Study Guides (238,613)
Canada (115,253)
Psychology (554)
PSYC 341 (26)

Reading 18

12 Pages
Unlock Document

McGill University
PSYC 341
Richard Koestner

Reading 7: Temporal dynamics of late second language acquisition: evidence from event-related brain potentials HIGHLIGHTS • the ways in which age of acquisition (AoA) may affect (morpho) syntax is second language acquisition (SLA) are discussed • event-related brain potentials (ERPs) provide an appropriate online measure to test some such effects • it is concluded that there is little evidence for a strict critical period in the domain of late acquired second language (L2) morphosyntax • proficiency rather than AoA (age of acquisition) seems to predict brain activity patterns in L2 (second language) processing including native-like activity at very high levels of proficiency • a strict distinction between linguistic structures that late L2 learners can vs. cannot learn to process in a native-like manner may not justified INTRODUCTION Important Questions: 1. Does late-acquired L2 involve the same neurocognitive mechanisms as found in native speakers? 2. Do the neurocognitive substrates of late-acquired L2 processing change with increasing L2 proficiency? 3. Is there evidence for a critical period in the acquisition of L2 grammar? 4. What might studies of the acquisition of artificial languages revel about natural language acquisition/processing? • children acquire full language competence despite impoverished input stimuli has led to the suggestion of universal grammar (UG) • universal grammar: an innate language-specific learning capacity ex: is believed to constrain the representation of language, limiting the kinds of grammars that children adopt • access to UG principles may be lost after a critical period (ending around puberty) due to brain maturation • if we compare young/early L2 learners with adult/late L2 learners the late L2 learners display selective problems in: 1. phonology (foreign accent) 2. morphosyntax ... BUT the following are relatively intact: 1. lexical learning 2. lexical-conceptual processing 3. certain aspects of syntax • fundamental difference hypothesis: states that young children acquire their L1 implicitly with UG mechanisms whereas late language learners depend largely on explicit, domain general cognitive functions • Most researchers agree on the existence of critical/sensitive periods in L1, these notions are CONTROVERSIAL in late L2 acquisition and processing: 1. not clear whether late L2 learners can reach native-like competence... whereas some studies suggest that later Age of Acquisition (AoA) of the target language virtually eliminates native like attainment... others suggest that such attainment is possible 2. whereas some studies have reported the predicted discontinuity of AoA effects after the assumed critical period... others have observed a repetitive AoA function incompatible with the claim of a critical period for language learning • one of the strongest pieces of evidence for the critical period in L2 grammar learning compatible with fundamental difference hypothesis - namely that late L2 learners stabilize at some point short of native-like grammar attainment called fossilization 3. AoA effects have been explained in a variety of ways including the loss of language-specific learning mechanisms, the advantage of small working-memory capacities in childhood (the 'less is more' hypothesis) and 'neutral commitment' or 'entrenchment' and consequent interference of L2 by earlier-learned knowledge • researchers in the second language acquisition (SLA) have recently suggested that cognitive neuroscience may provide fruitful answers to these issues • this is reasonable given that both the critical period hypothesis and the fundamental difference hypothesis are essentially neurobiological hypotheses, and their proponents generally imply claims about neuronal differences in language representation between L1 and (at least) late acquired L2 • Age of acquisition and critical period effects do not seem to affect all aspects of SLA to the same extent and can be distinguishes along several dimensions 1. they affect final attainment in L2 rather than rate of learning 2. they concern certain linguistic sub-domains (phonology and morphosyntax) more than others (lexical learning, semantic integration) • Shallow structure hypothesis: it should be possible for late L2 learners to process morphology in a native-like manner, this should not hold for (non-local aspects of) syntactic processing • we first review the relevance of work using event-related brain potentials (ERPs) to study language processing in both L1 and L2 (section II) ... in section III we review evidence suggesting that: 1. Late L2 learners can be shown to display native-like ERP patterns 2. the evidence in favour of the existence of principled limits on the range of L2 learnable properties of natural language - e.g. the shallow structure hypothesis are currently without compelling empirical support • Section IV we outline a hypothetical trajectory for L2 learning as indexed by the temporal dynamics of the emergence of various different ERP responses and suggest some directions for future research EVENT-RELATED BRAIN POTENTIALS (ERPs) • ERPs reflect the real-time electrophysiological brain dynamics of cognitive processes with an excellent time resolution in the range of milliseconds. • distinct ERP components (waveforms with either positive or negative polarity) have been identified for syntactic processing. • difficulties in semantic processing typically elicit centroparietal negativities that peak about 400 milliseconds post stimulus and depend particularly upon bilateral temporal lobe structures • syntactic processes have been linked to two ERP components: an early left anterior negativity (LAN) and a late centroparietal positive shift (P600) • violations of phrase structure and of morphosyntactic constraints often yield LANs-which have been hypothesized to be linked to rule-based automatic parsing • neural generators of the LAN have been identified in the (left) prefrontal cortex, notably Broca's area • if indeed LANs are generally linked to automatic, implicit grammar processing, one would predict that these early ERP components in particular should be the most difficult to elicit in late L2 learners (i.e. more likely to be constrained be AoA or critical periods than, for example, the p600) • ERP studies can be broadly divided into three kinds of experimental design: 1. group studies contrasting L2 learners with native speakers or different groups of L2 learners 2. longitudinal studies testing changes within participants while they acquire the target language 3. paradigms using artificial miniature languages that can be learned to high proficiency with a short period, allowing one to cover large ranges of proficiency changes within the same participants and at the same time control for many potential confounds • ERP studies argued in favour of the critical period hypothesis (i.e. a late age of acquisition results in fundamental differences in the way syntactic information is processed) • Weber-Fox and Neville explored whether the neurocognitive processes underlying semantic and syntactic processing in bilinguals are impacted by AoA • they found that when presented with semantically anomalous sentences, all of the bilinguals -regardless of AoA-elicited an N400 response, although this effect showed a delayed peak latency in those who learned English after age 11 • this suggests that semantic processing strategies rely on largely identical neurocognitive mechanisms in both one's L1 and L2, but that for late L2 learners lexical retrieval and/or semantic integration may be somewhat slower • in response to syntactic phrase structure violations the early L2 learners (AoA <11 yrs old) elicited a left lateralized negativity similar to that of native speakers • for the late learners (11-13 and 16 plus yrs), however this negativity was bilaterally distributed, and for the oldest group it was actually greater over the right hemisphere than the left • only the early bilinguals elicited a LAN similar to that of native speakers • similarly, the P600 elicited by the early learners (AoA <11 yrs) was identical to that of native speakers, was delayed in those who acquired English between 11-13 yrs, and was not present at all for the oldest group of L2 learners • while the neurocognitive basis of semantic processing may be relatively unaffected by AoA, the neural mechanisms underlying syntactic processes appear to be sensitive to delays in L2 acquisition • the fact that age-related differences were observed in both the LAN and P600 responses provided powerful evidence for the claim that syntactic processing relies on fundamentally different neural-cognitive structures in late L2 learners compared to native speakers and early L2 learners PROFICIENCY AND NATIVE-LIKE ATTAINMENT • AoA was negatively correlated with L2 proficiency in Weber-Fox and Neville (1996) as measured by standardized tests of English grammar, self-reported proficiency and acceptability judgement accuracy • it is thus impossible to determine whether the lack of LAN/P600 effects in late learners was due to AoA or proficiency • the same difficulty holds for Hahne and Friederici (2001) who found no differences between violation and control conditions but who did not report proficiency levels among the late Japanese-German bilinguals they tested • it was shown that at relatively low levels of L2 proficiency, whereas in learners with very high levels of L2 proficiency that had acquired their L2 early or after puberty • however these findings were relatively unspecific in terms of which psycholinguistic sub-processes were actually reflected by proficiency-dependent brain activation, and it is important to note that fMRI (and PET) lack the temporal resolution to detect fine-grained differences in the rapid dynamics of language processing • up until 2001 there had been no ERP study similar to Perani et al.'s neuroimaging studies ...although the available ERP evidence at the time generally supported the notion of a critical period for syntactic (but not semantic) processing, in all of the relevant studies second language learners were not even close to the proficiency level of native speakers • Friederici et al. (2002) presented the first ERP study directly addressing the AoA/proficiency confound... they trained adult (post-critical-period) participants in an artificial miniature language (BROCANTO) to a native-like level of proficiency in both production and perception (95% accuracy criterion) • unlike artificial grammars in previous studies, BROCANTO both conformed to UG requirements and could be used to actually communicate complex propositions • sentences referred to the moves of a complex (chess-like) computer-based board game for two players... participants practised the language BROCANTO by verbally communicating their moves while playing against each other at separate computer monitors... these practice sessions were distributed across several days and lasted up to 25 hours in total until participants reached criterion proficiency in both production and comprehension • this ERP finding was related to high performance in both a visual probe verification task (89% correct) and a grammaticality judgment task (93%)... in contrast, a control group that had received only vocabulary but no grammar training performed well only in the probe detection task (86%) but not in the grammaticality judgement task (58% correct)... as expected this control group did not display any ERP differences between grammatical and ungrammatical sentences • this study raised the following questions: 1. Can qualitative neurocognitive changes during language learning be monitored within participants using ERPs? 2. What are the limitation of SLA models based on artificial language paradigms? ... a number of recent studies have provided some preliminary answers... ERP STUDIES SHOWING NATIVE-LIKE PATTERNS IN L2 • key question raised by the BROCANTO findings is whether similar effects of native-like ERP patterns in late L2 acquisition can be replicated in natural language learners • an ERP study conducted in our own lab with visually presented stimuli examined late French and Chinese learners of English at two different proficiency levels (high vs. low) with respect to their processing of syntactic word category violations and contrasted them to native English speakers • all L2 learners acquired English af
More Less

Related notes for PSYC 341

Log In


Don't have an account?

Join OneClass

Access over 10 million pages of study
documents for 1.3 million courses.

Sign up

Join to view


By registering, I agree to the Terms and Privacy Policies
Already have an account?
Just a few more details

So we can recommend you notes for your school.

Reset Password

Please enter below the email address you registered with and we will send you a link to reset your password.

Add your courses

Get notes from the top students in your class.