Brien, C. & Sabourin, L.
University of Ottawa
Lexical processing relies on properties of words, which raises questions regarding homonym processing since homonyms include properties for more than one meaning. For instance, train contains syntactic (noun/verb) and semantic (a locomotive/to instruct) properties, each affecting sentence interpretation. A cross-modal lexical-decision task found both readings of noun-noun homonyms (bug: spy device/insect) accessed simultaneously, with contextual information aiding disambiguation at later stages (Swinney, 1979). A reading task found syntactic context influencing noun-verb homonym (train) processing (Folk & Morris, 2003).
Bilingualism adds complexity to this issue. Research suggests the earlier a speaker acquires a second language (L2) the more closely s/he resembles native speakers (Meisel, 1991) since all lexical items may be organized together. Studies have found early bilinguals recruiting the same locus of neurological processing as monolinguals while exhibiting slower reaction times (RT) (Fabbro, 2001), suggesting early bilinguals require more time to process in both languages. Late bilinguals have exhibited secondary language area recruitment and even longer RTs (Paradis, 1998).
The current study involves behavioural and event-related potential (ERP) paradigms investigating age of L2 acquisition effects on the disambiguation of homonyms. Using a cross-modal lexical-decision task, noun-noun and noun-verb homonym processing was compared. Participants were monolingual English speakers and Canadian French/English bilinguals who acquired their L2 at distinct periods of development. All groups revealed effects of word-relatedness (p<.001), and an interaction between homonym-type and word-relatedness (p=.005), suggesting disambiguating syntactic context influenced target-word processing.
Contrary to Fabbro (2001), results show that acquiring both languages simultaneously results in no processing differences compared to monolinguals (p=.987), supporting the hypothesis of both languages organized together. Further, these two groups show no syntactic priming (p=.314) compared to early and late bilinguals (p=.011), suggesting these later bilingual groups are dependent on contextual cues for lexical ambiguity resolution. This study will present preliminary ERP data investigating these effects.