Aspect selection in adult L2 Spanish and the

Competing Systems Hypothesis

When pedagogical and linguistic rules conflict*

Jason Rothman
The University of Iowa (USA)

Native-like use of preterit and imperfect morphology in all contexts by English

learners of L2 Spanish is the exception rather than the rule, even for successful
learners. Nevertheless, recent research has demonstrated that advanced English
learners of L2 Spanish attain a native-like morphosyntactic competence for the
preterit/imperfect contrast, as evidenced by their native-like knowledge of asso-
ciated semantic entailments (Goodin-Mayeda and Rothman 2007, Montrul and
Slabakova 2003, Slabakova and Montrul 2003, Rothman and Iverson 2007). In
addition to an L2 disassociation of morphology and syntax (e.g., Bruhn de Ga-
ravito 2003, Lardiere 1998, 2000, 2005, Prévost and White 1999, 2000, Schwartz
2003), I hypothesize that a system of learned pedagogical rules contributes
to target-deviant L2 performance in this domain through the most advanced
stages of L2 acquisition via its competition with the generative system. I call this
hypothesis the Competing Systems Hypothesis. To test its predictions, I compare
and contrast the use of the preterit and imperfect in two production tasks by na-
tive, tutored (classroom), and naturalistic learners of L2 Spanish.

Keywords: aspect, morphological performance, second language (L2)

acquisition, English/Spanish

1. Introduction

While first language (L1) acquisition is characterized by uniformity in route and

successful grammatical convergence, adult second language (L2) acquisition is
typified by variation in path and ultimate attainment, even under the most favour-
able of learning situations. This fact is incontrovertible, something which seem-
ingly obliges second language acquisition theories to address several related ques-
tions. One such question involves determining the significance of L2 variation and

Languages in Contrast 8:1 (2008), 74–106. doi 10.1075/lic.8.1.05rot

issn 1387–6759 / e-issn 1569–9897 © John Benjamins Publishing Company
 Aspect selection and the competing systems hypothesis 75

the weight assigned to it in terms of its implications for second language acquisi-
tion theories.
It is an observable fact that language use, albeit a first, second or third lan-
guage, varies at the macro (among people of the same language and dialects) and
at the micro level (for the same individual language user, both over time and in the
same moment). However, is L1 and L2 variation in language use the same? If so,
such variation does not need to be addressed by second language acquisition the-
ory specifically. In this case, variation can be accepted as a mere fact of language
without further consideration. However, if variations in L1 and L2 language use
manifest differently, the points at which variation between the two groups diverges
must be adequately addressed in theoretical terms.
Variation in the use of morphology is such a case in light of the following facts.
In L1 acquisition, there is an early mastery of bound morphology and a chrono-
logical emergence of inflection related to the development of syntax (Guasti 2002).
Ultimately, morphology comes to be used almost invariably by native adult speak-
ers, barring any particular pathology (e.g. Specific Language Impairment or severe
Down Syndrome; cf. Rondal 1993, van der Lely and Wexler 1998). Conversely,
L2 acquisition is typified by persistent problems in overt morphological use and
later mastery of bound morphology. And so, must this L1/L2 dissimilarity be un-
derstood as evidence of the so-called fundamental difference (cf. Bley-Vroman
1989, 1990) purported to underlie the competence of primary and non-primary
languages? If one takes the position that morphology drives syntactic compe-
tence, as Failed Features and Global Impairment models of L2 acquisition do (cf.
Beck 1998, Clahsen and Hong 1995, Franceschina 2001, Hawkins and Chan 1997,
Hawkins and Liszka 2003), then such differences can be interpreted accordingly.
However, there is good evidence for the dissociation of morphological and syn-
tactic development in adult L2 acquisition (cf. Bruhn de Garavito 2003, Lardiere
1998a and 1998b, 2000, 2005, Prévost and White 1999, 2000, Schwartz 2003), as
advocated by Full Access models (Schwartz and Sprouse 1996, White 1989, 2003).
In this case, L2 variation in morphological use cannot be considered sufficient evi-
dence that the processes of L1 and L2 acquisition are fundamentally different nor
can it be used to claim that L2 competence is morphosyntactically target-deviant.
Nevertheless, in the absence of L1 variation in the domain of overt morphologi-
cal use, L2 variation must be accounted for within a theoretical framework that is
both descriptively and explanatorily adequate.1
In this article, I address specific L2 variation in the use of preterit and im-
perfect morphology by highly advanced English learners of L2 Spanish. In light
of recent generative L2 research demonstrating that English learners of L2 Span-
ish and Portuguese acquire aspect-associated phrasal semantic entailments con-
ditioned upon the acquisition of L2 morphosyntactic features (Goodin-Mayeda
76 Jason Rothman

and Rothman 2007, Montrul and Slabakova 2003, Slabakova and Montrul 2003,
Rothman and Iverson 2007), I investigate the role that pedagogical rules of L1/
L2 grammar contrasts play in the L2 variation of preterit/imperfect morphologi-
cal use at the level of performance in advanced learners. Assuming full access to
Universal Grammar (UG), it is hypothesized that a system of learned pedagogical
rules contributes to target-deviant L2 performance through the most advanced
stages of L2 acquisition and thus explains some persistent performance problems
in this domain. To test this, I compare and contrast the use of the preterit and
imperfect in two production tasks for native Spanish, highly advanced classroom
learners, and highly advanced naturalistic learners of L2 Spanish.
This article is structured in the following manner. The background section is
divided into (i) an explanation of the morphosyntax of lexical and grammatical
aspect, highlighting their differences in English and Spanish and (ii) a discussion
of relevant L2 theories and studies on the L2 acquisition of the preterit and imper-
fect. Sections that describe the hypothesis, design and methodology of the present
study follow this. Lastly, I present the results, which are accompanied by a discus-
sion and conclusion section.

2 Background

In languages with fairly sophisticated aspectual systems, such as Spanish, aspect

plays an integral role in anchoring and determining the temporal role between
states and events (or eventualities) in the discourse. As Bonomi (1997) has point-
ed out, grammatical aspect also plays a deterministic role in relating appropriate
quantification over these eventualities. As a result, languages, like English, that do
not have articulated aspectual systems must dispense of other grammatical pro-
cesses to mediate the semantic intention of aspectual morphology. Moreover, the
fact that Spanish is a relatively morphologically rich language and English is not is
yet another confounding factor. That is, there are important morphosyntactic dif-
ferences between Spanish and English in terms of how their respective inflectional
systems encode person, number, tense, mood as well as aspect. These differences
are not just morphological and, as one might expect, thus have syntactic and se-
mantic consequences. In this section, I explain the differences between Spanish
and English with respect to grammatical aspect. In doing so, I highlight how as-
pect is realized differently in these languages and, therefore, what the English L2
learner must acquire to converge on a target-like grammar for L2 Spanish. I as-
sume a particular paradigm to explain these differences, the generative paradigm
(Chomsky 1981), as well as a particular model within this approach, namely the
Minimalist Program (Chomsky 1995, 2000)
 Aspect selection and the competing systems hypothesis 77

2.1 Grammatical and lexical aspect

Mastering the difference between the preterit and the imperfect involves much
more than learning the associated overt morpho-phonological forms (e.g., -é vs.
-aba). Selecting the preterit vs. imperfect, at least for native speakers, is consis-
tent with the aspectual perspective of the verb in context. The preterit form corre-
sponds to [+ perfective] aspect, which views the action from outside, such that it is
a closed, finished action. Accordingly, [+ perfective] aspect is bounded (Depraetere
1995) in that the action is seen as having an implicit beginning and end point.
(1) Roberto dijo la verdad.
“Roberto told the truth.”

According to Bonomi (1997), preterit morphology is the overt realization (identi-

fier) of an null existential operator. And so, the preterit correlates to only one read-
ing, that being an episodic one. This is not to suggest that the preterit cannot be
used after adverbial expressions with universal force (such as ‘always’ or ‘from time
to time’), but that the preterit itself signals a perspective that is only felicitous in a
context in which the temporal relation between eventualities is framed as closed
within the discourse.
Conversely, the imperfect, which corresponds to [- perfective] aspect, consid-
ers the action from within, whereby the action is ongoing, habitual, progressive or
not fully closed. Therefore, [-perfective] aspect is unbounded (Depraetere 1995)
as the focus is on the internal structure of the action without regard to any begin-
ning or end point.
(2) Roberto decía la verdad.
“Roberto was telling/used to tell the truth.”

It is important to note that the distinction between the preterit and the imperfect
is not a difference of tense, as both are used to indicate the past. In languages like
Spanish, it is somewhat difficult to tease this apart by looking only at the mor-
phological system, which seemingly encodes person, number, tense and aspect in
the same morphology. As can be appreciated by juxtaposing Spanish and Chinese
examples in (3) and (4) it can be seen that aspect is independent of tense, since
in Chinese the free morpheme le denotes [+ perfective] aspect regardless of time
(e.g., past, present or future).
(3) a. (Chinese) Zuótiān wŏ zuò wǎn le
Yesterday I do finish pfv. particle
zuòyè jiù lái le.
homework immanency adv. came pfv. particle.
“Yesterday, after I finished the homework, I came.”
78 Jason Rothman

b. (Spanish) Ayer después de que terminé la tarea,

Yesterday, after pro finish-1psg-pfv. the homework,
pro vine
I come-1psg-pfv
“Yesterday, after I finished my homework, I came.”
(4) a. (Chinese) Míngtiān wŏ zuò wǎn le
Tomorrow I do finish pfv. particle
zuòyè jiù lái.
homework immanency adv. come.
“Tomorrow, after I finish the homework, I will come.”

b. (Spanish) Mañana, cuando termine la tarea,

Tomorrow, when pro finish-1psg-subj. my home-work,
vendré
pro come-1psg-future
“Tomorrow, when I finish my home work, I will come.”

In both Spanish and English, lexical aspect is encoded in the lexical verb (e.g., [±
telic] features). That is, in both Spanish and English verbs are either telic or atelic
and this is learned lexically as part of verbal meaning.2 However, English and Span-
ish differ in terms of grammatical aspect realization. Following Giorgi and Pianesi
(1997) and others, it is assumed that grammatical aspect is realized as a functional
category, higher AspP (or outer AspP). Schematized in Figure 1, it can be seen
that both Spanish and English project higher AspP, English differs parametrically
in terms of which available features it associates with this functional category (see
Kempchinsky and Slabakova 2005 for greater detail).

Figure 1. AspP in English and Spanish.

Figure 1. AspP in English and Spanish.

Test 1: NS Deviation from NS Consensus by Individual

20%
 Aspect selection and the competing systems hypothesis 79

Spanish encodes both [± perfective] aspect features, which are checked in higher
AspP against preterit [+ perfective] or imperfect [- perfective] morphology. Con-
versely, English only has a [+ perfective] feature associated with this function-
al category. This means that the [± perfective] aspect distinction is not realized
morphologically in English. As a result, the simple past morphology of English is
most often associated with an episodic reading. English, therefore, often expresses
past characterizations or generalizations with modal verbs, such as ‘used’ to and
‘would’ or with the copula (‘be’) + gerund construction.
Although the simple past in English most often correlates to an episodic read-
ing, depending heavily on adverbial modifiers, as shown in (5) and (6), the past
morphology of English can readily convey an episodic or a habitual reading.
(5) I always walked to the park when I was younger.
(6) I walked to the park yesterday afternoon.

Conversely, in Spanish, only the imperfect (i.e. not the preterit) can correlate to
two types of readings, although they are different ones (habitual and progressive,
see Bonomi 1997). This can be a source of confusion for English learners of L2
Spanish since the Spanish form most closely related in function to the simple past
of English, the preterit, can only support an episodic reading, whereas the imper-
fect is poly-functional in that it can support different interpretations. This differ-
ence can be seen in the parallel sentences in (7) and (8).
(7) a. Yo compré una rosa para mi novia ayer.
b. #/*Yo compré una rosa a menudo para mi novia.
c. *Mientras compré una rosa para mi novia, ella llegó.
(8) a. I bought my girlfriend a rose yesterday.
b. I often bought a rose for my girlfriend.
c. #/* While I bought a rose for my girlfriend, she arrived.

These English/Spanish differences all fall out from particular-language differences

with respect to the composition of features they associate with the functional pro-
jection higher AspP.

2.2 Previous L2 studies

Although it was claimed in earlier work on adult acquisition of aspect (Coppieters

1987) that this distinction did not derive from principles of UG, recent theoretical
work has demonstrated that aspect distinction does derive from universal princi-
ples of grammar (cf. Bonomi 1997, Giorgi and Pianesi 1997, Lenci and Bertinetto
2000, Menéndez-Benito 2002, Schmitt 1996, Slabakova 2001, Smith 1991).3 As
80 Jason Rothman

discussed above and schematized in Figure 1, this grammatical aspectual contrast,

in accord with Minimalist Program assumptions (Chomsky 1995, 2000, Adger
and Smith 2005), is now understood in terms of features related to a functional
category. As a result, English learners of L2 Spanish must acquire a new L2 feature,
the [-perfective] feature associated with Spanish higher AspP, to achieve native-
like morphosyntactic competence in this domain. Within Minimalism, language-
to-language differences are isolated to the functional lexicon, which is to say, pa-
rameters are hypothesized to be set (or reset in L2 grammars) via the acquisition of
functional categories and their associated features. If adults continue to access UG,
as advocated by Full Access approaches (Schwartz and Sprouse 1996, White 1989,
2003), then English learners of Spanish should be able to acquire this contrast. Al-
though [± perfective] interpretable features have morpho-phonological corollaries
(i.e. preterit and imperfect Spanish verbal inflection), assuming a syntax-before-
morphology position, it is possible for an L2 learner to have a mental represen-
tation of these features that is not always evident in performance via invariably
proper use of their overt morpho-phonological forms (see e.g. Prévost and White
2000, Lardiere 2000, 2005). This is true because features are not morphological
forms themselves, but abstract properties that are mapped onto them (Schwartz
and Sprouse 1996). As a result, a function-to-form mapping problem, as suggested
by the Missing Surface Inflection Hypothesis (MSIH) (Prévost and White 2000),
may explain some errors in the use of Spanish aspectual morphology. This is espe-
cially important to consider in light of an important body of literature outside the
generative paradigm that has investigated the development of grammatical aspect
in L2 grammars of adult learners whose L1 is English.
The preterit and imperfect in adult L2 acquisition in general and in particular
L2 Spanish has been widely studied from non-generative perspectives to language
acquisition (cf. Andersen 1986, 1991, Bardovi-Harlig 2000, Andersen and Sharai
1994, 1996, Camps 2000, 2005, Hasbún 1995, Lafford 1996, Liskin-Gasparro 2000,
Salaberry 1999, 2000, 2002, Ramsay 1990). The majority of these studies focus
exclusively on the acquisition (emergence) of preterit and imperfect morphologi-
cal forms as they relate to interlanguage (IL) development following the Lexical
Aspect Hypothesis (Andersen 1986, 1991).4 Taken together, these studies have of-
fered discrepant results in terms of support for this hypothesis (see Bardovi-Harlig
2000, Salaberry 2000 for discussion). These studies examine the use of preterit and
imperfect morphemes in interlanguage development with different verb classes
and in particular discourse contexts. Based on analyses of L2 morphological usage
in these studies, the semantic value L2 learners assign to these morphemes is later
inferred. This methodology is perhaps unintentionally fallacious, related to the
longstanding discussion on performance — the production of language — versus
competence — the underlying knowledge of grammaticality — and the MSIH .
 Aspect selection and the competing systems hypothesis 81

Variation in language use is simply a fact of all output, non-native and native.
As a result, any given linguistic performance does not always accurately repre-
sent underlying competence. In the case of adult L2 language learners, however,
understanding the significance of differences between performance and compe-
tence seems to be more urgent. Nevertheless, in the absence of a uniform way
to gauge performance versus competence the usefulness of such terminology is
not immediately clear. After all, the likelihood that language use at any given mo-
ment, systematized or not, accurately depicts linguistic competence is tentative
at best. To assume that performance, even in the case that it demonstrates an ac-
curate system, is the window by which competence should be gauged weakens the
functional value of competence as a separate entity. This provision is heralded by
Full Access approaches to adult L2 acquisition, which assume a dissociation of
morphology and syntax in L2 acquisition (cf. Bruhn de Garavito 2003, Lardiere
1998a and 1998b, 2000, 2005, Prévost and White 1999, 2000, Schwartz 2003). In
light of the L2 tendency to variably use overt morphology despite sophisticated
knowledge of target-language syntax (see Lardiere 1998 a and 1998b, 2000, 2005,
2006), it is not clear that the over or under-usage of preterit or imperfect morphol-
ogy with a particular class of verbs in performance entails that an L2 learner’s
underlying competence in this domain is necessarily non-native-like. As a result,
it is somewhat dubious that analyses heavily dependent on overt production of
L2 morphology achieve their intended goal of determining underlying linguistic
competence.
Montrul and Slabakova (2003), Slabakova and Montrul (2003), Goodin-
Mayeda and Rothman (2007) and Rothman and Iverson (2007) have investigated
the possibility of L2 native-like morphosyntactic convergence in this domain,
as opposed to developmental patterns alone. Assuming the generative linguistic
paradigm, these studies demonstrated that advanced English learners of L2 Portu-
guese and Spanish achieve a competence level for the preterit/imperfect contrast
that is fundamentally native-like and that intermediate Portuguese and Spanish
learners also demonstrate knowledge of semantic entailments associated with the
acquisition of the necessary aspectual features.
Montrul and Slabakova (2003) tested L2 knowledge of available semantic in-
terpretation restrictions of the subject with generic pronouns associated with the
use of the preterit versus imperfect (cf. de Miguel 1992, Montrul and Slabakova
2003, Schmitt 1996), as in (9) and (1). In (9) only a specific subject interpretation
(speaker is implicated as a participant) is available as opposed to (9), in which both
a specific and a generic interpretation are available.
(9) Durante la dictadura, se vivió muy mal en Chile.
During the dictatorship, we lived very poorly. (+ specific)
82 Jason Rothman

*During the dictatorship, one (they) lived very poorly. (+ generic)

(10) Durante la dictadura, se vivía muy mal en Chile.
During the dictatorship, we lived very poorly. (+ specific)
*During the dictatorship, one (they) lived very poorly. (+ generic)
Montrul and Slabakova (2003) argue that the restriction of available semantic in-
terpretations in (9) as compared to (10) derives from a semantic universal (Chier-
chia 1995) that stipulates that habitual clause readings entail generic pronominal
subjects and episodic clause readings entail a specific pronominal subject. It is
argued that this principle must be accessed via the acquisition of the [- perfective]
feature associated with Spanish higher AspP.
Goodin-Mayeda and Rothman (2007) and Rothman and Iverson (2007) dem-
onstrated that advanced English learners of L2 Spanish and intermediate learners
of L2 Portuguese respectively acquired knowledge of another POS semantic en-
tailment associated with the acquisition of the preterit and the imperfect. Follow-
ing overt adverbial quantifiers (e.g., siempre que), there is an obligatory alternation
of [±accidental] interpretations of the preterit and imperfect respectively (cf. Lenci
and Berinetto 2000, Menéndez-Benito 2002), as in (11) and (12). Although (11) is
somewhat awkward without a context, it is completely (and only) felicitous with a
discourse context that denotes a sense of unintentionality.
(11) Siempre que fuimos a la universidad, estudiamos en la biblioteca.
“Whenever we went to the university, we ended up studying in the library.”
(12) Siempre que íbamos a la universidad, estudiábamos en la biblioteca.
“Every time we went to the university, we studied in the library.”

Except for the use of the preterit and imperfect, sentences (11) and (12) are iden-
tical. Moreover, given the presence of siempre que (an adverbial quantifier with
universal force) both sentences are necessarily interpreted as generalizations.
Nonetheless, a semantic distinction between the preterit and imperfect in this
context exists. The preterit, as in (11), within a proper context, is understood as
an unforeseen [+ accidental] generalization, whereas the imperfect, as in (9), is in-
terpreted as an expected [- accidental] generalization (Lenci and Bertinetto 2000,
Menéndez-Benito 2002).
The preterit/imperfect contrast is quite complex and involves convergence on
associated semantic knowledge that is not available from input and is not explic-
itly taught to tutored learners. Since English does not have the preterit/imperfect
distinction, L2 learners cannot transfer such knowledge from their L1. Under the
assumption that investigating associated semantic entailments provides stronger
evidence for determining linguistic competence than examining overt morpho-
logical use, generative studies have demonstrated target morphosyntactic compe-
 Aspect selection and the competing systems hypothesis 83

tence for the preterit/imperfect contrast in advanced and intermediate L2 Portu-

guese and Spanish.
In light of this, what explains target-deviant L2 use of the preterit and imper-
fect in these very same populations? While variation in language use is a ubiqui-
tous phenomenon for both L1 and L2 output, variation in bounded morphology
use is not terribly common among L1 speakers. Should it be assumed that L1/
L2 disparity in the use of the preterit and imperfect points to a necessary differ-
ence in competence? The Missing Surface Inflection Hypothesis (MSIH) claims
that the answer to this question is: not necessarily. However, the MSIH alone may
not be able to explain this disparity entirely, especially if one observes lingering
effects in highly advanced learners and it is discovered that there is a particular
pattern to these errors. The goal of this article is to shed some light on this line of
investigation by providing a complementary hypothesis, the Competing Systems
Hypothesis. In the next section, I detail the hypothesis, which attempts to integrate
the fact that, on the one hand, even highly advanced tutored English learners of L2
Spanish use the preterit and imperfect differently (although systematically) than
natives with the fact that, on the other hand, learners at this level have been shown
to have acquired associated POS semantic entailments.

3. Hypothesis

The linguistic concept of aspect is not introduced in the formal instruction of

Spanish as a foreign language, and with good reason. The formality of the above-
presented explanation of grammatical aspect is not appropriate for linguistically
naïve learners. As a result, the preterit/imperfect contrast is most often taught in
absolute terms and in line with English/Spanish contrast equivalencies. For ex-
ample, habitual acts, descriptions and generalizations, which are often expressed
via modal auxiliaries (i.e. would or used to) and the copula (‘be’) + gerund con-
struction in English, are taught as requiring the imperfect in Spanish. Equally,
one-time events, which most often take simple past morphology (‘-ed’) in English,
are taught to select the preterit in Spanish. In general, these explanations hold true,
but they are hardly absolute. Furthermore, the preterit/imperfect distinction of
particular verbs is more difficult to explain. Often, the preterit/imperfect distinc-
tion for these verbs is taught lexically in line with their translatability to English
equivalents. Needless to say, Spanish is not a translation of English. As a result, ex-
planations dependent on translation equivalency will, in due course, meet with the
inevitability of failing to account for what it seeks to explain with 100% accuracy.
For example, the verb saber is commonly taught as meaning ‘to know’ in the
imperfect and ‘to find out’ in the preterit (similar examples abound). Accordingly,
84 Jason Rothman

this translation based explanation is unable to account for the interpretation of

example (13), in which the preterit is not only used meaning ‘to know’ but also
follows siempre, one of the most frequently cited trigger words pedagogically as-
sociated solely with the imperfect.
(13) Siempre supe que un día me dejarías.
“I always knew that you would leave me.”

Although the preterit and imperfect forms of certain verbs are, by far, more likely
to be translated differently into English, it is inaccurate to claim that in every pos-
sible context, the preterit of these verbs will be equivalent to English translation
value X, whereas the imperfect of the same verbs is equivalent to value Y. In other
words, it is not that the meaning of the verb itself changes, but that the aspect may
be more accurately conveyed through a different English translation. Thus, instead
of saying that supe must mean ‘I found out’ and sabía must mean ‘I knew’, stu-
dents could be taught (I stress in an accessible way) that what is really happening
is that the verb saber, which is stative in the imperfect (and thus atelic), becomes
an achievement (and thus is telic) in the preterit. And so, in both forms the verb
still means ‘to know’ it is just that the preterit is inchoative in nature and marks
the beginning point of knowing (i.e. from that point on), which often happens to
nicely correlate to the English phrasal verb ‘to find out’. However, the meaning of
the verb itself does not actually change, as the translation technique of teaching the
difference would suggest.
Similar pedagogical conventions are prolific. For instance, particular words
are often taught to trigger either the preterit or imperfect exclusively, at least at ear-
ly levels of instructions. However, while this tendency holds true in general, it fails
to account for particular uses of both the preterit and the imperfect. In (14) and
(15), for example, the preterit is the most appropriate choice, notwithstanding the
accompanying trigger words mientras ‘while’ and todos los días ‘every day’, which
are often taught as being strictly associated with the use of the imperfect. So-called
trigger words do not automatically cue the preterit or the imperfect. Ultimately,
the most important considerations are the meanings that the speaker wishes to
communicate and the aspect that best fits this.
(14) Entre 1993 y 1995 mientras viví en Galicia, aprendí a hablar gallego.
“Between 1993 and 1995, while I lived in Galicia, I learned how to speak
Galician.”
(15) Durante aquella semana, les di de comer a sus perros todos los días.
“During that week, I fed their dogs every day.”
 Aspect selection and the competing systems hypothesis 85

Finally, similar to the emergence of the preterit and imperfect in children, L2

learners have been observed to use the preterit first with punctual verbs and the
imperfect with stative verbs. The Lexical Aspect Hypothesis (Andersen 1986,
1991) discussed above attempts to explain this and related patterns of morpho-
logical use via an L2 aspectual primacy position. However, it is possible that this
pattern has more to do with a pedagogical basis, which is to say the order of pre-
sentation and the repetition of ad hoc rules by which these verbal paradigms are
taught. The preterit is most often taught with achievements and accomplishments
verbs well before states, which are introduced first in the imperfect. Moreover, the
copula verbs ser and estar are often taught as defaulting almost exclusively to the
imperfect given their inherent function of describing. As a result, L2 learners tend
to have problems with sentences like (16) not in comprehension per se, but in their
production of similar sentences.
(16) No cabe el odio entre dos amigos que un día fueron novios.
“There should never be ill-will between two friends who were once lovers.”

There is good reason for the difficulty encountered by English-speaking learners of

Spanish as it relates to the acquisition of this aspectual distinction and their pro-
duction of preterit and imperfect morphology. Not only is aspect encoded differ-
ently in English and Spanish morphology, but, textbooks and formal instruction,
which must simplify the rules in order to make them manageable and accessible
even to beginning students, may promote faulty intuitions. In the present study, I
examine the usage of the preterit and the imperfect by highly successful English
learners of L2 Spanish to see how they perform in a general sense and in particular
with uses of the preterit and imperfect that seem to contradict pedagogical ex-
planations. Comparing advanced learners of two types, tutored (classroom learn-
ers) and untutored (naturalistic learners), to each other and to native speakers, I
explore and discuss the possibility that formal instruction plays a role in target-
deviant performance through the most advanced stages of adult L2 acquisition.
Based on the research I have discussed that shows advanced learners have so-
phisticated knowledge of related syntax and semantics despite errors in morpho-
logical production, I take the position that native-like ultimate attainment is pos-
sible at the underlying level. In addition to other interceding factors, the present
hypothesis claims that pedagogical simplifications as discussed above form a sepa-
rate system of learned knowledge and that this system can override linguistic com-
petence (the generative system) of the L2 learner at the level of performance. These
separate systems remain intact through advanced stages and essentially provide
two filters for linguistic performance. This means that even when interlanguage
reaches a steady-state in advanced learners that is representationally native-like
in particular domains, the learned knowledge system can intercede, especially in
86 Jason Rothman

highly monitored output, resulting in systematic errors. The prediction, therefore,

is that once grammatical properties have been acquired at the mental representa-
tion level this system can interfere with production, but not comprehension. This
explains how learners can have knowledge of semantic entailments of particular
morphosyntactic properties at proficiency levels where they do not use the corol-
lary overt forms entirely like natives. I call this hypothesis the Competing Systems
Hypothesis. If this hypothesis is on the right track and native-like attainment of
the preterit and imperfect in its underlying form (acquisition of the L2 feature
[-perfective]) is achievable by advanced English learners of L2 Spanish, then the
only errors in performance we should observe in highly successful learners should
coincide with contexts unaccounted for by pedagogical rules and language-to-lan-
guage translation equivalencies. Additionally, this hypothesis is falsifiable if the er-
rors observed are not isolated to tutored learners, since naturalistic learners would
not have such a separate system.

4. Methodology

4.1 Subjects

There are three participant groups. Group 1 is a Spanish native control group.
Group 2 is a group of highly advanced tutored English learners of L2 Spanish.
Group 3 is a group of highly advanced naturalistic English learners of L2 Spanish.
All of the subjects in groups 1 and 2 were sampled from university-level instruc-
tors of Spanish from two U.S. universities. The 20 participants in the native group
were from various Hispanic countries (Argentina, Chile, Colombia, Cuba, Mexico,
Peru, Puerto Rico and Spain) and came to the United States in their twenties, at the
earliest. While all members of the control group are Spanish-dominant, they are all
relatively fluent in English. It is important to note that dialectal variation among the
native speakers is not relevant for the purposes of this study, as aspect distinction
as in the preterit/imperfect contrast is not subject to notable dialectal differences5.
The average age of the NS group was 29.45 with a range of 21–48.
Group 2, the tutored L2 learners, was comprised of twenty subjects, with an age
range of 26–54 and a mean age of 34.5. Participants were selected if (a) English was
their L1, (b) they were not child bilingual speakers of another language, (c) they
reported having learned Spanish with 5 years of explicit instruction, (d) they teach
Spanish and are therefore very familiar with pedagogical Spanish grammars and
(e) a group of native speakers judged their Spanish as “near-native”. All but four of
the twenty subjects reported having lived in a Spanish speaking country for a year
or more (mean = 2.054 years, range = 0–7 years), while only one reported never
 Aspect selection and the competing systems hypothesis 87

having spent time abroad. The average total time since Spanish was first studied
in a consistent manner was 17.9 years. With the exception of one participant, all
began learning Spanish at least 11 years prior to data collection. While there is
intra-group difference in terms of time spent abroad and time since the individual
onset of learning Spanish (the latter of which understandably correlates to age) all
of the L2 learners received the same amount of formal education in Spanish. That
is, all L2 learners in this group reported five years of language courses (i.e., not
courses in literature, civilization or linguistics) in which they received pedagogical
instruction on the structure of the Spanish language, which include four years in
high school and one year in college.
Group 3, the untutored L2 learners, was comprised of 11 subjects, with an age
range of 34–62 and mean age of 45.3. All of the subjects learned Spanish in a natu-
ralistic environment as adults without ever receiving formal instruction in Span-
ish. Participants were selected for this group if (a) English was their L1, (b) they
were not child bilingual speakers of another language, (c) they never studied Span-
ish formally, (d) they learned Spanish exclusively via exposure to naturalistic input
(e) they resided in a Spanish speaking country (Chile, Mexico or Spain) at the time
of data collection and had lived there for at least 7 years (mean= 15.5, range 7–35
years) and (f) and a group of native speakers judged them as “near-native”.

4.2 Design

The present study employed two types of tests: a cloze paragraph multiple-choice
test and a fill-in-the-blank production task. The participants were asked to code
their tests with a 4–6 digit code of their choice. These numbers correspond to the
numbers reported in the following statistics for each individual.

4.2.1 Test 1.
Test 1, the cloze paragraph multiple-choice task, consisted of a well-known
fairytale written in Spanish adapted from a pedagogical website ([Link].
edu/~bknelson/exercises/ [Link]). The task provided binary choices, either
the preterit or imperfect, for 55 verbs from the story of Goldilocks and the Three
Bears, as exemplified in (17)
(17) 1. (Hubo, Había) una vez tres osos que 2. (vivieron, vivían) en el bosque: Papá
Oso, Mamá Osa, y Bebé Oso. Un día Mamá Osa 3. (hizo, hacía) una sopa
de arroz con pollo y 4. (puso, ponía) tres platos en la mesa. Como ya (fue,
era) mediodía, los osos 6. (se sentaron, se sentaban) para comer porque 7.
(tuvieron, tenían) muchísima hambre.
88 Jason Rothman

“Once upon a time three bears lived in the forest: Papa bear, Mama bear and
baby bear. One day Mama bear made a soup of rice and chicken and put
three bowls on the table. Since it was noon, the bears sat down to eat because
they were very hungry.”

The definition of a consensus among the control group for Test 1 was an agree-
ment (either preterit or imperfect selection) of at least 18 out of 20 of the control
group members for all stimuli. However, for 53 of 55 verbs the consensus was, at
least, 19 of 20. Furthermore, in 49 of 55 exemplars the native control agreement
reached 20 out of 20, or 100% agreement.

4.2.2 Test 2.
Test 2 was an aspect production task for which the subjects were required to fill in
blank spaces with either the preterit or the imperfect. Each of the twelve sentences
was specifically designed with a use of the preterit or imperfect that would con-
tradict traditional pedagogical explanations. For example, particular stative verbs
that often pose aspectual selection difficulties for English learners of L2 Spanish,
such as ser, haber, estar, tener and poder; less common uses of the preterit preced-
ing particular triggering adverbs (e.g., siempre and a menudo) as well as the imper-
fect with phrases like hacía años que and era obvio que were included.
(18) a. Hubo (haber) varias manifestaciones de los afro-americanos durante
los años sesenta en los [Link].
“There were various protests by the African-Americans during the
sixties in the United States.”
b. Tus padres siempre me cayeron (caer) bien hasta aquel día inesperado
que todo sucedió.
“You parents and I always got along well until that unforeseen day in
which it all happened.”

The definition of native consensus for Test 2 was agreement of at least 18 out of 20
for the control group, although it should be noted that in all but two of the twelve
stimuli, the native speaker consensus was 20 out of 20, or 100% agreement. In the
other two sentences there was 90% agreement.

5. Evaluation/results

This section is divided into two main sections, which correlate to each of the two
tasks. The results for each task are further divided into two subparts. First, I pro-
vide a descriptive analysis of the results for each group. Second, I provide a statisti-
cal comparison of the mean score performances of the tutored L2 learners and the
 Aspect selection and the competing systems hypothesis 89

naturalistic L2 learners against each other and against the native speakers for both
tasks. I employ a one-way ANOVA as an initial measure of inferential statistics,
which is followed up by Tukey pairwise comparisons. The statistics were conduct-
ed using the mean number correct for each group and the alpha was set at 0.05 for
both tests. Additionally, the data are presented in such a way to address whether
or not there is a pattern of tutored L2 learner divergence in preterit/imperfect con-
trast performance from both natives and naturalistic L2 learners.

5.1 Task 1: Descriptive analysis

The data are presented as number of correct selections of the preterit or imperfect
out of 55 total verbs (also provided as overall group percent deviation in Table 1,
2 and 3). A response was deemed “correct” if it matched the native consensus se-
lection of the preterit or imperfect. As can be seen in Table 1 and Figure 2 below,
there was very little intra-group variation for the NS group.
Table 1. Group 3: NS group average deviation = 0.72%
ID No. No. correct % divergence ID No. No. correct % divergence
of 55 of 55
1155 55 0% 5432 55 0%
1111 54 (#55) 1.8% 7771 55 0%
5454 55 0% 2391 55 0%
7331 55 0% 3116 55 0%
9910 55 0% 8553 54 (#21) 1.8%
7610 55 0 1899 55 0%
1234 54 (#3) 1.8% 1445 53 (#10 &11) 3.6%
1969 54 (#43) 1.8% 4792 55 0%
1028 53 (#10& 11) 3.6% 6619 55 0%
3778 55 0% 1976 55 0%

As can be seen in Table 2 and Figure 3 below, the tutored learners performed well
individually and as a group. The range of individual tutored learners’ divergence
from the native speaker use of the preterit and imperfect was between 0%-14.5%
(55–47 correct) on this task, which resulted in a group divergence rate of 4.6%
(52.6 of 55 correct). Ten of twenty tutored L2 participants performed native-like
on this task, that is, within the range of NS intra-group variation (at least 53 of 55
correct).
As can be seen in Table 3 and Figure 4 below, each of the naturalistic L2 learn-
ers performed within the range of individual NS intra-group variation, resulting in
a group divergence rate of 0.98% (an average of 54.6 of 55 correct).
 Figure 1. AspP in English and Spanish.
90 Jason Rothman

Test 1: NS Deviation from NS Consensus by Individual

20%
18%
16%
14%
12%
10%
8%
6%
4%
2%
0%
11

11

54

73

99

76

12

19

10

37

54

77

23

31

85

18

14

47

66

19
55

11

54

31

10

10

34

69

28

78

32

71

91

16

53

99

45

92

19

76
Figure2.2. NS
Figure NS Intra-group
Intra-group Deviation
Deviation from
from Consensus
Consensusby
byIndividual
Individual

Table 2. Group 2: tutored NNS group average deviation = 4.6%

ID No. No. correct % divergence ID No. No. correct % divergence

of 55 of 55
2104 51 7.2% 1719 54 1.8%
2105 52 5.5% 6045 47 14.5%
1806 53 3.6% 8021 55 0%
6304 52 5.5% 1517 53 3.6%
4585 54 1.8% 0609 54 1.8%
1092 48 12.7% 2183 50 9.1%
1028 52 6% 3729 52 5.5%
4609 55 0% 8954 54 1.8%
0317 52 6% 7976 54 1.8%
1226 51 7% 2791 53 3.6%

Table 3. Group 3: untutored NNS group average deviation = 0.98%

ID No. No. correct % divergence ID No. No. correct % divergence
of 55 of 55
4543 55 0% 1875 55 0%
1248 54 (#55) 1.8% 7223 53 3.6% (10&11)
3131 55 0% 8865 55 0%
2512 54 (#23) 1.8% 5426 53 3.6% (10&11)
9007 55 0% 7389 55 0%
1967 55 0%

While disaccord with NS consensus selection of the preterit and the imperfect was
low all around, it is patently clear that there was greater variation for the tutored L2
learner group. Of the 55 verbs, the tutored L2 learners chose the preterit or imper-
fect wrongly for 18 verbs. Of those 18 verbs, only 11 were missed by more than one
 Aspect selection and the competing systems hypothesis 91

Test 1: Tutored NNS Deviation from NS Consensus by Individual

Test 1: Tutored NNS Deviation from NS Consensus by Individual
20%
20%
18%
18%
16%
16%
14%
14%
12%
12%
10%
10%
8%
8%
6%
6%
4%
4%
2%
2%
0%
0%
21 2

21 2

18 1

63 6

45 4

10 1

10 1

46 4

31 3

12 1

17 1

60 6

80 8

15 1

60 6

21 2

37 3

89 8

79 7

27 2
04 10

05 10

06 80

04 30

85 58

92 09

28 02

09 60

7 17

26 22

19 71

45 04

21 02

17 51

9 09

83 18

29 72

54 95

76 97

91 79
4

1
Figure 3. Tutored NNS Deviation from NS Consensus by Individual
Figure3.3. Tutored
Figure Tutored NNS
NNS Deviation
Deviation from
from NS
NS Consensus
Consensus by Individual
by Individual

Test 1: Untutored NNS Deviation from NS Consensus by Individual

Test 1: Untutored NNS Deviation from NS Consensus by Individual

20%
20%
18%
18%
16%
16%
14%
14%
12%
12%
10%
10%
8%
8%
6%
6%
4%
4%
2%
2%
0%
0%
45 4

12 1

31 3

25 2

90 9

19 1

18 1

72 7

88 8

54 5

73 7
43 54

48 24

31 13

12 51

07 00

67 96

75 87

23 22

65 86

26 42

89 38
3

Figure4.4. Untutored
Figure Untutored NNS
NNS Deviation
Deviation from
from NS
NS Consensus
Consensus by
by Individual
Individual
Figure 4. Untutored NNS Deviation from NS Consensus by Individual

person. All but one of the items that the tutored L2 learners missed involved one
of the following verbs: ser, estar, poder, querer, haber, tener, saber. All of these verbs
were either copula verbs or among the verbs taught lexically in terms of English/
Spanish translation equivalencies. Furthermore, every appearance of the stative
verb ser except one had varying results among tutored NNSs. We would also like to
bring attention to verbs (10) and (11) from the story, reproduced in (19) below.
(19) Entonces Bebé Oso y Mamá Osa (10) (quisieron, querían) comer la sopa pero
no (11) (pudieron, podían) porque…
“Then Baby Bear and Mama Bear (attempted, wanted) to eat the soup, but
they (could not, were unable) because…”

For verb (10), sixteen of the twenty (80%) tutored L2 learners used the target-devi-
ant imperfect form of querer in sharp contrast with both the NS and naturalistic L2
92 Jason Rothman

Table 4. Target-deviancy by Item for Test

Verb Context NS NS Tut. Tut. Nat. Nat.
Item # Pret. Imp. L2 L2 L2 L2
Pret. Imp. Pret. Imp.
3 Un día Mamá Osa 95% 5% 100% 0% 100% 0%
(hizo, hacía) una sopa…
5 Como ya (fue, era) mediodía…… 0% 100% 10% 90% 0% 100%
6 Los osos (se sentaron, se sentaban) 100% 0% 95% 5% 100% 0%
7 porque (tuvieron, tenían) muchísima hambre 0% 100% 5% 95% 0% 95%
10 Entonces Mamá Osa y Bebé oso 90% 10% 20% 80% 82% 18%
(quisieron, querían) comer la sopa
11 pero no (pudieron, podían) porque.. 90% 10% 90% 10% 82% 18%
12 (estuvo, estaba) tan caliente como la sopa… 0% 100% 5% 95% 0% 100%
15 (Fue, Era) un día bonito del verano…. 0% 100% 10% 90% 0% 100%
21 Ella siempre (jugó, jugaba) cerca de 5% 95% 0% 100% 0% 100%
23 Ricitos de Oro (se acercó, se acercaba) a la casa 100% 0% 100% 0% 9% 91%
25 pero no (hubo, había) nadie en el interior 0% 100% 10% 90% 0% 100%
31 toda la sopa del plato pequeño porque (estuvo, 0% 100% 5% 95% 0% 100%
estaba) perfecta…
34 Le (gustó, gustaba) más la silla cómoda 100% 0% 80% 20% 100% 0%
35 pero la silla (fue, era) muy pequeña para ella. 0% 100% 5% 95% 0% 100%
39 y (tuvo, tenía) ganas de dormir. 0% 100% 10% 90% 0% 100%
41 (fue, era) tan cómoda que … 0% 100% 15% 85% 0% 100%
43 los tres osos (volvieron, volvían) del bosque 95% 5% 100% 0% 100% 0%
44 Cuando vieron la puerta abierta…los osos 100% 0% 85% 15% 100% 0%
(supieron, sabían) que alguien
45 (estuvo, estaba) en su casa en ese momento. 0% 100% 5% 95% 0% 100%
48 Qué sorpresa (tuvieron, tenían) allí 100% 0% 65% 35% 100% 0%
49 (Hubo, había) una chica en la cama de Bebé 0% 100% 5% 95% 0% 100%
Oso
55 siguió una senda que le (fue, era) conocida 5% 95% 20% 80% 9% 91%

learner groups, which used the preterit. We note that two NSs and two naturalistic
L2 learners also chose the imperfect for exemplar (10); however, they, unlike the
majority of the tutored L2 learners, chose the imperfect in (11) as well, rendering
their choice viable.
Again, I highlight the fact that there was relatively little group variation for
this task. However, in the case tutored learners’ selection of the preterit and im-
perfect varied from the native speaker group and the naturalistic L2 group there
was a pattern, which coincided with pedagogical simplification. That is, all of the
verbs in the sentences that have some degree of tutored L2 learner target-deviancy
are taught lexically in terms of English translation equivalents or they are copula
 Aspect selection and the competing systems hypothesis 93

verbs, which are taught as defaulting almost exclusively to the imperfect. Table 4
below shows the target-deviant responses for all three groups in context. As can
be seen in Table 4, the only group that demonstrates significant target-deviant re-
sponses for these verbs is the tutored L2 learners.

5.2 Task 1: Statistical analysis

Although there is little L1/L2 deviation for this task overall, I employed a one-way
ANOVA comparison of all three groups to determine if there were any statistically
significant difference between the three groups’ performances as they compared to
each other. The ANOVA revealed that between these three groups there was sig-
nificant variation (f=14.53, p<0.001). As a result, I followed up the ANOVA with a
series of Tukey pairwise comparisons for each relevant group comparison (NS vs.
tutored L2, NS vs. naturalistic L2 learners and tutored vs. naturalistic L2 learners).
A comparison of mean scores for the naturalistic L2 learners (54.46, sd=.82) and
NS group (54.6, sd=.68) reveals no significant differences (p<.005) in aspect selec-
tion between the two groups. On the other hand, the difference between the mean
score of the tutored L2 learner group (52.3, sd= 2.13) and the mean score for the
NS control yields a statistically significant number (p>.005). Comparing the mean
scores of the tutored L2 learner group versus the L2 naturalistic group reveals
a statistically significant difference between these two groups as well ( p>.005),
which is not surprising given the fact that the naturalistic L2 learners performed
like the native Spanish speakers.

5.3 Task 2: Descriptive analysis

Task 2 was a fill-in-the-blank task designed to elicit productions of the preterit

or the imperfect that did not correspond to pedagogical rules/expected English
translations. As can be seen in Table 5 and Figure 5 below, the NS group dem-
onstrated virtually no variation in choosing either the preterit or the imperfect
for the sentences in Task 2. As was the case in Task 1, the results are reported as
number of correct responses, which was determined against the native consensus
selection of the preterit or imperfect for each verb.
Task 2 showed greater deviation for the tutored L2 learner group. As can be
seen in Table 6 and Figure 6, L2 learners produced the preterit and imperfect dif-
ferently than the natives 25% of the time. As can be seen, this 25% is really indica-
tive of an averaging of a wide range of individual L2 performances, which spanned
0% to 58.3% divergence. In fact, four of the twenty learners (4585, 1028, 4609,
1517) perform within the range of native speakers on this task.
94 Jason Rothman

Table 5. NS group average deviation = 1.66%

ID No. No. correct % divergence ID No. No. correct % divergence
of 12 of 12
1155 11 (#10) 8.3% 5432 12 0%
1111 12 0% 7771 12 0%
5454 12 0% 2391 12 0%
7331 12 0% 3116 12 0%
9910 12 0% 8553 12 0%
7610 12 0% 1899 12 0%
1234 12 0% 1445 12 0%
1969 11 (#10) 8.3% 4792 12 0%
1028 12 0% 6619 11 (#8) 8.3%
3778 12 0% 1976 11 (#8) 8.3%

Test 2: NS Deviation from NS Consensus by Individual

Test 2: NS Deviation from NS Consensus by Individual
60%
60%
55%
55%
50%
50%
45%
45%
40%
40%
35%
35%
30%
30%
25%
25%
20%
20%
15%
15%
10%
10%
5%
5%
0%
0%
19 19
23 23

31 31

85 85

18 18

14 14

47 47

66 66
19 19

10 10

37 37

54 54

77 77
76 76

12 12
11 11

11 11

54 54

73 73

99 99

19 19

76 76
16 16

53 53

99 99

45 45

92 92
32 32

71 71

91 91
10 10

34 34

69 69

28 28

78 78
55 55

11 11

54 54

31 31

10 10

Figure 5. NS individual results

Figure5.5. NS
Figure NS individual
individual results
results

Test 2: Tutored NN S Deviation from NS Consensus by Individual

Test 2: Tutored NN S Deviation from NS Consensus by Individual
60%
60%
55%
55%
50%
50%
45%
45%
40%
40%
35%
35%
30%
30%
25%
25%
20%
20%
15%
15%
10%
10%
5%
5%
0%
0%
21 21

21 21

18 18

63 63

45 45

10 10

10 10

46 46

31 31

12 12

17 17

60 60

80 80

15 15

60 60

21 21

37 37

89 89

79 79

27 27
04 04

05 05

06 06

04 04

85 85

92 92

28 28

09 09

7 7

26 26

19 19

45 45

21 21

17 17

9 9

83 83

29 29

54 54

76 76

91 91

Figure6.6. Tutored
Figure Tutored NNS individual
individual results
Figure 6. Tutored NNS
NNS individual results.
results
 Aspect selection and the competing systems hypothesis 95

Table 6. Tutored NNS group average deviation = 25%

ID No. No. correct % divergence ID No. No. correct % divergence
of 12 of 12
2104 10 16.7% 1719 8 33.3%
2105 9 25% 6045 9 25%
1806 10 16.7% 8021 10 16.7%
6304 7 41.7% 1517 11 8.3%
4585 11 8.3% 0609 9 25%
1092 8 33% 2183 6 50%
1028 11 8.3% 3729 7 41.7%
4609 12 0% 8954 10 16.7%
0317 8 33.3% 7976 10 16.7%
1226 5 58.3% 2791 8 33.3%

In Table 7 and Figure 7, the individual and group data for the naturalistic L2
learner group are provided. Unlike the tutored L2 learner group, there is virtually
no difference in performance between the NS control and naturalistic L2 learn-
ers, individually or as a group. That is, these naturalistic L2 learners produced the
preterit and imperfect with native-like accuracy.
The sentences in Test 2 were carefully designed to be inconsistent with peda-
gogical explanations of the preterit/imperfect contrast taught in language course
to English learners of L2 Spanish. In light of this, we turn to the tutored learners’
responses to individual sentences in an effort to further investigate whether or not

Test 2: Untutored NNS Deviation from NS Consensus b

60%
55%
50%
45%
40%
35%
30%
25%
20%
15%
10%
5%
0%
45

12

31

25

90

19

18

72

88

54

73
43

48

31

12

07

67

75

23

65

26

89

Figure7.7. Untutored
Figure Untutored NNS
NNS Individual
Individual Results
Results.

Test 2: Rate of NNS Deviation by Question

80%
96 Jason Rothman

Table 7. Untutored NNS Group Average Deviation = 2.27%

ID No. No. correct % divergence ID No. No. correct % divergence
of 12 of 12
4543 11 (#8) 8.3% 1875 12 0%
1248 12 0% 7223 12 0%
3131 12 0% 8865 12 0%
2512 12 0% 5426 11 8.3% (#8)
9007 12 0% 7389 12 0%
1967 11 (#10) 8.3%

a pattern of L2 target-deviancy can be explained in terms of pedagogical oversim-

plification, as can be seen in Figure 8 and Table 8 below.
The most problematic item on this test, number 3, repeated as (20) below, with
a tutored L2 learner group divergence rate of 80% involved the use of the preterit
following the word siempre, perhaps the most common word taught as a trigger as-
sociated strictly with the use of the imperfect. L2 tutored learners overwhelmingly
answered with the imperfect (16 of 20 L2 learners) while both the NS group and
naturalistic L2 learners completed this sentence with the preterit 100% of the time.

Table 8. Target-deviancy by Item for Test 2.

Item # Context NS NS Tut. Tut. UnT. UnT.
Pret. Imp. NNS NNS NNS NNS
Pret. Imp. Pret. Imp.
1 …no cabe el odio entre dos amigos que un día 100% 0% 60% 40% 100% 0%
fueron (ser) novios.
2 Era obvio que escuchaba (escuchar) la radio 0% 100% 10% 90% 0% 100%
3 Tus padres siempre me cayeron (caer) bien 100% 0% 20% 80% 100% 0%
hasta aquel día…
4 Hubo (haber) varias manifestaciones durante 100% 0% 70% 30% 100% 0%
los años 60s…
5 Yo nunca pensé (pensar) que fueras a dejarme.100% 0% 65% 35% 100% 0%
6 Miguel se quedó (quedarse) mudo……. 100% 0% 95% 5% 100% 0%
7 ….[mudo] en cuanto escuchó (escuchar) los 100% 0% 85% 15% 100% 0%
acontecimientos del incidente.
8 Era (ser) un error decirle mi contraseña, por 10%* 90*% 40% 60% 18%* 82%*
eso no se la dije.
9 Hacía años que no comía tacos. 0% 100% 20% 80% 0% 100%
10 …¿Por qué dijiste que……. 10% 90% 70% 30% 9% 91%
11 ..[que] me amabas cuando estabas abrazán- 0% 100% %0 100% 0% 100%
dome y besándome….
12 …pero al final me obligaste a irme. 0% 100% 5% 95% 0% 100%
 5%
0%

45

12

31

25

90

19

18

72

88

54

73
43

48

31

12

07

67

75

23

65

26

89
Aspect selection and the competing systems hypothesis
Figure 7. Untutored NNS Individual Results 97

Test 2: Rate of NNS Deviation by Question

80%

70%

60%

50%

40%

30%

20%

10%

0%
1 2 3 4 5 6 7 8 9 10 11 12
Question Number

Figure 8. Tutored NNS Deviation by Item

Figure 8. Tutored NNS Deviation by Item

(20) Tus padres siempre me ____cayeron [+perfective] (caer) bien hasta aquel día
inesperado en que todo sucedió.
“I always liked your parents until that unexpected day when everything
occurred.”

The next most disparate responses, numbers 1, 4, 5, and 8 also involved the use of
preterit with trigger words such as siempre and/or verbs like ser and haber.
Interestingly, items 11 and 12, which served as fillers utilizing verbs that are
not introduced lexically for the imperfect/preterit distinction (amar and obligar),
were the only sentences for which there was no tutored L2 learner divergence from
NS consensus responses.
We note that there is slight intra-group variation in the NS control and natu-
ralistic L2 learner group response for sentence 8 reproduced here in (21). How-
ever, this variation is different as compared to the tutored L2 learners.
(21) ________ (ser) un error decirle mi contraseña, por eso no se la dije.
“________ (be) an error to tell him/her my password, so I did not tell him/
her it.”

Tutored L2 learners diverged from the NS consensus by producing the preterit as

opposed to the imperfect (fue versus era), which actually results in a truth-value
semantic violation. Conversely, whenever a native or untutored L2 learner pro-
vided an answer that was different than the NS consensus of the imperfect it was
either the conditional (sería) or the conditional perfect (habría sido). While both
98 Jason Rothman

alternative forms are possible using either form represented an instance of incom-
pliance with the directions of the task and is thus counted as divergent.

5.4 Task 2: Statistical analysis.

In the previous section, we reported a 25% tutored L2 difference and a 2.27%

naturalistic L2 difference in preterit vs. imperfect production as compared to na-
tive speakers consensuses. To check the statistical relevance of these differences I
employed a one-way ANOVA comparison for all three groups. It became imme-
diately clear that there were statistically significant differences between the three
groups’ performances as they compared to each other (f= 34.07, p= 0.000). In an
effort to tease apart which of the possible group-to-group comparisons resulted in
significant differences, I followed up the ANOVA with a series of Tukey pairwise
comparisons for each relevant group comparison (NS vs. tutored L2, NS vs. natu-
ralistic L2 learners and tutored vs. Naturalistic L2 learners). A comparison of the
mean score of the naturalistic L2 learner (11.73, sd= .47) and the mean score of
the NS group (11.8, sd=.41) revealed no significant differences (p<.005) in aspect
production between these two groups. On the other hand, the mean score diver-
gence between the tutored L2 learner group (8.95, sd.=1.82) and the NS control
(11.8, sd=.41) yielded a statistically significant difference (p>.005). Equally, com-
paring the mean score of the tutored L2 learner group (8.95, sd.=1.82) versus the
L2 naturalistic group (11.73, sd= .47) revealed a statistically significant difference
between these two groups (p>.005), on par with the tutored L2 vs. native control
difference.

5.5 Discussion of Results

Bringing together the results of both tasks, a pattern to tutored L2 learner target-
deviancy was revealed. In line with the predictions of the Competing Systems
Hypothesis, tutored L2 learners demonstrated variation in selecting between the
preterit and imperfect in three contexts only: (a) with commonly used stative
verbs (e.g. ser, haber), (b) verbs whose preterit and imperfect contrast is taught
lexically by means of English translation equivalents (e.g., sabía vs. supe; quería
vs. quise) and (c) after adverbials that are taught as default triggers to either the
preterit or the imperfect form (e.g., siempre, cuando). Conversely, the naturalistic
learners of L2 Spanish performed like the NSs on both tasks. It is reasonable to
suppose that formal instruction is the variable between the two L2 learner groups
that accounts for the difference in their performance. If instruction is the criti-
cal variable, it is logical to believe that oversimplified pedagogical rules taught to
L2 learners form a system of linguistic knowledge that they use to monitor their
 Aspect selection and the competing systems hypothesis 99

output and, thus, affects their performance. The fact that not even one naturalistic
learner demonstrated the observed pattern of target-deviancy that the vast major-
ity of tutored learners did is explained by the fact that they, not having received
pedagogical instruction, have no such system of learned linguistic rules of explicit
comparisons to their L1, English. It should be noted that four of twenty tutored L2
learners performed completely native-like on both tasks, which suggests that the
encroachment of pedagogical explanations on linguistic performance is not an in-
evitability for all L2 learners. Comparing variables such as time lived in a Spanish
speaking environment, sex, age and years since Spanish was first studied revealed
that there is no means by which these four learner could be reliably differentiated
from their peers.
Based on previous research (Goodin-Mayeda and Rothman 2007, Montrul
and Slabakova 2003, Slabakova and Montrul 2003, Rothman and Iverson 2007),
the present study began with the assumption that adults have full access to UG
in general and that in particular English-speaking learners of L2 Spanish at the
advanced level acquire the necessary morphosyntactic feature [- perfective] to at-
tain native-like competence in the domain of [± perfective] aspect distinction (i.e.
native-like competence for the preterit/imperfect contrast). While I also argued
that performance, even when it shows systematicity, is not the most accurate tool
to gauge underlying linguistic competence, I note, nonetheless, that the whole of
the evidence provided for both L2 learner groups herein is consistent with the pos-
sibility of L2 native-like morphosyntactic competence. While the performance of
several of the tutored L2 learners diverged from that of NSs (and the naturalistic
L2 learners), there was a clear pattern to this target-deviancy, which I have argued
stems from a separate system of learned linguistic knowledge. I join Schwartz
(1993), Schwartz and Gubala-Ryzak (1992) among others in positing that only
natural positive evidence leads to grammatical competence. In other words, while
explicit positive evidence (pedagogical rules), constitute an indispensable facet of
the most common situation of second language learning (the classroom), these
rules do not lead to underlying linguistic competence. Rather, these rules serve the
communicative component crucially, in that they provide L2 learners with both a
learning strategy and a mechanism that facilitates communication throughout in-
ter-language development. It is reasonable to believe that these pedagogical rules
are consciously accessed in discourse as an output monitor by many L2 learners,
resulting in surface morphological errors despite a morphosyntactic competence
that is fundamentally native-like. If this is true, the observed pattern of L2 devia-
tion is explained accordingly.
The role of instruction, therefore, is understood as a facilitator and motivator
of language acquisition, especially for people that are not learning a target L2 in a
native speech community. However, instruction is clearly neither a sufficient nor
100 Jason Rothman

a necessary variable to acquire an L2 as an adult, otherwise, naturalistic learners

would be unable to acquire properties different from their L1. The data presented
here suggest that pedagogical explanations would greatly benefit from under-
standing and keeping in mind the grammatical underpinnings to the preterit and
imperfect and, to the extent that this is possible, incorporate linguistic rules into
teaching. However, assuming full access to UG, the key variable, that is, the only
compulsory variable is sufficient access to quality input. The benefit of teaching the
preterit and imperfect in a linguistically savvy way is not because such an explicit
explanation is necessary, but that pedagogical rules based on linguistic rules would
not differ from what the learner’s actual grammar dictates. As a result, competition
between these separate systems would be rendered negligible to null.
Since the data show L2 target-deviancy primarily with stative verbs, it may be
suggested that Andersen’s (1986, 1991) Lexical Aspect Hypothesis (LAH) equally
explains the observed target-deviancy in tutored learners’ performance. Impor-
tantly, however, target-deviancy in this study is not limited exclusively to stative
verbs. Furthermore, the LAH assumes that L2 learners will eventually learn that
stative verbs can be used with preterit morphology. Given the proficiency level
of our tutored L2 participants, as well as the time it has been since they began
learning Spanish (average= 17 years), it seems reasonable to argue that they have
achieved a steady-state for the L2 grammar, at which point the LAH does not pre-
dict the observed target-deviancy. Even if an argument could be made that these
learners are still in the process of learning and that this explains why they perform
the way they do in accord with the predictions of the LAH, one would still have
to address why the naturalistic learners do not manifest the same performance
pattern. That is, the LAH is purported to apply to L2 learners in general; there is
no distinction made in terms of its applicability to tutored versus naturalistic L2
learners. In addition, there is no basis on which to distinguish the tutored learn-
ers from the naturalistic learners in this study (e.g., years of exposure to Spanish),
nor is there a basis on which to argue a correlation between years of exposure to
L2 Spanish and the attainment of an end-state grammar of Spanish. It is also im-
portant to recall that the performance of several L2 learners in the tutored group
was identical to that of the NSs on both tests. As previously discussed, the lin-
guistic background of these individual participants compared to other tutored L2
learners who performed in a variably target-deviant manner, it becomes clear that
such a correlation is not tenable. This is especially telling in light of the fact that
none of the untutored learners demonstrated target-deviancy even though some
had been learning Spanish, albeit in a natural environment, for only seven years.
Taken together, all the evidence points to instruction as the significant variable
that explains the difference of the two L2 groups and supports the tenability of the
Competing Systems Hypothesis.
 Aspect selection and the competing systems hypothesis 101

6. Conclusions

The present study has provided evidence in support of the view that pedagogi-
cal rules of oversimplification can result in L2 performance variation, perhaps in-
definitely. It goes without saying that teaching adult learners of a foreign language
in an explicit manner is beneficial for their success. Insofar as adult L2 learners
transfer and use their L1 grammar to make hypotheses about the target L2 initially
and throughout inter-language development, it is logical to teach these learners
in terms of explicit L1/L2 contrasts when such comparisons facilitate the acqui-
sition of the L2 grammar. However, in the case that the contrastive explanation
does not accurately depict the L1/L2 difference in all contexts, lingering effects
on performance may ensue for at least some learners. In providing evidence that
certain pedagogical conventions may add to target-deviant performance for many
L2 learners, the goal is to highlight the long-term effects of such practices and to
suggest that L2 learners will benefit from instruction on linguistic rules that seek
to describe underlying linguistic competence. Moreover, the Competing Systems
Hypothesis adds to other recent proposals from the generative L2 paradigm that go
beyond the untenable notion of UG inaccessibility to explain particular instances
and causes of L2 target-deviancy (cf. Prévost and White 2000, Slabakova 2006,
Goad and White 2006 among many others). In the present situation, instructing
L2 learners to fully consider the entire context as it relates to what [± perfective]
aspect provides to the semantic value of any given sentence could nullify the nega-
tive effects that pedagogical rules and misconceived L1 translation equivalencies
engender at the level of performance.
To test the verifiability of the Competing Systems Hypothesis as well as to iso-
late other areas of grammar in which native-like performance may be affected by a
separate system of learned pedagogical rules, additional studies are warranted. The
focus of this study was limited to the domain of the preterit/imperfect contrast;
however, if the hypothesis is on the right track one should observe similar effects
in other domains. In fact, any domain in which pedagogical rules do not coincide
with underlying linguistic competence should be subject to similar effects.

Notes

* The data for the tutored L2 learner group and the design of the empirical tasks were done
with Liz Goodin-Mayeda for which I owe her a debt of gratitude. The data for the tutored learn-
ers and some of the native speakers was presented with her at AATSP 2005 and PAMLA 2005.
I am grateful for many comments from several colleagues; however, I am especially indebted
to Judith Liskin-Gasparro for her detailed comments and close reading, as well as Roumyana
102 Jason Rothman

Slabakova and Paula Kempchinsky for discussion about their work on aspect and suggestions
for parts of the analysis I provide. I am also grateful for the insightful suggestions of LiC anony-
mous reviewers. Despite this help, any and all errors are entirely my own.

1. It should be noted that there has been much recent discussion of variation at a macropara-
metric and a microparametric level, (where macro and micro are understood differently from
their previous use in this section) from formal approaches to language. For example, Adger
and Smith (2005) argue that the architecture of the Minimalist Program (MP) in particular is
apt to account for micro-grammatical variability due to the way it views language-to-language
differences. In light of the MP’s manipulation of grammatical features, which essentially allows
variable phonological outputs with the same semantic interpretation, variation in micro- and
macro-language use can be accounted for without assuming a variationist approach. Within L2
acquisition Sorace (2000, 2003, 2004, 2005), for example, have paid much attention to isolating
the source of variation in L2 language use and have argued that vulnerable interfaces (e.g., syn-
tax/semantics and syntax/pragmatics interfaces) are the source of much difficulty and perhaps
fossilization.

2. This is not to suggest that verbs that are inherently atelic, like correr ‘to run’ cannot take on
a telic meaning in certain contexts, for example when proceeded by DPs such as una milla ‘a
mile’.

3. The difference between Spanish and English in this respect is much more complicated lin-
guistically than space permits us to explore here. For a more theoretical explanation of how
Spanish and English aspectual encode different meanings, for example, see Kempchinsky and
Slabakova (2005).

4. The Lexical Aspect Hypothesis (see Bardovi-Harlig, 2000 for a discussion of relevant litera-
ture) purposes aspectual primacy in L2 acquisition whereby verbal morphology is hypothesized
to initially mark inherent aspect distinction only. Eight stages for the emergence of preterit and
imperfect morphological usage associated with the type of verb (states, activities, achievements
and accomplishments) are proposed. It is argued that either the preterit or imperfect morphol-
ogy is used exclusively for each verbal class initially. It is hypothesized that the L2 learner will
eventually learn that both preterit and imperfect morphology can be used with each class of
verb.

5. In Spain, the present perfect in lieu of the preterit is often used, especially in discourse, to re-
port past actions that have recently taken place and have [ + perfective] aspect. Therefore, some
of the sentences in the study could have been answered with this tense as opposed to the preterit.
However, given the binary choices of the preterit and imperfect, in such cases, the preterit is the
only logically available choice. Although 4 of the 20 natives in this study were from Spain, no
native speaker participants used or reported preferring the present perfect instead of the two
choices available to them.
 Aspect selection and the competing systems hypothesis 103

References

Adger, D. and Smith, J. 2005. “Variation and The Minimalist Programme”. In Syntax and Varia-
tion: Reconciling the Biological and the Social, L. Cornips and K. Corrigan (eds), 149–172.
Amsterdam and Philadelphia: John Benjamins.
Andersen, R. 1986. “El desarrollo de la morfología verbal en el español como segundo idioma”.
In Adquisición del lenguaje/Adquisição da linguagem, J. Meisel (ed), 115–138. Frankfurt:
Vervuert.
Andersen, R. 1991. “Developmental Sequences: The Emergence of Aspect Marking in Second
Language Acquisition”. In Crosscurrents in Second Language Acquisition and Linguistic The-
ories, T. Huebner and C. A. Ferguson (eds), 305–324. Amsterdam and Philadelphia: John
Benjamins.
Andersen, R. and Sharai, Y. 1994. “Discourse Motivations for some Cognitive Acquisition Prin-
ciples”. Studies in Second Language Acquisition 16:133–56.
Andersen, R. and Sharai, Y. 1996. “The Primacy of Aspect in First and Second Language Ac-
quisition: The Pidgin-Creole Connection”. In Handbook of Second Language Acquisition, R.
Bhatia and W. Ritchie (eds), 527–570. Malden, NJ: Academic Press.
Bardovi-Harlig., K. 2000. Tense and Aspect in Second Language Acquisition: Form, Meaning, and
Use. Malden, MA: Blackwell.
Beck, M.-L. 1998. “L2 Acquisition and Obligatory Head Movement: English-speaking Learners
of German and the Local Impairment Hypothesis”. Studies in Second Language Acquisition
20:311–348.
Bley-Vroman, R. 1989. “What is the Logical Problem of Foreign Language Acquisition?”. In Lin-
guistic Perspectives on Second Language Acquisition, S. Gass and J. Schachter (eds), 41–64.
Cambridge: Cambridge University Press.
Bley-Vroman, R. 1990. “The Logical Problem of Foreign Language Learning”. Linguistic Analysis
20:3–49.
Bonomi, A. 1997. “Aspect, Quantification and When-Clauses iIn Italian”. Linguistics and Phi-
losophy 20:469–514.
Bruhn De Garavito, J. 2003. “The (Dis)Association between Morphology and Syntax: The Case
of L2 Spanish”. In Linguistic Theory and Language Development in Hispanic Languages, S.
Montrul and F. Ordóñez (eds), 398–417. Somerville, MA: Cascadilla Press.
Camps, J. 2000. “preterit and imperfect in Spanish: The Early Stages of Development”. In Applied
Linguistics at the End of The Millennium, R. Leow and C. Sanz (eds), 1–19. Sommerville,
MA: Cascadilla Press.
Camps, J. 2005. “The Emergence of the imperfect in Spanish as a Foreign Language: The Asso-
ciation between imperfective Morphology and State Verbs”. International Review of Applied
Linguistics in Language Teaching 43:163–192.
Chierchia, G. 1995. “Individual-Predicates as Inherent Generics”. In The Generic Book, G. Carl-
son and F. Pelletier (eds), 176–224. Chicago: University of Chicago Press.
Chomsky, N. 1981. Lectures on Government and Binding. Dordrecht: Foris.
Chomsky, N. 1995. The Minimalist Program. Cambridge, MA: MIT Press.
Chomsky, N. 2000. “Minimalist Inquiries: The Framework”. In Step by Step: Essays on Minimalist
Syntax in Honor of Howard Lasnik, R. Martin, D. Michaels and J. Uriagereka (eds), 89–156.
Cambridge, MA: MIT Press.
104 Jason Rothman

Clahsen, H., and Hong, U. 1995. “Agreement and Null-subjects in German L2 Development:
New Evidence from Reaction Time Experiments”. Second Language Research 11:57–87.
Coppieters, R. 1987. “Competence Differences between Native and Near-Native Speakers”. Lan-
guage 63:544–573.
De Miguel, E. 1992. El aspecto de la Sintaxis del español: perfectividad e impersonalidad. Madrid:
Ediciones de la Universidad Autónoma de Madrid.
Depraetere, I. 1995. “On the Necessity of Distinguishing between (Un)Boundedness and (A)
Telicity”. Linguistics and Philosophy 18:1–19.
Franceschina, F. 2001. “Morphological or Syntactic Deficits in Near-Native Speakers? An As-
sessment of some Current Proposals”. Second Language Research 17:213–247.
Giorgi, A. and Pianesi, F. 1997. Tense and Aspect: From Semantics to Morphosyntax. Oxford:
Oxford University Press.
Goad, H. and White, L. 2006. “Ultimate Attainment in Interlanguage Grammars: A Prosodic
Approach”. Second Language Research 22:243–268.
Goodin-Mayeda, C.E. and Rothman, J. 2007. “The Acquisition of Aspect in L2 Portuguese and
Spanish: Exploring Native/Non-native Performance Differences”. In Romance Languages
and Linguistic Theory, Baauw, S. Dirjkoningen, F. and Pinto, M. (eds), 131–148. Amsterdam
and Philadelphia: John Benjamins.
Guasti, M.T. 2002. Language Acquisition: The Growth of Grammar. Cambridge, MA: MIT Press.
Hasbún, L. 1995. “The Role of Lexical Aspect in the Acquisition of the Tense/Aspect System in
L2 Spanish”. Unpublished Ph.D. Dissertation, Indiana University.
Hawkins, R. and Chan, C. 1997. “The Partial Accessibility of Universal Grammar in Second Lan-
guage Acquisition: The Failed Functional Features Hypothesis”. Second Language Research
13:187–226.
Hawkins, R. and Liszka, S. 2003. “Locating the Source of Defective Past Tense Marking
in Advanced L2 English Speakers”. In The Lexicon-Syntax Interface in Second
Language Acquisition, R. Van Hout, A. Hulk, F. Kuiken, and R. Towell (eds), 21-
44. Amsterdam and Philadelphia: John Benjamins.
Kempchinsky, P. and Slabakova, R. (Eds.) 2005. Aspectual Inquiries. Dordrecht: Kluwer. Lafford,
B. 1996. “The Development of Tense/Aspect Relations in Spanish Narratives: Evidence to
Test Competing Theories”. Paper presented at the Second Language Research Forum, Tuc-
son, AZ.
Lardiere, D. 1998a. “Case and Tense in the “Fossilized” Steady-State”. Second Language Research
14:1–26.
Lardiere, D. 1998b. “Dissociating Syntax from Morphology in a Divergent L2 End-state Gram-
mar”. Second Language Research 14:359–375.
Lardiere, D. 2000. Mapping Features to Forms in Second Language Acquisition. In Second Lan-
guage Acquisition and Linguistic Theory, J. Archibald (ed), 102–129. Malden, MA: Black-
well.
Lardiere, D. 2005. “On Morphological Competence”. In Proceedings of the 7th Generative Ap-
proaches to Second Language Acquisition Conference, L. Dekydtspotter, R.A. Sprouse and A.
Liljestrand (eds), 178–192. Somerville, MA: Cascadilla.
Lardiere, D. 2006. Ultimate Attainment in Second Language Acquisition: A Case Study. Mahwah,
NJ: Lawrence Erlbaum Associates.
Lenci, A. and Bertinetto, P. M. 2000. “Aspect, Adverbs and Events: Habituality and Perfectivity”.
In Speaking of Events, J. Higginbotham, F. Pianesi and A. C. Varzi (eds), 245–289. Oxford:
Oxford University Press.
 Aspect selection and the competing systems hypothesis 105

Liskin-Gasparro, J. 2000. “The Acquisition of Temporality in Spanish Oral Narratives: Exploring

Learners’ Perceptions”. Hispania 83:830–844.
Menéndez-Benito, P. 2002. “Aspect and Adverbial Quantification in Spanish”. In Proceedings of
the 32nd North Eastern Linguistics Society, M. Hirotani (ed), 365–382. University Of Mas-
sachusetts, Amherst: GLSA.
Montrul, S. and Slabakova, R. 2003. “Competence Similarities between Natives and Near-native
Speakers: An Investigation of the preterit/imperfect Contrast in Spanish”. Studies in Second
Language Acquisition 25:351–398.
Prévost, P. and White, L. 1999. “Finiteness and Variability in SLA: More Evidence for Missing
Surface Inflection”. Proceedings of the Annual Boston University Conference on Language
Development 23:575–586.
Prévost, P. and White, L. 2000. “Missing Surface Inflection or Impairment in Second Language
Acquisition?” Second Language Research 16:103–133.
Ramsay, V. 1990. “Developmental Stages in the Acquisition of the Perfective and the imperfec-
tive Aspects by Classroom L2 Learners of Spanish”. Unpublished Ph.D. Dissertation, Uni-
versity Of Oregon.
Rondal, J.A. 1993. “Down’s Syndrome”. In Language Development in Exceptional Circumstances,
D. Bishop and K. Mogford (eds), 165–177. Hove: Lawrence Erlbaum.
Rothman, J. and Iverson, M. 2007. “Beyond Morphological Use: What Semantic Knowledge
Tells us about Aspect in L2 Portuguese”. Paper presented at the XXXVII Linguistic Sympo-
sium on Romance Languages, University Of Pittsburgh.
Salaberry, R. 1999. “The Development of Past Tense Verbal Morphology in Classroom L2 Span-
ish”. Applied Linguistics 20:151–178.
Salaberry, R. 2000. Spanish Past Tense Aspect: L2 Development in a Tutored Setting. Amsterdam
and Philadelphia: John Benjamins.
Salaberry, R. 2002. “Tense and Aspect in the Selection of Spanish Past Tense Verbal Morphol-
ogy”. In The L2 Acquisition of Tense-Aspect Morphology, R. Salaberry and Y. Shirai (eds),
397–416. Amsterdam and Philadelphia: John Benjamins.
Schmitt, C. 1996. “Aspect and the Syntax of Noun Phrases”. Unpublished Ph.D. Dissertation.
University Of Maryland.
Schwartz, B. 1993. “On Explicit and Negative Data Effecting and Affecting Competence and
‘Linguistic Behavior’”. Studies in Second Language Acquisition 15:147–163.
Schwartz, B. 2003. “Child L2 Acquisition: Paving the Way”. In BUCLD 27 Proceedings, B. Beach-
ley, A. Brown, and F. Conlin (eds), 25–50. Somerville: MA: Cascadilla Press.
Schwartz, B. and Gubala-Ryzak, M. 1992. “Learnablity and Grammar Reorganization in L2 Ac-
quisition: Against Negative Evidence Causing the Unlearning of Verb Movement”. Second
Language Research 8:1–38.
Schwartz, B. and Sprouse, R. 1996. “L2 Cognitive States and the Full Transfer/Full Access Mod-
el”. Second Language Research 12:40–72.
Slabakova, R. 2001. Telicity in Second Language. Amsterdam and Philadelphia: John Benja-
mins.
Slabakova, R. 2006. “Is there a Critical Period for Semantics?”. Second Language Research
22:302–338
Slabakova, R. and Montrul, S. 2003. “Genericity and Aspect in L2 Acquisition. Language Acqui-
sition 11:165–196.
Smith, C. 1991. The Parameter of Aspect. Dordrecht: Kluwer.
106 Jason Rothman

Sorace, A. 2000. “Syntactic Optionality in Non-native Grammars.” Second Language Research

16:93–102.
Sorace, A. 2003. “Near-nativeness”. In The Handbook of Second Language Acquisition,
C. Doughty and M. Long (eds), 130–153. Oxford: Blackwell Publishers.
Sorace, A. 2004. “Native Language Attrition and Developmental Instability at the Syntax-Dis-
course Interface: Data, Interpretations and Methods”. Bilingualism: Language and Cognition
7:143–145.
Sorace, A. 2005. “Syntactic Optionality at Interfaces”. In Syntax and Variation: Reconciling the
Biological and the Social, L. Cornips and K. Corrigan (eds), 46–111. Amsterdam and Phila-
delphia: John Benjamins
Van Der Lely, H. and Wexler, K. 1998. “Introduction to the Special Issue on Specific Language
Impairment in Children”. Language Acquisition 7:83–85.
White, L. 1989. Universal Grammar and Second Language Acquisition. Amsterdam and Phila-
delphia: John Benjamins.
White, L. 2003. Second Language Acquisition and Universal Grammar. New York: Cambridge
University Press.

Author’s address
Jason Rothman
The University of Iowa
Department of Spanish and Portuguese
111 Phillips Hall
Iowa City, IA 52242
USA
jlrothma@[Link]