working memory and nativelikeness in the processing of focus structure robert reichle 1 annie...

30
Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages Center for the Interdisciplinary Study of Language and Literacy Northern Illinois University 2 Department of Linguistics University of Kansas

Upload: tatum-norrod

Post on 14-Dec-2015

220 views

Category:

Documents


3 download

TRANSCRIPT

Page 1: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

Working Memory and Nativelikeness

in the Processing of Focus Structure

Robert Reichle1

Annie Tremblay2

Caitlin Coughlin2

1Department of Foreign LanguagesCenter for the Interdisciplinary Study of Language and Literacy

Northern Illinois University

2Department of LinguisticsUniversity of Kansas

Page 2: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

2

Second Language ProcessingSecond Language Processing

• For late L2 learners, L2 processing is difficult• Corresponding features/structures in the

native language do not ensure success in the L2

• Variability is typical• Native-like attainment is not typical• Why?• (e.g., Babcock et al., 2008; Chen et al., 2007; Clahsen et al., 2010; Coughlin & Tremblay,

to appear; Hopp, 2006, 2010; Jiang, 2004, 2007; Keating, 2009; Lew-Williams & Fernald, 2007; Neubauer & White, 2009; Sagarra & Herschensohn, 2010; Sato & Felser, 2008; Silva & Clahsen, 2008)

Page 3: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

3

Proposed ExplanationsProposed Explanations

• Cognitive Processing Hypothesis (CPH) (McDonald, 2006; see also Hopp, 2010)

o Relative to native speakers, L2 learners have

• Lower L2 working memory (WM) capacity• Poorer L2 decoding abilities• Slower L2 processing speed

What about general L2 proficiency and general WM capacity?

All correlated with L2 proficiencyAll correlated with L2 proficiency

Page 4: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

4

Proposed ExplanationsProposed Explanations

• Cognitive accounts:Quantitative (computational) difference

• Shallow Structure Hypothesis:Qualitative (representational) difference

• Declarative/Procedural Model:Qualitative-to-quantitative difference

Page 5: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

5

Event-Related Potentials (ERPs)Event-Related Potentials (ERPs)

• Can shed light on the qualitative vs. quantitative difference

o (e.g., graphs from Osterhout et al., 2006, p. 204; Friederici, Steinhauer, & Frisch, 1999; Friederici, 2002; Hahne & Friederici, 1999, 2001; Hagoort & Brown, 1999; Kaan et al., 2000; Kutas & Hillyard, 1980; Osterhout, Bersick, & McLaughlin, 1997; Osterhout & Holcomb, 1992; Osterhout & Mobley, 1995; Steinhauer, Halter, & Friederici, 1999)

5

Lexical/semantic processing Morphosyntactic/grammatical processing

Page 6: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

6

Event-Related Potentials (ERPs)Event-Related Potentials (ERPs)

• Left anterior negativity(LAN)o phrase structure and

morphosyntax violations are often associated with a left anterior negativity effect

o index increased working memory load stemming from the storage of syntactic dependencies in working memory (e.g. Cowles, 2003; Kluender & Kutas, 1993; Münte, Schiltz, & Kutas, 1998)

(e.g., figures from Neville et al., 1991; Friederici, Pfeifer, & Hahne, 1993; Kluender & Kutas, 1993; Münte, Heinze, Matzke, Wieringa, & Johannes, 1998; Osterhout & Holcomb, 1992)

Page 7: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

7

• Proficiency effects on sensitivity to agreement violations (inflectional morphology)

o Artificial language learning: with language training, N400 P600

o (e.g., Morgan-Short et al., 2010)

o Real language learning: with increasing proficiency, N400 P600

o (e.g., Bowden et al., 2007; Hahne et al., 2006; Osterhout et al., 2006; Rossi et al., 2006; Steinhauer et al., 2006; Gillon Dowens, Vergara, Barber, & Carreiras, 2010; McLaughlin, Osterhout, & Kim, 2004; McLaughlin, Tanner, Pitkänen, Frenck-Mestre, Inoue, Valentine, & Osterhout, 2010; Tanner, Osterhout, & Herschensohn, 2009; Ojima, Nakata, & Kakigi, 2005)

Event-Related Potentials (ERPs)Event-Related Potentials (ERPs)

Page 8: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

8

Focus MarkingFocus Marking

• Focus indicates something new or relevant about the topic

• Focus is marked in different ways across languages (Lambrecht 1994, 2001)

• English: syntactically word accent• French: syntactically word accent

Page 9: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

9

Focus MarkingFocus Marking

Page 10: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

10

• LANs and focus structureo LANs for contrastive focus compared to informational focuso Interpreted as an index of increased WM load as the set of

referents is kept active in WM (Cowles, 2003)

o Instead of fully left lateralized, broadly distributedo This (L)(A)N is found in L1 French; L2 proficiency found to

modulate effect in L2 French (Reichle, 2010; Reichle & Birdsong, under revision)

• Since focal status must be kept active in WM, we predicted a similar LAN for licensed informational focus compared to no focus

Focus MarkingFocus Marking

Page 11: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

11

• Syntax-discourse interfaceo Sorace has proposed that phenomena at the syntax-

discourse interface can pose problems for near-native L2 learners (e.g. Sorace, 2000; Sorace, 2003; Sorace & Serratrice, 2009)

• Thus, this type of focus marking in L2 French may be difficult for learners

Focus MarkingFocus Marking

Page 12: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

12

Present StudyPresent Study

• Proficiency may not be the only factor in determining whether L2 learners show sensitivity to focus structure

• Aims of this studyo Determine whether Cognitive Processing

Hypothesis can account for L2 learners’ ERP signatures

o Can WM capacity explain individual variability?

Page 13: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

13

Present StudyPresent Study

• Hypotheses:

1. In discourse contexts where focus is licensed, an LAN would index the increased WM load required to track the focal referent across discourse, compared to sentences with no preceding focus marking.

2. These LANs would be modulated by WM capacity and L2 proficiency.

Page 14: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

14

ParticipantsParticipants

•10 English L2 learners of French (mean age: 22)

•7 L1 French controls (mean age: 26)

English cloze test: Brown (1980)French cloze test: Tremblay (2011); Tremblay & Garrison (2010)

Page 15: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

15

MaterialsMaterials

• 4 conditions (36 sentences/condition)

o 25% of sentences followed by a true-or-false comprehension question

Page 16: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

16

WM TasksWM Tasks

• Adapted from Waters & Caplan (1996)• Similar sentences, but different lexical

items and different words to recall• Control for frequency and length of the

lexical items to be recalled• (for details, see Coughlin & Tremblay, 2012)

• Dependent variables: word recall in French and Englisho Correlate between the two languageso Do not correlate with proficiency

Page 17: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

17

ProceduresProcedures

Session 1•ERP session in the L2•WM task in the L2•Cloze test in the L2•Language background questionnaire

Session 1•ERP session in the L2•WM task in the L2•Cloze test in the L2•Language background questionnaire

Session 2•ERP session in the L1•WM task in the L1

Session 2•ERP session in the L1•WM task in the L1

Page 18: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

18

Data AnalysisData Analysis

• Time window for negativity: 300-500 ms.• ANOVAs:

2 (Condition) x 3 (Laterality) x 3 (Anteriority)

• Linear regressions between individual variables and effect of negativity/positivity

Page 19: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

19

ResultsResultsL1 English

Licensed: Who is it exactly who makes the soup?Extraneous: Who makes the soup?Target sentence: It is the chef who makes the soup.

Page 20: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

20

ResultsResultsL1 English

Licensed: Who is it exactly who makes the soup?Extraneous: Who makes the soup?Target sentence: It is the chef who makes the soup.

Condition * Hemisphere: p < .01Condition * Anteriority: p = .07

Page 21: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

21

ResultsResultsL1 English

Licensed: Who is it exactly who makes the soup?Extraneous: Who makes the soup?Target sentence: It is the chef who makes the soup.

Condition * Hemisphere * EngWordSpan: p = .06

Page 22: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

22

ResultsResultsL2 French

Licensed: Qui c’est qui fait la soupe?Extraneous: Qui fait la soupe?Target sentence: C’est le cuisinier qui fait la soupe.

Page 23: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

23

ResultsResultsL2 French

Licensed: Qui c’est qui fait la soupe?Extraneous: Qui fait la soupe?Target sentence: C’est le cuisinier qui fait la soupe.

Condition: p = .1

Page 24: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

24

ResultsResultsL2 French

Licensed: Qui c’est qui fait la soupe?Extraneous: Qui fait la soupe?Target sentence: C’est le cuisinier qui fait la soupe.

F3, Fz: Condition*Language*Site, p = .088

Page 25: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

25

DiscussionDiscussion

• English L2 learners of Frencho Quantitively different in the L1 as a function of

WM capacityo Qualitatively the same ERP effect (LAN) in L1

and L2, but with different distributiono In L2, trended towards a significant negativity

effect but did not display a significant effect at the group level

o This marginal negativity was widespread rather than anterior or lateralized, as it was in the L1

Page 26: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

26

DiscussionDiscussion

• English L2 learners of Frencho Consistent with findings that L2 LANs are not

fully localized (laterality/anteriority) until higher levels of proficiency (e.g. Steinhauer et al., 2009)

o Consistent with notion that processing factors such as WM capacity quantitatively modulate brain signatures in the L1, but no direct evidence for the L2 processing account put forth by McDonald (2006)

o Reinforce the notion that phenomena at the syntax-pragmatics interface are likely to pose processing difficulties in the L2

Page 27: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

27

DiscussionDiscussion

• English L2 learners of Frencho In L1 English, WM capacity positively correlated

with LAN effect size. Suggests that higher-WM participants better able to keep the focal status of referent in WM across sentences.

o Lack of similar correlation in L2 French suggests all learners encountered processing difficulty

o Marginal Language*Condition interaction at anterior sites suggests quantitative difference in effect size between L1 and L2

Page 28: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

28

Further ResearchFurther Research

• Higher-proficiency English L2 learners of French• Native speakers of English in a French-speaking

environment• To test other aspects of Cognitive Processing

Hypothesis, look at other individual variables that might correlate with quantitative ERP differences

• Test other areas of morphosyntactic processing, such as SV agreement (Reichle, Tremblay, & Coughlin, to appear; Reichle, Tremblay, Frenck-Mestre, & Coughlin, in progress)

Page 29: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

29

Merci!Merci!

Page 30: Working Memory and Nativelikeness in the Processing of Focus Structure Robert Reichle 1 Annie Tremblay 2 Caitlin Coughlin 2 1 Department of Foreign Languages

30

AcknowledgmentsAcknowledgments

• All the research participants at NIU• NIU Center for the Study of Family Violence and

Sexual Assault• NIU Center for the Interdisciplinary Study of

Language and Literacy• University of Illinois Humanities/Arts Scholarship

Support