Review of The Lexicon–Syntax Interface in Second Language Acquisition
|
|
|
|
|
Review:
|
Date: Wed, 18 Feb 2004 10:45:25 -0800 (PST)
From: Ahmad R. Lotfi <arlotfi@yahoo.com>
Subject: The Lexicon-Syntax Interface in Second Language Acquisition
van Hout, Roeland, Aafke Hulk, Folkert Kuiken, and Richard Towell,
eds. (2003) The Lexicon-Syntax Interface in Second Language
Acquisition, John Benjamins Publishing Company, Language
Acquisition & Language Disorders 30.
Ahmad Reza Lotfi, Azad University at Khorasgan
SYNOPSIS
"The Lexicon-Syntax Interface in Second Language Acquisition" is a
selection of papers presented atthe NWCL/LOT Expert Seminar held
in Amsterdam on March 30-31, 2001. In addition to the introductory
paper by Richard Towell, and the concluding remarks by Roeland van
Hout, Aafke Hulk and Folkert Kuiken, the volume contains 8
articles on generativist and psycholinguistic studies of the
interface between the lexicon and syntax in second language
acquisition.
Richard Towell in his introductory chapter "Introduction: Second
language acquisition research in search of an interface" reviews
the relation between linguistics and psychology in the past
focusing on both linguistic (specifically generativist) and
psychological dimensions of second language acquisition, and the
extent to which such different perspectives of SLA research can be
complementary. It is within this framework that he outlines the 8
articles (chapters 2-9) in this volume.
Roger Hawkins and Sara Liszka's "Locating the source of defective
past tense marking in advanced L2 English speakers" is an analysis
of optionality in Chinese English interlanguage with regard to
marking tense in their spontaneous oral production in such
sentences as "The police CAUGHT the man and TAKE him away." They
claim that such errors are due to the difficulty Chinese speakers
have in assigning the formal feature [past] to the category
T(ense) as such a feature is not selected in their L1. The
participants in their study--Chinese (n=2), Japanese (n=5), and
German (n=5) advanced L2 speakers of English--were expected to
produce present/past tense forms for (irregular) nonce verbs like
*spling-->splung*. Also their spontaneous oral data were elicited
by means of story-retelling and personal experience recounting
tasks. While the non-native participants with different L1
backgrounds were not significantly different in their performance
on the inflection test, the Chinese informants were significantly
less accurate in inflecting thematic regular/irregular verbs in
oral tasks. This suggests that "Chinese speakers cannot establish
[+/-past] on T in English precisely because this feature is absent
in their L1" (p.40).
"Perfect projections" by Norbert Corver assumes a minimalist
interface perspective on L2 knowledge. For Corver, "L2-products of
inter-language grammars are typically 'target-imperfect' but
'interface-perfect'" (p. 48) in the sense that despite their
idiosyncracy once compared with native products, they consist of
features interpretable at interface levels. An analysis of L2-
expressions produced by Turkish L2-learners of Dutch is taken to
support this view.
Ineke van de Craats in "L1 features in the L2 output" also focuses
on the nature of L2ers' grammatical knowledge at the L2-initial
state, and the possible role L1 transfer may play in this respect.
The data in this case come from a longitudinal study of 8 adults
(followed for 2.5 years) with Turkish and Moroccan Arabic L1
backgrounds learning Dutch as a second language. The results
suggest that while L2ers are more aware of the fact that
morphological and lexical properties of words differ across
languages, they originally assume L1 formal feature constellation,
which gradually changes in favour of the L2 ones.
Nigel Duffield's "Measures of competent gradience" is concerned
with lexical and syntactic gradient effects in a revised model of
competence in which (plausibly universal) underlying competence
(UC) "is categorical, and consists of formal (phonological and
syntactic) principles autonomous from the lexicon" and (largely
language-specific) surface competence (SC) which "is intimately
determined by the interaction of contextual and specific lexical
properties with the formal principles delivered by UC ..." (p.
101). A review of empirical studies employing a variety of tasks
such as online/offline grammaticality judgements, sentence-
matching paradigm, and also a replication of the Marlsen-Wilson et
al. experiments with L2ers seems to support this dual model in
that explicit tasks tap SC, whereas implicit tasks tap UC.
Ton Dijkstra in "Lexical storage and retrieval in bilinguals"
argues that syntactic parsing is qualitatively and quantitatively
different between monolinguals and bilinguals. The participants in
the studies reported in this article are late bilinguals highly
proficient in English but speaking Dutch as their strongest native
language. The results suggest that the bilingual word
identification system is largely automatic to the effect that
intentional and attentional factors do not influence the process
of word recognition. Despite that, L2 lexicon is more slowly
activated. Also experimental and contextual factors may influence
the retreival patterns.
"Inducing abstract linguistic representations: Human and
connectionist learning of noun classes" by John N. Williams
focuses on the similarities and differences between human and
connectionist learning of word classes. Williams and Lovatt (2003)
show that an arbitrary noun class system with masculine and
feminine genders is learnable to both man and a connectionist
network (via distributional information). "The problem is,
however, that the networks only seem to account for learning
amongst those participants who already possessed knowledge of
other gender languages. Yet none of the networks contained any
prior knowledge" (p.167). This is in harmony with the observation
that gender is a persistent problem in second language
acquisition.
Laura Sabourin and Marco Haverkort in their "Neural substrates of
representation and processing of a second language" compare the
results obtained with different methods including an off-line
grammaticality judgement task, on-line EEG measurements, and the
evidence from aphasia studies. They notice that while the
differences between aphasics and unimpaired users are
*quantitative*, the difference between native speakers and L2ers
in the processing of language is qualitative. "[L]inguistic
processing (as reflected by the P600) can only occur in the L2
when the processing strategy from the L1 can be used relatively
directly in L2 processing" (p. 193).
David W. Green in "Neural basis of lexicon and grammar in L2
acquisition" focuses on the differential representation hypothesis
in contrast with the convergence hypothesis concerning the
question of whether or not L1 and L2 lexical and grammatical
knowledge are represented differently in the brain. Based on the
ERP data and haemodynamic methods, he argues that "as proficiency
in L2 increases, the networks mediating L2 converge with those
mediating language use in native speakers of that language" (p.
212).
Roeland van Hout, Aafke Hulk and Folkert Kuiken conclude the
volume with the remark that with the shift of the theory from
principles and parameters into minimalist syntax the importance of
the lexicon has been crucially increased. While the use of lexical
items must still take place within a syntactic system, the driving
force of language acquisition shifts to the lexicon.
CRITICAL EVALUATION
The volume is an insightful and meticulously selected and
organised collection of papers on the latest developments in the
generative/psycholinguistic studies of L2 grammar. Towell's
introduction to the volume goes much beyond an ordinary
introduction to papers in a collection: it beautifully
contexualises the contributions in generative and psychological
paradigms of research in general and the implications for SLA
research in particular. Van Hout, Hulk, and Kuiken's concluding
remarks also add to the value of the collection with some of their
comments to reside for a long time in the reader's memory. The
editing and proof-reading of the papers have been admirably
careful with a minimum of typographic mistakes--actually, the only
mistake I could find in the whole volume was one in Sabourin and
Haverkort's article(p. 185) where they report that there were 39
(?) participants in total, 23 native speakers of Dutch and 14
second language speakers.
Minimalist syntax, however, is still conceptually too complicated
and also somehow controversial in both theoretical and empirical
issues to be straight-forwardly applied in SLA research without
problems unanticipated. For instance, Hawkins and Liszka claim
that "there is a syntactic (i.e. semantically uninterpretable)
tense feature which, for the sake of exposition, we call [+/-
past], which is available in the universal inventory F, but which
is optional" (p. 25). Apparently, the authors have confused
interpretable/noninterpretable features and semantic/formal ones.
Corver (p. 51) hypothesises syntactic-formal features to mediate
sound and meaning at the word level. Most probably,
'morphological-formal' is a better term for that purpose. He also
considers *van* in Dutch(e.g. in "Ik niet trouwen van Yvette" p.
61) to be a *case-suffix*, which is rather surprising. Also he
claims that "[i]t is the element carrying the [+interpretable]
feature that agrees with the element carrying the [-interpretable]
feature" (p. 62) while it is usually just the opposite in the
standard literature on the issue. On page 64, we read: "numbers
above one are inherently marked for the property [+plural]." This
ignores the case of duality in such languages as Arabic. Also he
argues that "[f]rom the perspective of the target language,L2-
expressions often seem highly imperfect" (p.65).Such a use of the
term 'imperfect' is quite different from what Chomsky and other
minimalist syntacticians originally intended. Corver should use
less marked terminology like 'unnatural' or 'idiosyncratic'.
Van de Craats distinguishes "lexical knowledge as defined by UG
from language-specific knowledge of lexical items and their
lexical entries" (referred to as the vocabulary by van de Craats)
(p. 72). If we accept Chomsky's definition of the lexicon as a
list of exceptions ("whatever does not follow from general
principles. These principles fall into two categories: those of
UG, and those of a specific language." 1995:235), then UG cannot
define the lexicon anymore.
Sabourin and Haverkort suggest that "[i]t is possible that
successful second language learners have native-like *knowledge*,
just like the aphasics (suggesting that access to Universal
Grammar (UG) for a second language is possible. However, they may
actually *process* this knowledge in a non-native-like manner"(p.
179). However, Native-like knowledge and UG are not the same at
all. Even some intelligent creatures from the other side of the
galaxy may happen to pickup some native-like knowledge of English
(if this knowledge is essentially learnable even without a UG,
which seems to be the case although it seems to play a
facilitating role for human infants--one without which L1
acquisition would be impossible) with no access to any UG. Then
even if empirically speaking UG and native-like knowledge happen
to be the same, they are conceptually quite distinct.
Finally, Green advocates a convergence hypothesis according to
which, "as proficiency in L2 increases, the networks mediating L2
converge with those mediating language use in native speakers of
that language" (p. 212). Even if higher levels of L2 proficiency
are associated with more native-like neurological representations
of the language, one still cannot claim that L2 proficiency is the
CAUSE(and not an EFFECT) of such a process of neurological
levelling.
REFERENCE
Chomsky, N. 1995. The Minimalist Program. The MIT Press.
|
| |
ABOUT THE REVIEWER:
ABOUT THE REVIEWER
Dr. Ahmad R. Lotfi, Assistant Professor of linguistics at the
English Department of Azad University (Khorasgan, IRAN) where he
teaches linguistics to PhD candidates of TESOL. His research
interests include (minimalist) syntax, second language
acquisition, and Persian linguistics.
|
|
|
|
|
|
