Electrophysiological Studies of Cognates a comparison of FrenchEnglish Bilinguals and EnglishFrench

1
Electrophysiological Studies of Cognates a
comparison of French/English Bilinguals and
English/French BilingualsKatherine J.
Midgley1,2, Jonathan Grainger2 Phillip J.
Holcomb1Tufts University, Medford, MA1
Laboratoire de Psychologie Cognitive - CNRS,
Université de Provence, Aix/Marseille2


results

In bilinguals, how does the processing of
cognates, or words that share orthographic and
semantic overlap across languages (e.g., in
English and French regime and régime), differ
from that of words without such overlap (e.g.,
tree and arbre)? One reason for this questions
importance is that during the process of becoming
bilingual cognates from ones first language (L1)
are thought to provide some of the earliest
footholds into the establishment of the target
language (L2) vocabulary. For example, the
Revised Hierarchical Model of Kroll and Stuart
(1994) predicts that L2 word processing will
initially be very dependent on the L1
word/semantic system, but that as one becomes
more proficient in L2 this dependence will become
less pronounced. In the present ongoing study
we are interested in two interrelated
questions. 1) What ERP component(s) is sensitive
to differences in cognate status for L1 speakers
of French learning English and are they the same
for L1 speakers of English learning French? 2)
Does this sensitivity vary as a function of L2
proficiency?
Figure 2. L2 Cognates vs. L2
Non-cognates in both language groups at the
anterior electrode site Fz (finding A).


discussion
AB The results of this study suggest that
cognate status in L2 modulated two ERP components
in bilingual learners. The first was a posterior
positivity starting as early as 225 ms and
extending as late as 400 ms at some sites. This
component was largest (most positive) to L2 items
that were not cognates and was more prominent in
L1 English speakers. The second component was a
later negativity that started at about 300 ms and
extended as late as 700 ms at some sites. This
negativity, which resembles the N400, tended to
have a more anterior distribution than the
earlier positivity (especially in the L1 English
group), but like the earlier positivity was
larger to non-cognates than cognates. It was more
prominent in the L1 French speakers.
Figure 3. L2 Cognates vs. L2
Non-cognates in both language groups at the
posterior electrode site Pz (finding B).
2
Figure 4. L2 Cognates vs. L2
Non-cognates in both language groups at the
posterior electrode site Pz (finding C).

C Among the L1 English speakers those that were
relatively more proficient in L2 produced ERPs to
cognates in L2 that were dominated by the
posterior positivity. Their ERPs in L1 produced
negativities in this same time range. On the
other hand, English speakers that were less
proficient in L2 produced ERPs to cognates in L2
that more like their L1 responses (i.e.,
dominated by the negativity). This seemingly
counterintuitive pattern makes sense within a
framework such as that suggested by the Revised
Hierarchical model of Kroll and Stuart (1994).
This view assumes that as L2 learners become more
proficient in their second language they start to
process words in L2 within their newly evolving
L2 lexical-semantic system. Since this system is
still relatively incomplete, the negativity
(i.e., the N400) to words processed within this
system is comparatively small. In contrast, less
proficient L2 speakers rely more heavily on their
L1 lexical-semantic system to process words in
L2. For cognates this results in an ERP pattern
(i.e., a larger N400) that resembles their L1
response.

A Both language groups show an increased anterior
negativity in the 300 to 500 ms range for L2
non-cognates compared to L2 cognates with the
French group showing a more prominent
effect. B Both groups show an earlier
positivity to L2 non-cognates compared to L2
cognates at the Pz electrode site (200 to 400 ms
range). The English group shows only the
positivity while the French group shows this
positivity and the later negativity. C In the
more proficient group the L2 exact cognates were
patterned with the L2 close cognates that is
they exhibit more of the early positivity than do
the L1 exact cognates or the L2 exact cognates of
the low proficiency group.
electrode montage



Kroll, J. F., Stewart, E. (1994). Category
interference in translation and picture naming
Evidence for asymmetric connections between
bilingual memory representations. Journal of
Memory and Language, 33, 149-174.
figure 1
contact kj.midgley_at_tufts.edu
This research was supported by NIH Grants HD25889
and HD043251 and by the CNRS, France.