See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/316094089

Universal Grammar and Second Language Acquisition

Chapter · November 2012

DOI: 10.1002/9781405198431.wbeal1235

CITATIONS READS

5 9,278

1 author:

Teresa Parodi
University of Cambridge
27 PUBLICATIONS 437 CITATIONS

SEE PROFILE

All content following this page was uploaded by Teresa Parodi on 04 June 2020.

The user has requested enhancement of the downloaded file.

 Encyclopedia of Applied Linguistics 2013

Universal Grammar and Second

Language Acquisition
TERESA PARODI

This entry will introduce universal grammar (UG) with respect to second language (L2)
acquisition. It will first deal with the notion of UG and how it works in the adult language
and in the acquisition of the first language (L1) before turning to the acquisition of an L2.

General Points

The UG framework takes a cognitive approach to language. A grammar is understood

as the knowledge we have of our own language (“of” the language, not “about” it; for
example, speakers of a language are able to produce a negated utterance, even if they
cannot explain the linguistic process leading to it). This knowledge allows us to produce
and to understand utterances in this language. It also allows us to identify sounds, words,
sentences, and combinations of them that are part of this language or are not.
Native speakers know not only what expressions mean, but also whether the form of a
given expression is or is not consistent with the grammar of their language. And it is here
that UG plays a role. UG is assumed to be part of the innate language faculty. The role of
UG is to determine in advance what is or is not a possible grammar. UG provides an
inventory of possible grammatical categories (e.g., nouns, verbs, plural, or tense as both
morphological and semantic categories) as well as an inventory of possible grammatical
operations; that is, it determines the way in which grammars can function. Some aspects
of language hold across languages (so-called principles of UG); others are language specific.
Language variation is not random: Languages do not form questions following linear
ordering rules of the type “the question word should be the third word in the interrogative
clause,” which would yield did Taroo what eat? as well as did your what friend eat? Languages
vary along certain lines, known as parameters of language variation. For example, all
languages have question words, but their distribution varies: They can appear in the position
where non-question words appear in declarative clauses, as in the Japanese sentences of
(1), or in a position at the edge of the clause, as in the English sentence (2).

(1) a. Taroo-wa keiki-o tabemasita

Taroo-top cake-acc ate
b. Taroo-wa nani-o tabemasita-ka
Taroo-top what-acc ate?

(2) What did Taroo eat?

UG in L1 Acquisition

Before getting into L2 acquisition it is helpful to have a brief look at how UG works in L1
acquisition, focusing in particular on the initial state, the developmental process, and the

The Encyclopedia of Applied Linguistics, Edited by Carol A. Chapelle.

© 2013 Blackwell Publishing Ltd. Published 2013 by Blackwell Publishing Ltd.
DOI: 10.1002/9781405198431.wbeal1235
 Encyclopedia of Applied Linguistics 2013
2 universal grammar and second language acquisition

end state. The last is clear: The outcome of L1 acquisition is inevitably the adult version
of that particular language. How do children get there? The obvious answer is that children
learn their L1 through exposure to linguistic experience: Children need to be exposed to
their L1 and receive sufficient input. However, the linguistic data children are exposed to
are both under- and over-informative. They are under-informative in that not all possible
words, forms, and expressions will be represented in what children hear. Absence from
the input could mean either that that expression has just not occurred by chance or that
the expression is impossible in that particular language: Children are able to identify those
expressions that belong to the latter set, and to establish some kinds of grammatical
knowledge, in spite of the fact that the input is under-informative. An example is the
identification of possible antecedents for anaphors like himself, as in example (3). In this
case there are two participants, Tom and Max; himself could in principle refer to either of
them, but in English it can only apply to Max.

(3) Tom said that Max hit himself.

On the other hand the input may be potentially misleading about the target language,
as it may contain false starts, repetitions, speech errors, and so forth (over-informative
input). UG mitigates the effects of the under- and over-informative input (poverty of the
stimulus) by offering a framework of categories and operations that constrain the search
or hypothesis space within which grammatical representations are formulated and, thus,
guiding both interpretation and production. The initial state in L1 acquisition is, then, UG
paired with exposure to linguistic experience. UG guides development, so that the child’s
grammar is at all stages a possible human grammar, as opposed to a wild collection of
words without any particular organization or organized according to individual rules,
which would make the child unintelligible for everyone else. UG guidance accounts for
the fast rate of acquisition and for the similarity of developmental stages across individuals
learning the same language and across different languages.

UG in L2 Acquisition

If we now turn to L2 acquisition, the central question is whether in this case the acquisition
process is guided by UG as the acquisition of the L1 is. What would be the evidence? We
can compare with the acquisition of the L1 and its initial, intermediate and final stages.
Given the end point in the acquisition of the L1, the most straightforward evidence
would be if learners would achieve in the L2 a level of proficiency equivalent to that of a
native speaker of that language. Arguably, most speakers of an L2 do not become native-
like. This in itself, however, does not exclude the possibility that UG is constraining the
process: Recall that children acquiring their L1 go through stages that are different from
the adult grammar but, as said above, each of them represents a natural grammar. The
same could apply to developmental stages of L2 grammars. We will explore this in what
follows.
With respect to the initial stage, recall that for the L1 the assumption is that the initial
stage consists of UG plus exposure to data of the relevant language. Learners of an L2 by
definition already have an L1, which represents UG as it is instantiated in that particular
language. But L2 learners will need to go beyond that instantiation toward the L2 target:
Transfer from the L1 alone will not do. They need to acquire phenomena which are not
part of the L1. And the question which arises here is: If the L1 is constrained by UG, will
these constraints also hold in the L2?
 Encyclopedia of Applied Linguistics 2013
universal grammar and second language acquisition 3

Does UG Constrain the Development in the L2?

This question was at the center of the debate about UG access in the 1980s (White, 1985;
Clahsen & Muysken, 1986, 1989; Bley-Vroman, Felix, & Ioup, 1988). Specifically the question
was whether it is possible to construct a grammar beyond that of the L1. As such, it is
directly related to that of parameter (re)setting: Is it possible to modify the parametric
specification of one’s own language? Is the outcome a natural grammar or a “wild” one?
L2 learners, particularly at early developmental stages, show effects of transfer from the
L1. For example, speakers of Chinese, a language without articles, leave out articles in
English, a language in which they are obligatory. Initially transfer effects were taken as
evidence against the involvement of UG in L2 acquisition (no UG access; Clahsen &
Muysken, 1986). Later it was acknowledged that, as the L1 is UG constrained, whatever
is transferred from it must by definition also be UG constrained (indirect access to UG;
Clahsen & Muysken, 1989).
But is UG actively constraining the L2 beyond what is instantiated in the L1? Are the
production and comprehension of structures and phenomena not attested in the L1 also
constrained by UG?

Evidence of UG Constraints in the L2

Evidence of UG constraints in the L2 are in the first place the systematicity of the interlanguage
and, furthermore, a systematicity which represents a possible human grammar.
Learners make “errors”: they produce or understand language in ways which diverge
from that of native speakers. But, as first claimed by Selinker (1972), learners’ errors are
systematic, which points to a rule-based system. This is Selinker’s idea of an interlanguage,
a grammatical system in its own right, as opposed to “habits” taken over from the L1
(contrastive analysis hypothesis; Lado, 1957) or a random collection of “errors.” The study
of L2 in the UG framework adopts the basic idea of interlanguage (Selinker, 1972): Errors
are systematic, which indicates that the system is rule based.
Assuming, then, that interlanguage is rule based, the question arises whether inter-
language representations are also possible human grammars, an instance of a possible natural
language.

Are L2 Grammars Possible Natural Language Grammars?

Clustering of Morphological and Syntactic Properties

One way of answering this question is by observing specific phenomena in an L2 and
testing whether they are compatible with a UG analysis. There are numerous examples of
studies with this focus. A straightforward case is the clustering of morphological and
syntactic indications of finiteness on the verb. Finite verbs are marked for agreement, tense,
or both, depending on language-specific properties. Finite and nonfinite verbs have dif-
ferent distributions in the sentence. This is clearly visible in English (or German or French)
in that finite verbs, marked for person agreement and tense, precede negation, while
nonfinite ones follow negation. See (4a–c).

(4) a. Kim has not seen Robin English

finite nonfinite
b. Kim hat Robin nicht gesehen German
finite nonfinite
c. Kim a pas vu Robin Colloquial French
finite nonfinite
 Encyclopedia of Applied Linguistics 2013
4 universal grammar and second language acquisition

Children learning their L1 have been observed to be very accurate very early with the
clustering of morphological and syntactic markings: Verbs marked for tense and agreement
do not occur in the slot for nonfinite verbs (Clahsen, 1988; Pierce, 1989; Meisel, 1994; among
many others). Negation consistently follows finite verbs and precedes nonfinite ones, as
in the French examples (5a–b):

(5) a. marche pas b. pas casser

walk.3sg not not break.infinitive
‘s/he doesn’t walk’ ‘don’t break (it)’

It is less clear how this clustering works in L2 acquisition. One of the reasons for this
is the fact that L2 learners, particularly adult learners, have lasting problems with bound
morphology (e.g., realization of affixes), which, in turn, makes it difficult to distinguish
finite and nonfinite verbs. Some authors (Clahsen & Muysken, 1986, 1989; Meisel, 1991,
1997) claim that the clustering does not exist in L2 acquisition, a fundamental difference
with L1 acquisition. According to these authors L2 learners fail to mark agreement and
tense consistently and the distribution of negation follows semantic, rather than syntactic,
criteria. Their observations are based on the fact that in L2, German verbs which according
to the context should be finite lack the morphological marking and, in some cases, occur
with the wrong distribution. An example would be (6), produced by a speaker of Italian
(Clahsen, Meisel, & Pienemann, 1983), the German equivalent of I no buy, where the negation
immediately precedes the verb to which it applies; this distribution is, however, incorrect
in a finite context. According to the authors mentioned, the learners follow a strategy based
on linear order and semantics, rather than structural criteria. This, in turn, means that any
grammatical knowledge in the L2 is drawn from the L1; given that the L1 is assumed to be
constrained by UG, this constraint applies also to the L2 in cases of transfer from the L1.

(6) ich nein kauf

I no buy.root
Target: ‘ich kaufe nicht, ich habe nicht gekauft’

A different interpretation of these data is offered by Parodi (2000), who observes that
the omission of morphological markings affects a subset of verbs, namely, full lexical verbs
(or “thematic” verbs). Modals, auxiliaries, copula be, and possessive have (“non-thematic”
verbs), on the other hand, display correct subject–verb agreement in over 90% of the
contexts over a two-year period. Crucially, these verbs also occur with the expected dis-
tribution for finite verbs, that is, to the left of negation as in example (7). This is in striking
contrast to thematic verbs, with a rate of 59% of correct agreement and a fluctuating dis-
tribution. In other words, the clustering of morphological and syntactic properties does
hold, but is initially limited to the subset of non-thematic verbs, and extends to thematic
verbs in the course of development.

(7) aber er hat nich die papier

but he have.3sg not the papers
‘but he has no documents’

Other authors (Lardiere, 2000; Prévost & White, 2000) acknowledge the omission of tense
and agreement markings, but claim that the distribution of verb forms is still not random.
Their claim is that verbs which show the morphological marking never occur with the
wrong distribution. According to these authors this indicates that syntactic constraints are
operating in the interlanguage.
 Encyclopedia of Applied Linguistics 2013
universal grammar and second language acquisition 5

These examples illustrate how researchers compare L1 and L2 acquisition asking to what
extent UG constrains the latter; that is, whether the interlanguage is a possible human
grammar.

Poverty of the Stimulus Effects

Schwartz and Sprouse (2000), however, observe that UG-compatible analyses of inter-
language data may become obsolete if the theory they are formulated in evolves. These
authors point out that for this reason, possibly the best way of clarifying the role of UG
in L2 is by focusing on effects of poverty of the stimulus, that is, properties which are not
part of the L1 and which are underdetermined in the L2 input. Poverty of the stimulus
effects can be evident in the speakers’ knowledge of grammaticality and ungrammaticality:
Native speakers know which expressions (sounds, words, sentences) are part of their gram-
mar or not, that is, which expressions are “grammatical” in their system. For example, a
speaker of English knows that while (8a) is possible in certain registers, (8b) is not. In
technical terms we would say that, although subjects are obligatory in English, they can
be left out in a colloquial context (8a). Omitting the subject, however, is impossible in an
embedded clause (8b).

(8) a. _ seems that it’s raining

b. it seems that _ is raining

Poverty of the Stimulus in the Syntax

In the L2 acquisition literature there are many studies which address the question of
grammaticality in syntactic distribution. The domain of anaphors can offer an example,
specifically about the binding of anaphors with their antecedents. There is crosslinguistic
variation in the way anaphors bind to local or nonlocal antecedents. Recall example (3),
here repeated as (9). In English himself can only refer to Max (local antecedent). In a com-
parable Japanese example like (10) (Thomas, 1995, p. 222), on the other hand, zibun can refer
to either Tom or Max. Notice, crucially, that both Tom and Max are subjects, respectively
of the matrix and the subordinate clause.

(9) Tom said that Max hit himself.

(10) Tom-wa Max-ga zibun-no butta to iimasita.

Tom-topic Max-nom self-gen hit comp said

English allows only for local binding of anaphors, so the same holds when the antecedent
is a nonsubject: Long-distance binding is impossible. In example (11) himself can only mean
Max. Interestingly, Japanese does not allow for long-distance binding either if the antecedent
is a nonsubject. Consider example (12) (Thomas, 1995, p. 221): In this case the long-distance
binding of zibun can refer to Tom’s (i.e., the subject’s) book, but not to Kim’s.

(11) Tom said that Max showed a picture of himself.

(12) Tom-wa Kim-ni Max-ga zibun-no hon-ga suki-da to iimasita.

Tom-top Kim-dat Max-nom self-gen book-acc like-cop comp said
‘Tom told Kim that Max likes self’s book.’

What examples (10) and (12) show is that anaphors can bind to a long-distance antecedent
provided this is a subject. Long-distance binding of a nonsubject is not possible. The reason
 Encyclopedia of Applied Linguistics 2013
6 universal grammar and second language acquisition

Table 1 Response “true” in percentages

Groups Long-distance subject *Long-distance nonsubject

L2 low proficiency (n = 34) 55 54

L2 advanced (n = 24) 57 14
Native speakers (n = 34) 89 18

is that antecedent and anaphor must be in the structural configuration known as c-command.
This configuration holds for a long-distance reflexive like Japanese zibun, which can be
c-commanded by a noun phrase (NP) subject of a matrix clause. Locally binding anaphors
can have subject or nonsubject antecedents. But the c-command configuration does not
hold in a bi-clausal sentence like (12), if the anaphor zibun is interpreted as having a non-
subject antecedent outside its own clause (Thomas, 1995).
Given the facts just described, the Japanese input is under-informative for a learner.
While it is clear that zibun can bind a long-distance antecedent, the input does not provide
clear evidence about the constraints on interpretation, namely, that zibun can only refer to
a long-distance subject. The question arises whether learners of Japanese whose native
language does not allow long-distance binding can learn the properties of zibun and disallow
binding with a nonsubject in a matrix clause.
Thomas (1995) tested this question with English learners of Japanese at two proficiency
levels, low and advanced. The test methodology was a truth-value judgment based on the
context provided by a sequence of pictures accompanied by written utterances. Each scen-
ario was followed by a statement which the participants had to evaluate as true or false.
There were four items per condition. The aim was to find out whether learners of Japanese
are aware of the possibility of long-distance binding of an anaphor to a long-distance
subject as opposed to a nonsubject. Table 1 (from White, 2003, p. 49) illustrates the results.
The figures represent percentages of “true” responses.
The table shows that low-proficiency learners accept subject and nonsubject antecedents
at more or less the same rate. High-proficiency learners, however, appear to be making a
clear difference between subject and nonsubject antecedents. Thomas’s (1995) hypothesis
concerns a specific subset of learners, namely, those who admit long-distance binding of
zibun. The question is whether the grammar of individual learners shows the requirement
that the long-distance antecedent should be a subject. The criterion used to determine the
individual consistency in a particular feature of zibun is that of three or four out of four
responses. Thomas (1995) identifies 23 out of 54 participants who consistently accept long-
distance binding. Out of these 70% reject long-distance nonsubject antecedents of zibun.
The study illustrates a case in which learners achieve knowledge of properties of an
L2 grammar which are not explicit in the input, that is, a case of poverty of the stimulus
in the domain of syntax. UG is assumed to make up for the under-informativeness of the
input by providing a framework of categories and operations that limit the search space
and, thus, to guide both production and interpretation,

Poverty of the Stimulus in the Syntax–Semantics Interface

The syntax–semantics interface is another domain in which native speakers have knowledge
of (un)grammaticality. We find an example in the domain of argument structure, specifi-
cally with respect to intransitive verbs (13, 14). These verbs have only one argument, which
is realized as a subject. This only argument is in some cases an agent, as in laugh (13), and
in others a theme, as in open (14). While agents make prototypical subjects, themes are
 Encyclopedia of Applied Linguistics 2013
universal grammar and second language acquisition 7

more closely associated with objects (X opened the door). There is a parallel between the
theme argument as a subject of an intransitive verb (14) and passive subjects (15). (In [15]
the door, initially the object of the verb open, is realized as subject.) Intransitive verbs with
a theme subject have been called “unaccusatives” (Burzio, 1986).

(13) Kim laughed. (intransitive, subject: agent)

(14) the door opened. (intransitive, subject: theme, unaccusative)

(15) the door (theme) was opened. (passive sentence)

There is a range of syntactic and morphological effects of unaccusativity in different lan-

guages, which makes it a potentially revealing field in which to test the role of UG in L2
acquisition. It is interesting to see whether L2 learners are sensitive to unaccusativity effects.
An indication that this is indeed the case comes from errors of passivization, as observed
by Zobl (1989), Oshita (2000, 2001), and Balcom (1997), among others. In examples (16) and
(17), from learners’ production, the intransitive verbs are used with passive morphology
(Oshita, 2001). Further evidence comes from cases like (18) (Zobl, 1989), in which the subject
of the unaccusative verb appears postverbally, a type of inversion only attested with unac-
cusative verbs. These results are confirmed by experimental studies, such as Hirakawa’s (2000)
on the acquisition of L2 Japanese, which shows that English and Chinese learners make a
correct distinction between unaccusative and other intransitive verbs.

(16) my mother was died when I was just a baby.

(17) an accident was happened

(18) I was patient until dried my clothes

The studies mentioned show that L2 learners are sensitive to thematic roles and what
their default mapping to the syntactic realization is: In cases in which the mapping of
thematic roles to syntax is not the default option (i.e., the subject is an agent), the learners
resort to passivization or a linear order which reflects the default association of the theme
role with the object function. In other words, the learners are making use of linguistic
knowledge of aspects not explicitly present in the input.

Poverty of the Stimulus and Properties Not Present in Either L1 or L2

A further type of poverty of the stimulus evidence comes from a learner grammar which
could display properties not present in the L1 or the L2, but reflecting general linguistic
properties present in other languages. Clearly, in this case the learners must have abstracted
those linguistic properties with the help of UG.
An example is found in the already-mentioned study by Parodi (2000) on the acquisition
of German by speakers of Romance languages. Parodi finds a consistent difference between
modals and auxiliaries (“non-thematic” verbs) on the one hand, and full lexical or “thematic”
verbs on the other. These verb classes differ with respect to both morphological properties
and syntactic distribution in the learners’ system, although neither the Romance L1s nor
the L2 German carry out such a distinction. Interestingly, this distinction between thematic
and non-thematic verbs is present in other natural languages, and the most immediate
example is English. In this language, non-thematic verbs (modals and auxiliaries) inflect
differently from thematic verbs: They have no -s in the third person in the present tense,
 Encyclopedia of Applied Linguistics 2013
8 universal grammar and second language acquisition

for example. In addition, this has a syntactic correspondence: Non-thematic verbs appear
to the left of negation, a position blocked for thematic verbs. (See [19].)

(19) Tim will not arrive in time.

*Tim arrives not in time

These results are taken as an indication that learners are sensitive to the properties of finite
and nonfinite verbs and their distribution, and establish a relation between morphology and
syntax, although not in the way their L1 or the L2 does it. This is, again, a case in which
the linguistic knowledge available to L2 learners goes beyond the evidence in the input.

Summary

The UG approach views language as a mental system and offers a principled account of
language learning; in this particular case, of the acquisition of the L2. The account is based
on linguistic analyses which make specific predictions about the acquisition process, which
can be tested and proved true or false. It looks at how words and constituents are combined
to create novel utterances and offers explanations on the properties of the interlanguage:
how the learners arrive at the target-like forms in production or the target-like interpretation,
as well as why the learners’ production and interpretation deviate from those of native
speakers of a certain language.
The approach also offers ways of accounting for crosslinguistic variation and how this
will work in specific combinations of L1 and L2; that is, it makes specific (testable) claims
about the role of the L1 and that of the input (Braidi, 1999, p. 74). Exposure to primary
linguistic data is imperative and the type of data (naturalistic immersion, classroom
material) will undoubtedly shape the learners’ output alongside the L1.
The fact that this approach takes a cognitive perspective does not mean denying the
existence of other components in language and its acquisition, such as communicative
(types of exchange: conversation, a formal written text, etc.) and sociolinguistic aspects
(gender, socioeconomic level of the speakers, what the language is used for—e.g., business,
education, family language). There are, however, some aspects which are independent of
communicative or sociolinguistic aspects. For example, the way a language marks tense
(e.g., by suffixes on the verb or by adverbials) and the way an embedded clause is struc-
tured are not likely to be affected by the socioeconomic level of the speakers. A change of
register is likely to lead to a different lexical choice, but not to a different grammatical
choice, such as to mark tense on a noun instead of on a verb. In this sense the approach
is complementary to a sociolinguistic approach: They focus on different linguistic aspects.
The UG perspective on L2 acquisition offers linguistically informed explanations of
why learners behave as they do, beyond the direct comparison of language strings. These
explanations can form the background for psycholinguistic or neurolinguistic testing. They
also provide valuable insights which can be incorporated into the design of classroom and
test materials.

SEE ALSO: Critical Period; Crosslinguistic Influence in Second Language Acquisition;

Generative Grammar; Input Processing in Second Language Acquisition

References

Balcom, P. (1997). Why is this happened? Passive morphology and unaccusativity. Second Language
Research, 13, 1–9.
 Encyclopedia of Applied Linguistics 2013
universal grammar and second language acquisition 9

Bley-Vroman, R., Felix, S., & Ioup, G. (1988). The accessibility of universal grammar in adult
language learning. Second Language Research, 4, 1–32.
Braidi, S. (1999). The acquisition of second language syntax. London, England: Edward Arnold.
Burzio, L. (1986). Italian syntax. Dordrecht, Netherlands: Reidel.
Clahsen, H. (1988). Parameterized grammatical theory and language acquisition: A study of
the acquisition of verb placement and inflection by children and adults. In S. Flynn &
W. O’Neil (Eds.), Linguistic theory in second language acquisition (pp. 47–75). Dordrecht,
Netherlands: Foris.
Clahsen, H., Meisel, J. M., & Pienemann, M. (1983). Deutsch als Zweitsprache: Der Spracherwerb
ausländischer Arbeiter. Tübingen, Germany: Narr.
Clahsen, H., & Muysken, P. (1986). The availability of universal grammar to adult and child
learners: A study of the acquisition of German word order. Second Language Research, 2,
93–119.
Clahsen, H., & Muysken, P. (1989). The UG paradox in L2 acquisition. Second Language Research,
5, 1–29.
Hirakawa, M. (2000). Unaccusativity in second language Japanese and English (Unpublished doctoral
dissertation). McGill University, Canada.
Lado, R. (1957). Linguistics across cultures. Ann Arbor: University of Michigan Press.
Lardiere, D. (2000). Mapping features to forms in second language acquisition. In J. Archibald
(Ed.), Second language acquisition and linguistic theory (pp. 102–29). Oxford, England: Blackwell.
Meisel, J. M. (1991). Principles of universal grammar and strategies of language learning: Some
similarities and differences between first and second language acquisition. In L. Eubank
(Ed.), Point counterpoint: Universal grammar in the second language (pp. 231–76). Amsterdam,
Netherlands: John Benjamins.
Meisel, J. M. (1994). Getting FAT: Finiteness, agreement and tense in early grammars. In
J. M. Meisel (Ed.), Bilingual first language acquisition: French and German grammatical develop-
ment (pp. 89–129). Amsterdam, Netherlands: John Benjamins.
Meisel, J. M. (1997). The acquisition of the syntax of negation in French and German: contrasting
first and second language development. Second Language Research, 13, 227–63.
Oshita, H. (2000). What is happened may not be what appears to be happening: A corpus study
of “passive” unaccusatives in L2 English. Second Language Research, 16, 293–324.
Oshita, H. (2001). The unaccusative trap in second language acquisition. Studies in Second Language
Acquisition, 23, 279–304.
Parodi, T. (2000). Finiteness and verb placement in second language acquisition. Second Language
Research, 16, 355–81.
Pierce, A. (1989). On the emergence of syntax: A crosslinguistic study. Cambridge, MA: MIT Press.
Prévost, P., & White, L. (2000). Missing surface inflection or impairment in second language
acquisition? Evidence from tense and agreement. Second Language Research, 16, 103–33.
Schwartz, B. D., & Sprouse, R. (2000). When syntactic theories evolve: Consequences for L2
acquisition research. In J. Archibald (Ed.), Second language acquisition and linguistic theory
(pp. 156–86). Oxford, England: Blackwell.
Selinker, L. (1972). Interlanguage. International Review of Applied Linguistics, 10, 209–31.
Thomas, M. (1995). Acquisition of the Japanese reflexive zibun and movement of anaphors in
logical form. Second Language Research, 11, 206–34.
White, L. (1985). The pro-drop parameter in adult second language acquisition. Language Learning,
35, 47–62.
White, L. (2003). Second language acquisition and universal grammar. Cambridge, England: Cambridge
University Press.
Zobl, H. (1989). Canonical typological structures and ergativity in English L2 acquisition. In
S. Gass & J. Schachter (Eds.), Linguistic perspectives on second language acquisition (pp. 203–21).
Cambridge, England: Cambridge University Press.

View publication stats