We hypothesised that, in the noun-first conditions, English-speaking
monolingual participants would have difficulties in ignoring No-Go stimuli and
withholding decision making because this would involve anticipating critical
information from an adjective presented after a noun, which violates English
grammar. For example, following the picture of a red book, and after seeing the
first word car, monolingual participants were expected to have difficulties in
waiting for the adjective, as instructed. In sum, monolingual participants were
expected to make a decision regarding response execution in all cases, that is
whether the noun presented a first word matched and mismatched the picture.
This would translate in N2 peaks of similar amplitude in the two conditions.
Critically, we hypothesised that, in noun-first conditions, participants in
the Welsh-English bilingual group may be better able to withhold decision-
making following a mismatch noun than English monolinguals. If Welsh gram-
mar is indeed co-activated in the mind of bilingual participants, during the
processing of English sentences, then we may be able to find evidence that the
bilingual group is more open to the possibility of an adjective following a noun
than the monolingual group. Co-activation of Welsh grammar would mean
that, upon seeing a No-Go first word stimulus (car in our example), bilingual
participants could be more open to the possibility of an adjective (e.g. red) pre-
sented in second position and resolving decision making for this trial. In other
words, co-activation of Welsh, of the type found in studies on lexical activation,
may enable bilingual participants to anticipate the adjective. The N2 in this
case was expected to be greater for nouns presented as first words and mis-
matching the picture, because this required withholding decision making
(No-Go stimuli), whereas a matching noun indicated a Go trial (Go stimulus).
In sum, we expected that the ERPs of the monolingual group would reveal
no difference in the N2 domain between Go and No-Go conditions when the
first word was a noun, while in the bilingual group we expected a significant
difference in the amplitude of the N2 component between Go and No-Go
conditions, which would constitute evidence of parallel activation of Welsh in
the bilingual group. If, on the other hand, Welsh grammar is not activated,
bilingual participants would behave like monolinguals, their ERPs revealing
no significant difference between Go and No-Go conditions in the N2 domain.
Finally, note that, in our Go/No-Go design, the critical stimulus was the
word preceding the syntactic violation, and not the word that is responsible
for the violation. Consequently, the usual measures linked to the study of
syntactic violations (i.e. (E)LAN and P600) do not apply, and the expected
modulations fall in the N2 range.
Design and procedure
Prior to the experiment, participants were instructed to wash their hair
using baby shampoo so as to minimise scalp impedance. They were then
accompanied to a soundproofed room with dimmed lights, where they were
222 Part 5: The Bilingual Brain