCATE,
CIRcumNAviGAtion, etc. However, these relatively more prominent syllables are not necessarily
felt equally prominent. We can illustrate this with columns of stars (where the numbers of stars
correspond to levels of relative prominence within the word):

(1) ) )
) ) ) ) ) )
) ) ) ) ) ) ) )
) ) ) ) ) ) ) ) ) ) ) ) ) )
$rHf?!qds $djrj?!lit9mHjdHs $r29j?l$mzuH!fdHR?m

Most analyses would agree that the last syllable of excommunicate is more stressed than the
second and the fourth syllables, but less stressed than the first, and the first is less stressed than
the third, which is the strongest stress in the word.1 Given this, the question is how many degrees
of stress must be distinguished in English phonologically?
The traditional answer is that four degrees of stress are needed to describe the stress
patterns in words and these degrees are distinguished by a combination of factors: (i) the
loudness of the syllable, (ii) the pitch change occurring on the syllable, and (iii) the vowel quality
of its nucleus. This is shown in (2):

(2) Stress degrees

1 2 3 0
pitch change + ! ! !
loudness + + ! !
full vowel + + + !

1
Note that in this form, the star notation does not allow comparison of relative prominence
across words. The third syllable of cigarette is not less prominent (less stressed) than the third
syllable of excommunicate.
English Phonological analysis – Chapter 6: Word Stress Page 5 of 32

Cìgarétte .$rHf?!qds. and cìrcumnàvigátion .$r29j?l$mzuH!fdHR?m. have the stress patterns 2 0 1

and 2 0 2 0 1 0, respectively, and the word èxcommúnicate /$djrj?!lit9mHjdHs. 2 0 1 0 3,
exemplifies all four degrees of stress: primary (1ry), secondary (2ry), tertiary (3ry) and zero (0).
As can be seen in (2), 1ry stress (the ‘main’ stress of the word) is distinguished from 2ry
stress(es) by pitch change: when we say the word in isolation, a 1ry stressed syllable (also called
the ‘tonic’ syllable) is associated with a change in pitch (of any direction) while a 2ry stressed
syllable is not. 1ry stress and 2ry stress, called the major stresses (which I will abbreviate as
‘M’), are distinguished from 3ry stress and zero stress, called the minor stresses (which I will
abbreviate as ‘m’), by loudness (i.e. rhythmic prominence): the former are relatively loud
compared to the latter. 0 stress is distinguished from all the others by vowel quality: only reduced
vowels can occur in a zero stressed syllable. In English (as opposed to Hungarian) only a
restricted set of vowels (?+ h+ t+ H+ 'i(T) can occur in a zero stressed syllable2 – otherwise the
syllable must have a full vowel (this is called the rule of Vowel Reduction).
Even if we accept that all four degrees are distinguishable phonetically, it is not obvious
that all of them are necessary phonologically. Indeed, there are analyses that only distinguish two
degrees of word stress: stressed and unstressed.

6.2.1 1ry stress vs. 2ry stress

The difference between 1ry stress and 2ry stress can be shown to be a sentence/phrase level
distinction rather than a word-level one. If we say a word in isolation, we say it as a sentence,
with the appropriate tone on the tonic syllable of the sentence (e.g. falling if it is a statement). If
the same word occurs in a sentence consisting of more than one word, it may occur in such a
position that pitch change does not occur on any of its syllables, therefore the place of 1ry stress
in a word is the place of potential pitch change:

2
Syllabic nasals and liquids can also occur as the nuclei of 0 stressed syllables, e.g. the second
syllable of written Z!qHsm<\. We ignore this here for the sake of simplicity.
English Phonological analysis – Chapter 6: Word Stress Page 6 of 32

(3) 2 1
Cigarette.

2 2 2 2 1
I quit smoking cigarettes for good.

This means that at the word level, phonologically, we only need to know the location of major
stresses because the difference between the two kinds of major stresses, the actual place of pitch
change (the 1ry stressed syllable) is determined at sentence/phrase level. The location of the 1ry
stressed syllable is predictable: in a neutral sentence it is the rightmost major-stressed syllable
(see Chapter 7 on sentence stress). Since we are looking at isolated words here, all we need to
know is the location of major stresses – the rightmost one will be the potential bearer of pitch
change in a sentence, i.e. the rightmost one is the 1ry stress. For ease of reference, we will
continue to use the terms 1ry and 2ry stress as shorthand for rightmost and non-rightmost major
stress in a word, respectively.

6.2.2 3ry stress vs. other stresses

If we want to find out whether 3ry stress is phonologically different from the other stresses
(major stresses and zero stress), we have to find phonological regularities that treat them
differently, i.e. we must demonstrate that 3ry stress patterns differently from the major stresses
on the one hand, and zero stress on the other.
Notation: 3ry stress is unindicated in spelling and transcription by any special phonetic
symbol. A 3ry stressed syllable is one that has a full vowel but bears no stress mark.

6.2.2.1 3ry stress vs. major stress(es)

In isolated words, 3ry stressed syllables can be found (i) immediately preceding major-stressed
syllables,(ii) word-medially after a major stress and (iii) in word-final position.
English Phonological analysis – Chapter 6: Word Stress Page 7 of 32

(4) i. tormént .sN9!ldms., còndensátion .$jPmcdm!rdHR?m.

ii. órgasm .!N9fzy?l., còndensátion .$jPmcdm!rdHR?m.
iii. róbot .!qntaPs., séparateV .!rdo?qdHs.

In some positions 3ry stress is in complementary distribution with 2ry stress. (i) 3ry stress never
occurs on the second syllable preceding a major stress (where major stresses do) and (ii) 3ry
stress can occur after the last major stress in the word, i.e. the 1ry (where by definition a major
stress cannot). This is exemplified in (5i, ii) and summarised in a table in (5iii).

(5) i. àcadémic .$zj?!cdlHj., cìrcumnàvigátion .$r29j?l$mzuH!fdHR?m.3

ii. róbot .!qntaPs., séparateV .!rdo?qdHs.

iii.
position
2 σ before 1ry stress after 1ry stress
2ry (=major) stress àcadémic !
3ry stress ! róbot

A position in which a 3ry stress and a major stress seem to contrast is word-initial position
preceding a major stress ( # _ M ) because they are treated differently by the Rhythm Rule. The
Rhythm Rule4 (also called Rhythmic Stress Deletion/Shift or Iambic Reversal) is an optional
postlexical5 rule that can downgrade a major stress for rhythmic reasons if in the sentence it
occurs ‘too close’ to other major stresses. The downgraded syllable retains its full vowel. The
word èverlásting (2 0 1 0) has two major stresses, but the second one can be downgraded to 3ry
stress by the Rhythm Rule in a sentence like She shivered in the everlasting rain because it

3
The third syllable of cìrcumnàvigátion can lose its major stress but not its full vowel
(cìrcumnavigátion) as a result of the Rhythm Rule, see below and Chapter 7.
4
See Chapter 7
5
See Chapter 8
English Phonological analysis – Chapter 6: Word Stress Page 8 of 32

occurs between two major stresses that are close.6

(6) 2 2 2 1 2 2 3 1
She shivered in the everlasting rain = She shivered in the everlasting rain

As a result, the prominence relations of èverlásting , whose third syllable is more prominent than
the first in isolation and in some sentences (The rain is everlasting.), are reversed by the Rhythm
Rule: èverlasting, as in (6).
This cannot happen to a word like Octóber, which looks very similar superficially, having
a full-vowelled syllable followed by a more prominent full-vowelled syllable .Pj!snTa?.. If we
put this word in a shifting context where the Rhythm Rule could apply, we find that the rule does
not apply and the prominence relations do not change. The second syllable always remains more
prominent that the first. We can explain this by claiming that this is due to a difference in stress:
the first syllable of everlasting has 2ry stress, but the first syllable of October has 3ry stress. A
major stress can be downgraded to 3ry stress by the Rhythm Rule, but a minor stress (like 3ry
stress) cannot be upgraded to become a major stress for rhythmic reasons.

(7) 2 3 2 1 2 2 3 1
She shivered in the October rain = *She shivered in the October rain.
.Pj!snTa?. ).!PjsnTa?.

This means that there is a regularity that treats these two types of stresses differently, so they
have to be distinguished: 3ry stress exists as a separate stress-degree.
A full-vowelled syllable that is two syllables before a major stress can always undergo the
Rhythm Rule, i.e. behaves like a major stress (in accordance with (5iii) above), but a full-
vowelled syllable that immediately precedes a major-stressed syllable can behave in two ways.
Some words (8i) undergo the Rhythm Rule, others do not (8ii):

6
Defining what counts as close enough is a complicated and interesting question we will not
discuss here.
English Phonological analysis – Chapter 6: Word Stress Page 9 of 32

(8) i. sàrdíne $r@9!ch9m = sàrdine sándwich $r@9ch9m !rzmvHcY

ii. Octóber Pj!snTa? = Octòber ráin Pj$snTa? !qdHm

iii.
position
2 σ before after 1ry stress 1 σ before
1ry stress 1ry stress
2ry (=major) stress àcadémic ! sàrdíne
3ry stress ! róbot Octóber

In this position then 3ry stress and a major stress appear in contrast: the first syllable of
Octóber ( # 3 1 0 ) seems to have 3ry stress and the first syllable of sardíne ( # 2 1 ) seems to
have major stress (8iii). However, even this position of potential contrast (# _ M ) disappears
if we consider the larger environment within the word. All words in which there is a full-
vowelled syllable (F) followed by a major stress and are bisyllabic ( # F M #) seem to behave like
sardine and undergo the Rhythm Rule (9i); and all trisyllabic words in which a completely
unstressed syllable follows the major stress ( # F M 0 # ) seem to behave like October and fail
to undergo the Rhythm Rule (9ii):

(9) i. Two-syllable words # F M #

sàrdíne $r@9!ch9m sàrdine sándwich $r@9ch9m !rzmvHcY
dìréct $c`H!qdjs dìrect débit $c`Hqdjs !cdaHs
Bàhréin $a@9!qdHm Bàhrein Ísland $a@9qdHm !`Hk?mc
cartóon $j@9!st9m càrtoon nétwork $j@9st9m !mdsv29j

ii. Three-syllable words # F M 0 #

Octóber Pj!snTa? Octòber ráin Pj$snTa? !qdHm
Titánic s`H!szmHj Titànic’s bánd s`H$szmHjr !azmc
salvátion rzk!udHR?m Salvàtion Ármy rzk$udHR?m !@9lh
Montána lPm!szm? Montàna béar lPm$szm? !ad?
English Phonological analysis – Chapter 6: Word Stress Page 10 of 32

Given this regularity, the Rhythm Rule provides no argument for distinguishing 3ry stress from
major stress. There is no need to say that words like (9i) are different from those in (9ii) because
their initial syllables have different stress (major in the former, minor in the latter) since we can
distinguish them with reference to their syllabic structure ( # F M # vs. # F M 0 # ) and say that
the Rhythm Rule does not apply in the latter because of the syllabic environment. Then, we could
maintain that both groups have the same kind of stress, i.e. 2ry stress, in their first syllable.

There are other rules too that group 3ry stress together with major stresses in English. One of
them is the phonotactic constraint that restricts Zg\ to word-initial position and to a position before a
full vowel. As can be seen in (13), it makes no difference if this full vowel occurs in a major-stressed
syllable or in a . 3ry stressed one:

(10) The distribution of .g.

.g. pronounced /h/ not allowed

before a σ 1ry 2ry 3ry any 0 any

with _ stress

position any #_ _C _#

a.héad appre.hènsibility Só.ho helló Grah/am Joh/n Shah/

?!gdc $zoqH$gdmrH!aHk?sh !rnTgnT g?!knT fqdH?l cYPm R@9

These arguments suggest that the traditional distinction between 3ry stress and major stresses is
phonologically questionable.

6.2.2.2 3ry stress vs. zero stress

The difference between a 3ry stressed syllable and a zero stressed syllable is that the former
contains a full vowel while the latter contains a reduced one. This seems a neat distinction, but
unfortunately, in addition to the vowels that only occur in unstressed syllables .?+ h+ t., there are
vowels that can occur in both unstressed syllables and under major stress .H+ 'i(T., i.e. there is
an overlap between full and reduced vowels:
English Phonological analysis – Chapter 6: Word Stress Page 11 of 32

(11) full V reduced V

? ! machíne l?!Rh9m
T búsh !aTR ámulet !zliTk?s
H kiss !jHr áttic !zsHj

In some cases (e.g. áttic) this makes it impossible to tell if a syllable has 3ry stress or zero stress
– so there is no difference. These syllables are indeterminate between 3ry and zero stress, as there
is no contrast between them, we can safely (and arbitrarily) consider them unstressed.
There seem to be no rules of English phonology that consistently group 3ry stress with 0
stress.

6.3 The predictability of stress in English words

As opposed to a language like Hungarian, which has fixed stress (always on the first syllable of
a word), English stress is free in that stress can be on any syllable of the word.
There are two conflicting views about English stress. One view, the ‘no-pattern view’
maintains that there is no stress pattern in English, i.e. English stress is not predictable.
According to the other view, the ‘pattern-with-exceptions view’, there is a very intricate pattern
(with many rules and exceptions).

The ‘no-pattern view’

According to proponents of the no-pattern view no rules that are sufficiently general can
be formulated about the place of word stress, therefore English word stress is lexical, it has to
be memorised for every word by native speakers.
There are some regularities, e.g. 1ry stress has to fall on one of the last three syllables of
an English word, i.e. within ‘the final three-syllable window’, but the place of stress is
unpredictable within the limits of these regularities.
Take syllable weight as an example. English stress is often claimed to be weight-sensitive,
English Phonological analysis – Chapter 6: Word Stress Page 12 of 32

i.e. the weight of the syllables influences the place of 1ry stress. The proponents of the no-pattern
view would argue that this is not true because it can be shown that within the final three-syllable
window, a syllable of any weight in any position (ultimate: σσσ#, penultimate: σσσ#,
antepenultimate: σσσ#) may be 1ry-stressed. Consider (12) below (where V is a short vowel, V9
is a long vowel and 1ry stressed syllables are emboldened). There are no word-final stressed light
syllables in English (*V#, e.g. ).oz".), this is why the cell in the top right corner is empty.

(12) A syllable of any weight may be stressed (within the final σσσ# window)

antepenult stress σσσ# penult. stress σσσ# ult. stress σσσ#

LIGHT ánimal !z-mH-l?k vanílla u?-!mH-k? !

HEAVY dígnity !cHf-m?-sh enígma ?-!mHf-l? pìcturésque $oHj-sR?!-qdrj

Furthermore, not only is the location of 1ry stress independent of the weight of the 1ry stressed
syllable itself, but the location of 1ry stress is also independent of the weight of any other syllable
within the final three-syllable window.
(13) below shows that any syllable may have 1ry stress within the final three-syllable
window independently of the weight of the ultimate syllable (1ry stressed syllables are
emboldened and the ultimate syllables are underlined).

(13) The location of stress and the weight of the ULT

LIGHT ULT HEAVY ULT

σσσ# cínema !rH-m?-l? dy3 namite !c`H-m?-l`Hs

σσσ# enígma ?-!mHf-l? potáto o?-!sdH-snT

σσσ# ! Jàpanése $cYz-o?-!mh9y

(14) shows that any syllable may have 1ry stress within the final three-syllable window
independently of the weight of the penultimate syllable (1ry stressed syllables are emboldened
and the penultimate syllables are underlined).
English Phonological analysis – Chapter 6: Word Stress Page 13 of 32

(14) The location of stress and the weight of the PENULT

LIGHT PENULT HEAVY PENULT

σσσ# ánimal !z-mH-l?k cárpenter !j@9-o?m-s?

σσσ# vanílla u?-!mH-k? enígma ?!-mHf-l?

σσσ# cìgarétte $rH-f?-!qds chìmpanzée $sRHl-o?m-!yh9

(15) shows that any syllable may have 1ry stress within the final three-syllable window
independently of the weight of the antepenultimate syllable (1ry stressed syllables are
emboldened and the antepenultimate syllables are underlined).

(15) The location of stress and the weight of the ANTEPENULT

LIGHT ANTEPENULT HEAVY ANTEPENULT

σσσ# ánimal !z-mH-l?k scénery !rh9-m?-qh

σσσ# vanílla u?-!mH-k? Novémber mnT!-udl-a?

σσσ# cìgarétte $rH-f?-!qds mútinéer $lit9-sH-!mH?

The ‘pattern-with-exceptions view’

According to the pattern-with-exceptions view, there is a stress pattern in English, stress is

(mostly) predictable: it is the result of several factors (partly morphological and phonological)
and there are exceptions (English stress is partly lexical).
The argument is that
(a) some of the stress patterns we have seen in (12, 13, 14, 15) above are rare or less
frequent than others. The words ánimal and vanílla are very similar in terms of weight structure:7
L L σ # and both of them are nouns, but ánimal has antepenultimate and vanílla has penultimate

7
Remember: H stands for a heavy syllable, S for a superheavy syllable, L for a light one and
σ is a syllable of any weight. For the definitions of heavy, light and superheavy syllables, see
4.5.1 on syllable weight.
English Phonological analysis – Chapter 6: Word Stress Page 14 of 32

1ry stress. The stress pattern of the former is much more frequent than that of the latter, therefore
ánimal can be seen as regular and vanílla as an exception;
(b) word-class matters in stress placement. The verb objéct and the noun óbject have
different stress patterns both of which are regular and follow the stressing of verbs and nouns;
(c) native speakers have intuitions about the place of stress: if they are required to guess
the place of stress in the nonsense noun phalidon they are much more likely to say .!ezkHc?m. or
.e?!k`Hc?m. than .)ez!kHc?m., .)!ezk`Hc?m. or .)$ezkH!cPm.. This also shows that stress in English
is sensitive to syllable weight (in this case the weight of the penultimate syllable which gets the
stress if it is heavy (see the details later)).
(d) some suffixes determine the place of 1ry stress (e.g. -ity places 1ry stress on the
immediately preceding syllable: cívil but civílity) and the way they do is related to their
phonological shape. Compare civíl-ity and ánimal – both are nouns, both have the weight
structure8 L L σ # in the final three-syllable window (underlined above) and both have
antepenultimate stress.
Therefore, in the rest of this chapter we follow the pattern-with-exceptions view and
examine the regularities of stress placement in English words

6.3.1 Determining the place of primary stress within the word

6.3.1.1 Primary stress in words with a short-vowelled ult

If we first examine underived words whose last syllable does not have a long vowel, we find the
following two patterns:

8
For the definitions of heavy (H), light (L) and superheavy (S) syllables, see 4.5.1 on syllable
weight.
English Phonological analysis – Chapter 6: Word Stress Page 15 of 32

(16) THE NOUN PATTERN

i. 1ry stress falls on a H penult, if there is one !H σ #
examples: agénda, appéndix, horízon

ii. otherwise 1ry stress falls on the antepenult !σ L σ #

examples: América, ásterisk

(17) THE VERB PATTERN

i. 1ry stress falls on a Superheavy ult, if there is one !S #
examples: prevént, condúct

ii. otherwise 1ry stress falls on the penultimate σ !σ σ #

examples: inhábit, imágine

Nouns (typically), suffixed adjectives (typically) and some unsuffixed adjectives follow the noun
pattern and verbs (almost always) and some unsuffixed adjectives and adjectives ending in -ic
follow the verb pattern.

(18) Examples
THE NOUN PATTERN
(a) Nouns
i agénda, veránda, synópsis, uténsil, horízon, aróma, macaróni, Minnesóta
ii América, cínema, vértebra, ánimal, vénison, président

(b) Adjectives (suffixed)

i dialéctal, moméntous, triúmphant, contíngent, mediéval, inhérent, anecdótal
ii pérsonal, dángerous, máximal, vígilant, dífferent

(c) Adjectives (unsuffixed)

i pérfect, éarnest, ádult (BrE), éxpert, áwkward, bástard
ii móribund, dérelict, dífficult, mánifest, táciturn
English Phonological analysis – Chapter 6: Word Stress Page 16 of 32

THE VERB PATTERN

(a) Verbs
i prevént, collápse, tormént, eléct, adópt, reláx
ii imágine, astónish, embárrass, prómise, detérmine

(b) Adjectives (unsuffixed)

corrúpt, ovért, diréct, defúnct, absúrd, adúlt (AmE)

(c) Adjectives in -ic

atómic, cosmétic, económic, fanátic, sadístic, terrífic

The two basic patterns, the noun pattern and the verb pattern are phonologically the same.
The only difference between them is the domain of metrification. Specifically, extrametricality
(see 6.1.2) is different for noun pattern items and verb pattern items. In noun pattern items it is
the last syllable of the word that is extrametrical, while in verb pattern items it is only the last
consonant (if there is one).

(19) Noun pattern extrametricality: the last syllable is extrametrical

σ# = <σ> #

(20) Verb pattern extrametricality: the last consonant is extrametrical

C# = <C> #

Thus the main stress rule that accounts for the basic patterns can be given as follows:

(21) Main Stress Rule (MSR)

Within the domain of metrification
i. the rightmost syllable gets 1ry stress if it is heavy
ii. otherwise: the syllable preceding the rightmost syllable gets 1ry stress if the
rightmost syllable is light
English Phonological analysis – Chapter 6: Word Stress Page 17 of 32

This gives us the following metrifications for the patterns examined (and correctly predicts the
place of primary stress).

(22) N-pattern i. Heavy penult !H<σ># agén<da> ?!cYdmc?

ii. antepenult !σL<σ># Améri<ca> ?!ldqHj?

V-pattern i. Superheavy ult !H<C># prevén<t> oqH!udms

ii. penult !σL<C># inhábi<t> Hm!gzaHs

Note that the weight of the rightmost syllable is determined only within the domain of
metrification, i.e. disregarding the extrametrical material. For example, the superheavy ult of
prevent counts as heavy for metrification and the heavy ult of inhabit counts as light because the
last consonant in both is extrametrical and only ven and bi are considered by the Main Stress
Rule.

6.3.1.2 Primary stress in words with a long-vowelled ult

Underived words that have a long vowel in their ult behave in the following way:

(i) long vowels in final syllables are regularly stressed in bisyllabic words (even in nouns), and
(ii) a word regularly has antepenultimate stress if it is longer than two syllables and has a long
vowel in its final syllable (even if it is a verb):

(23) i 2 σ brocádeN .aq?!jdHc., canóeN /j?!mt9., sedáteV /r?!cdHs., obéyV .?!adH.

ii 2+ σ dy3 namiteN .!c`Hm?l`Hs., pédigreeN .!odcHfqh9., óperateV .!Po?qdHs., pétrifyV
.!odsqhe`H.

Note that this stress pattern is independent of morphological class membership: in words of this
type there is no difference between the stressing of nouns and verbs.
This pattern is handled by two rules: Long Vowel Stressing (LVS) and the Alternating
English Phonological analysis – Chapter 6: Word Stress Page 18 of 32

Stress Rule (ASR):

(24) Long Vowel Stressing (LVS)

Stress long vowels in final syllables

LVS overrides extrametricality and prevents it from applying: brocáde aq?!jdHc, not
*bró<cade>.

(25) The Alternating Stress Rule (ASR)

σ σ !σ # = !σ σ σ # (if a word-final syllable has been stressed by some rule
and it is preceded by two or more syllables, move the
stress to the antepenultimate syllable)

ASR applies to the output of LVS, e.g. dynamíte = dy3 namite .!c`Hm?l`Hs.. ASR only applies
to final stressed syllables and it applies to all final stressed syllables, not only to those stressed
by LVS but also to those stressed by the general Main Stress Rule (MSR): genufléct = génuflect.
Sample derivations are shown in (26), and table (27) summarises the working of LVS and ASR:9

9
Word-final .nT. does not seem count as a long vowel. vé<to> not *vetó; potá<to> not
*pótato.
English Phonological analysis – Chapter 6: Word Stress Page 19 of 32

(26) Long Vowel Stressing and/or the Alternating Stress Rule in derivations

Verb Noun Verb Verb

UR #sedate# #caraway# #operate# #genuflect#

LVS #sedáte# #carawáy# #operáte# –

Extr – – – #genuflec<t>#
MSR #sedáte# #carawáy# #operáte# #genufléc<t>#
ASR – #cáraway# #óperate# #génuflec<t>#

SR Zr?!cdHs\ Z!jzq?vdH\ Z!Po?qdHs\ Z!cYdmiTekdjs\

(27) Summary of Long Vowel Stressing and the Alternating Stress Rule

i. Words with a long-vowelled ult

2+ σ 2σ
Nouns ánecdote ballóon
LVS + ASR LVS
Verbs décorate debáte
LVS + ASR LVS

ii. Words with a short-vowelled ult

2+ σ 2σ
Nouns
not relevant10 not relevant

Verbs génuflect eléct

MSR + ASR MSR

10
Not relevant because the short-vowelled ult of nouns is extrametrical and cannot get stress.
English Phonological analysis – Chapter 6: Word Stress Page 20 of 32

6.3.1.3 Some complications

6.3.1.3.1 Word-medial /s/ plus consonant clusters (sC)

In some words a medial sC cluster appears to syllabify as an onset (V.sC), in others as a

coda+onset sequence (Vs.C). If the vowel preceding the sC cluster is short, then its syllable
should count as light under the former syllabification, but as heavy under the latter one (because
a closed syllable is heavy even if its vowel is short). Compare the nouns mínister .!lHmHrs?. and
asbéstos .zr!adrs?r.. The correct stress pattern is predicted if they are analysed like this:

(28) Syllabification of sC clusters and stress

i. .r. in onset ii. .r. in coda
mi.ni.ster = míni<ster> as.bes.tos = asbés<tos>
an.ce.stor = ánce<stor> A.las.ka = Alás<ka>
or.che.stra = órche<stra> Fran.cis.can = Francís<can>
in.du.stry = índu<stry> a.spi.dis.tra = aspidís<tra>
Pro.te.stant = Próte<stant> con.tes.tant = contés<tant>

This is a problem for syllabification, because the same sequence of segments must be syllabified
in two different ways and it is unpredictable which sC words syllabify in which way.

6.3.1.3.2 Some ‘prefixes’ of Latin origin in verbs

(e.g. o=, ex=, im=, con=, re=, inter=, contra=, intro=, re=11, etc.)
These are not proper prefixes12 in the sense that in English they do not have an identifiable
meaning and the bases that they precede do not have an identifiable meaning either (e.g. omit,

11
not the prefix re- ‘again’
12
This is indicated here by the special boundary symbol “=”, see 8.1.2
English Phonological analysis – Chapter 6: Word Stress Page 21 of 32

explain, confess, intervene, etc.). However, they may interfere with the stress rules (MSR, ASR)
discussed. Typically, they fall outside the domain of 1ry stress placement, may not receive 1ry
stress (although they receive 2ry stress regularly) and only the base is visible to the stress rules
discussed above. E.g. the verb omít should be *ómit according to the MSR (*ómi<t>, compare
édi<t>), but it must be analysed as o=mít to get the actual stressing .?!lHs.; the verb ìntervéne
should be *íntervene according to LVS and ASR (*íntervene, compare óperate), but it must be
analysed as inter=véne to get the actual stressing .$Hms?!uh9m.. The problem is that, as there is no
real morphological motivation for analysing these word-initial sequences as prefixes, the analysis
is circular: the ‘prefix’ analysis is only motivated by the anomalous stressing it is designed to
explain. According to this ‘explanation’ omít has final stress because is has a ‘=’ boundary in the
middle; and it must have a ‘=’ boundary in the middle because it has final stress!

6.3.1.3.3 Conversion and stress

In English, verbs can be freely derived from nouns by conversion (zero derivation) without any
change in pronunciation (including stress). Take, for instance, the nouns plátypus
(Ornithorhynchus anatinus, ‘kacsacsőrű emlős’) and chihuáhua (‘a Mexican breed of dog ’), both
of which follow the noun pattern: pláty<pus> .!okzs?o?r/ and chihuá<hua> .sRH!v@9v?.- In the
pair of sentences below the word for the same animal is a noun in one sentence and a verb in the
other – nevertheless the stressing do not change, it remains chihuahua .sRH!v@9v?. and platypus
.!okzs?o?r.:

(29) Don't you chihuáhuaV my plátypusN! Don't you plátypusV my chihuáhuaN!

(‘Don’t call my platypus a chihuahua!’) (‘Don’t call my chihuahua a platypus!’)

The problem is that in this way there is a great number of verbs that actually follow the noun
stress pattern. A possible analysis is to say that conversion applies after the stress rules:13

13
Note that it is sometimes not possible to tell when conversion happens and when it does not:
deceptively similar pairs of words may be related by conversion in one instance and have
different (noun or verb) stress patterns in the other: e.g. có<mment>N; có<mment>V .!jPldms.
(conversion) vs. ré<cord>N .!qdjN9c.; recór<d>V .qH!jN9c. (noun and verb patterns respectively).
English Phonological analysis – Chapter 6: Word Stress Page 22 of 32

(30) UR #platypus#NOUN

Extr #platy<pus>#NOUN
MSR #pláty<pus>#NOUN
conversion #plátypus#VERB

SR Z!okzs?o?r\

6.3.1.4 Primary stress in derived words: suffixes

Phonologically, there are two kinds of suffixes: (i) stress-placing suffixes determine the place of
primary stress in the word, they may ‘overwrite’, i.e. not preserve, the stress of the base,
(ii) stress-neutral suffixes leave the original stress of the base intact.14

6.3.1.4.1 Stress-neutral suffixes15

The following are some common stress-neutral suffixes: -able, -ly, -ing, -ed, -es, -er, -ist, -ism,
-ful, -less, -ness, -hood, -ishadj , -ment, -wise.
Stress-neutral suffixes are outside the domain of metrification (see 6.2.1), i.e. neutral
suffixes are disregarded when primary stress placement is determined and the rest of the word
is metrified without the suffix:

(31) #electr+ic+ity#wise# = #electríci<ty>#wise# (by MSR: N-pattern)

14
See BEP 12.15-33 for examples.
15
These suffixes are also called ‘strong-boundary’, ‘#-boundary’, ‘Level 2'
English Phonological analysis – Chapter 6: Word Stress Page 23 of 32

#erot+ic#ism# = #eróti<c>#ism# (by MSR: V-pattern)

#class+ifi#able# = #clássify#able# (by LVS + ASR)

6.3.1.4.2 Stress-placing suffixes16

Words (only) containing stress-placing suffixes are metrified in the same way as underived
words, i.e. (i) the reason why stress-placing suffixes influence the placement of stress is that they
are metrified together with the base they are added to and (ii) (ideally) the way a particular
stress/placing suffix influences the placement of primary stress derives (a) from the phonological
shape of the suffix and (b) its morphological properties (i.e. whether it derives nouns, verbs, etc.).
For example the suffix -ity places the stress on the syllable preceding the suffix because it
derives a noun, so its final syllable will be extrametrical according to (19) and its initial syllable
is light, -ity is -Lσ, so it has to be the rightmost syllable of a bisyllabic left headed foot according
to the MSR: (σL)<σ>. Therefore, the syllable preceding it will be stressed:

(32) electr+ic+ity = electríci<ty> (by MSR: N-pattern)

6.3.1.4.3 Types of stress-placing suffixes17

A. Pre-stressed 1
1ry stress falls on the syllable preceding the suffix. There are two subclasses according to
suffix shape:

16
These suffixes are also called ‘restressing’, ‘non-neutral’, ‘weak-boundary’, ‘+boundary’,
‘level 1', ‘stress-fixing’
17
Some of these ‘suffixes’ are ‘endings’ rather than suffixes proper, i.e. (i.e. they do not have
any meaning).
English Phonological analysis – Chapter 6: Word Stress Page 24 of 32

SHAPE
(i) +Lσ -uble, -ity, -ety, -erie, -ion18, -ular, -logy, -meter, -graphy, -poly, -tomy, -pathy,
-thesis, -gamy
These suffixes consist of a light syllable followed by another syllable. They follow
the noun pattern and stress placement follows from their shape: abíli<ty>
confórmi<ty>

(ii) +H -ic, -id, -ishV

These suffixes consist of a single heavy syllable. They follow the verb pattern and
stress placement follows from their shape: anatómi<c>, militarísti<c>

B. Pre-stressed 1/2
1ry stress falls on the syllable preceding the suffix if it is H, but on the second syllable
preceding the suffix if the syllable preceding the suffix is L. There are two subclasses
according to suffix shape:

Suffix shape
(i) +σ -age, -al, -ous, -ive, -ure, -ant, -ance, -ent, -ence
These suffixes consist of a single vowel-initial syllable. They follow the
noun pattern and stress placement follows from their shape: medíci<nal>,
parén<tal>

(ii) +σσ -ative, -ature, -ible, -ary, -ory

1ry stress placement does not follow from the shape of these suffixes.

C. Pre-stressed 2
1ry stress falls on the 2nd syllable preceding the suffix

Suffix shape
+(C)V9(C) -ateV, -ize, -ite, -ene, -ine, -cide, -oir, -ose, -tude, -(i)fy

18
Note that -ion counts as two syllables for stress purposes, e.g. definíti<on>
English Phonological analysis – Chapter 6: Word Stress Page 25 of 32

As all these suffixes consist of a single, long-vowelled syllable, they

regularly get 1ry stress by Long Vowel Stressing and the Alternating Stress
Rule. (Of course, if there is only one syllable before the suffix, only LVS
can apply and the suffix gets the stress). Thus, stress placement follows
from their shape: rádiate, sedáte

D. Auto-stressed
1ry stress falls on the suffix itself. Auto-stressed verbal endings that consist of a
superheavy syllable are not special in bisyllabic words – they simply follow the basic verb
pattern: -ainV maintái<n>. Otherwise, auto-stressed suffixes are exceptional and 1ry stress
placement does not follow from their shape. There are three subclasses according to shape:

Suffix shape
(i) +(C)V9(C) -ade, -ese, -ique .h9j., -ee .h9., e.g. lemonáde
(ii) +VCC -esque .drj.,; e.g. picturésque
(iii) +VC -elle, -enne, -esse, -esce, -ette; e.g. novelétte

6.3.1.5 Primary stress patterns unaccounted for by the analysis

There are some primary stress patterns in English which are not predicted by this analysis. We
consider them irregular here,19 e.g. ceméntN .r?!ldms., vaníllaN .u?!mHk?., téndencyN .!sdmc?mrh.,
céremonyN .!rdq?l?mh., harássV .g?!qzr.20, rábbi .!qza`H., kàngaróo .$jzMf?!qt9., ellípsoid
.H!kHorNHc..

6.3.1.6 Summary of primary stress patterns

19
There are analyses which postulate additional rules to account for (some of) these patterns.
.
20
HarassV has a regularly stressed pronunciation in traditional RP: .gz!q?r.. According to
Wells (2008), while there is a 68% to 32% preference for .gz!q?r. in RP, there is a 60% to 40%
preference for .g?!qzr. among ‘younger’ people.
English Phonological analysis – Chapter 6: Word Stress Page 26 of 32

1ry stress is
(i) calculated right-to-left from a strong (#) boundary,
(ii) weight-sensitive,
(iii) not stress preserving (1ry stress placement can change the 1ry stress of the base in
words derived by + boundary suffixes) and
(iv) partially unpredictable (there are exceptional patterns).

6.3.2 Determining the place of stresses preceding the primary stress

If we distinguish between 2ry stress and 3ry stress, both of these stresses can precede 1ry stress,
e.g. còndensátion .$jPmcdm!rdHR?m. (where the . 3ry stress is underlined). In this subsection we
examine the predictability of these stresses.

6.3.2.1 Determining the place of 2ry stress within the word

The place of 2ry stress is predictable and is due to the interaction of constraints, i.e. restrictions
on the occurrence of 2ry stress, some of which are violable. These constraints are not equally
important, they are ranked, i.e. some are more important to obey than others. These constraints
are:

(33) NO STRESS CLASH * #...MM...#

There should be no adjacent major stresses (2ry and 1ry)
(= *#...21...#; *#...22...#)

This constraint does not hold in bisyllabic words with a final major (i.e. 1ry) stress, e.g. sàrdíne,
prìncéss, fòurtéen, etc. see Section 6.2.2.1. So #MM# is permitted.21

(34) EARLY STRESS *#mm...

21
We will not analyse #MM# words here.
English Phonological analysis – Chapter 6: Word Stress Page 27 of 32

There must be a major stress on one of the two syllables at the

beginning of a word
(= *#00; *#30; *#03)

This constraint seems to be inviolable, i.e. there are no exceptions.

(35) STRESS PRESERVATION A derived word has to preserve the placement of the major
stress(es) of its base.

We have seen that 1ry stress (i.e. the rightmost major stress) is not stress-preserving, compare
átom and atómic. In 2ry stress placement, however, there is a tendency to preserve the major
stresses of the base if possible. The reason why there is a difference between the placement of
2ry stress in chàracterístic and orìginálity is that their bases (cháracter and oríginal22) have their
1ry stresses on different syllables, and the derived words both preserve the place of the 1ry stress
of their bases. However, stress preservation is not always possible. STRESS PRESERVATION is a
violable constraint: it applies as long as NO STRESS CLASH and EARLY STRESS are not violated.
It is not possible to preserve the major stress of Japán in the derived form Jàpanése, because the
hypothetical form that would preserve it (*Japànése) would violate NO STRESS CLASH. It is more
important to obey NO STRESS CLASH and EARLY STRESS than STRESS PRESERVATION. This can
be expressed as the ranking of these constraints:

(36) NO STRESS CLASH, EARLY STRESS >> STRESS PRESERVATION

What happens in underived words in which 1ry stress falls later than the third syllable from the
beginning of the word? In these words EARLY STRESS and NO STRESS CLASH permit 2ry stress
placement either on the first or the second syllable of the word, but STRESS PRESERVATION
cannot decide between these two candidates23 since there is no base whose major stress should
be preserved. In this case the place of 2ry stress is unpredictable (and therefore lexical): compare
Wìnnipesáukee .$vHm?o?!rN9jh. and Monòngahéla .l?$mPMf?!gh9k?.-

22
Notice that it is the immediate base that counts: the base of orìginálity is oríginal not órigin.
23
unlike in the case of derived words
English Phonological analysis – Chapter 6: Word Stress Page 28 of 32

6.3.2.1.1 Summary of secondary stress patterns

2ry stress is
(i) calculated from the location of 1ry stress,
(ii) weight insensitive,
(iii) iterative (there may be more than one 2ry stress in a word),
(iv) stress preserving (if possible) and
(v) partially unpredictable in underived words (when 1ry stress is later than the third
syllable from the beginning of the word).

6.3.2.2 The location 3ry stress before the 1ry stress

A 3ry stressed syllable is a syllable with a full vowel not bearing 2ry or 1ry (i.e. major) stress.
In other words, it is a non-major-stressed syllable in which Vowel Reduction has not applied. So
the question is if we can predict when Vowel Reduction applies to a syllable without major
stress. The answer is mostly negative. There are some tendencies, but 3ry stress is mainly lexical.
Immediately preceding the 1ry stress, 3ry stress can occur (i) word-initially, e.g. tormént
.sN9!ldms. and (ii) word-medially còndensátion .$jPmcdm!rdHR?m.. In both positions there are
some tendencies that predict the presence/absence of reduction, but they are not categorical.
In word-initial closed syllables immediately preceding the 1ry stress Vowel Reduction may
be suspended (there is free variation between 3ry and zero stress), but in open syllables in the
same position there is a strong tendency for it to apply. We call this the closed-syllable tendency
for 3ry stress.

(37) i. initial open syllable ii. initial closed syllable

A.mérica )z!ldqHj? ?!ldqHj? Mon.tána lPm!szm? l?m!szm?
a.trócious )z!sqnTR?r ?!sqnTR?r Oc.tóber Pj!snTa? ?j!snTa?
lamént )kz!ldms k?!ldms segméntV rdf!ldms r?f!ldms
English Phonological analysis – Chapter 6: Word Stress Page 29 of 32

However, the closed syllable tendency is only a tendency because we can find words with initial
unreduced vowels in an open syllable e.g. va.cátion .udH!jdHRm+ u?!jdHRm.24, and with compulsory
vowel reduction in an initial closed syllable e.g. con.trástV .)jPm!sq@9rs+ j?m!sq@9rs..
Word-medial 3ry stress can be seen as the result of stress preservation, where the
unreduced vowel immediately preceding the 1ry stress preserves the vowel quality of the base
word in which the same syllable is major stressed, compare còndensátion .$jPmcdm!rdHRm.
(because of condénse .j?m!cdmr.) and còmpensátion .$jPlo?m!rdHRm. (from cómpensate; no
related word has stress on -pen-). However, stress preservation here is also only a tendency since
there are words which do not preserve the vowel quality of the major stress of their bases in this
way. Compare the words in (a), which do, with the similar words in (b), which do not.

(38) a. stress preservation: 3ry stress in the derived word

augmént N9f!ldms àugmentátion $N9fldm!sdHRm
impórt Hl!oN9s ìmportátion $HloN9!sdHRm
condémn j?m!cdl còndemnátion $jPmcdl!mdHRm

b. no stress preservation: zero stress in the derived word

consúlt j?m!rUks cònsultátion $jPmr?k!sdHRm
infórm Hm!eN9l ìnformátion $Hme?!ldHRm
consérve j?m!r29u cònservátion $jPmr?!udHRm

We have to conclude that 3ry stress is mainly lexical when it precedes 1ry stress.

24
According to Wells (2008), there is a 91% to 8% preference for .udH!jdHRm. in AmE and
there is a 61% to 39% preference for it even in BrE.
English Phonological analysis – Chapter 6: Word Stress Page 30 of 32

6.3.3 3ry stress after the primary stress

3ry stress is also unpredictable (lexical) when it is after the 1ry stress. We can find syllables with
unreduced vowels (a) in a non-final syllable immediately following the 1ry stress (39i), (b) in a
final syllable immediately following the 1ry stress (39ii) and (c) in a final syllable not
immediately following the 1ry stress (39iii):

(39)25 i. 130 órgasm !N9fzy?l, sárcasm !r@9jzyl

ii. 13 róbot !qntaPs, fórmat !eN9lzs, ellípsoid ?!kHorNHc
iii. 103 séparateV !rdo?qdHs, cáravan !jzq?uzm, récognize !qdj?fm`Hy

However, we can find syllables with reduced vowels in the same positions:

(40) i. 100 vanity !uzm?sh, elítism H!kh9sHyl,

ii. 10 clímate !jk`Hl?s, ábbot !za?s
iii. 100 séparateAdj !rdo?q?s, Ánglican !zMfkHj?m

Therefore, we conclude that 3ry stress following the 1ry stress is (also) essentially lexical.

25
Rarely, (39i) and (39iii) can combine with the result of having two 3ry stresses after the 1ry
stress, e.g. démarcate /!ch9l@9jdHs.
English Phonological analysis – Chapter 6: Word Stress Page 31 of 32

6.4. Checklist

i suprasegmental feature
i metrification
i weight sensitivity
i weight-insensitive
i extrametricality
i degrees of English stress
i pitch change
i loudness
i 1ry stress
i 2ry stress
i major stresses
i 3ry stress
i zero stress
i minor stresses
i reduced vowel
i full vowel
i Vowel Reduction
i Rhythm Rule
i Rhythmic Stress Deletion/Shift
i Iambic Reversal
i ‘no-pattern view’
i ‘pattern-with-exceptions view,’
i 1ry stress: the noun pattern
i 1ry stress: the verb pattern
i Noun pattern extrametricality
i Verb pattern extrametricality
i Main Stress Rule (MSR)
i Long Vowel Stressing (LVS)
i Alternating Stress Rule (ASR)
English Phonological analysis – Chapter 6: Word Stress Page 32 of 32

i .r. plus consonant clusters

i ‘prefixes’ of Latin origin
i conversion
i stress-placing suffixes
i stress-neutral suffixes
i types of stress-placing suffixes
i Pre-stressed 1 suffix
i Pre-stressed 1/2 suffix
i Pre-stressed 2 suffix
i Auto-stressed suffix
i NO STRESS CLASH
i EARLY STRESS
i STRESS PRESERVATION
i ranking of constraints
i closed-syllable tendency