Doctoral Dissertation Research: Consequences of Translation Ambiguity for Bilingual Word Meaning

University Of Pittsburgh, Pittsburgh PA

Investigators

Abstract

Does a second language influence how bilinguals process their native language? The current project examines this, specifically focusing on the way that bilinguals interpret the meanings of words. This project capitalizes on the prevalence of words that can be translated in more than one way across languages. For example, the English words "watch" and "clock" are both translated into Spanish as "reloj". Preliminary research suggests that these words are thought of differently by bilinguals and monolinguals. Therefore, this project tests whether bilinguals consider word pairs like "watch-clock" to be more or less related than monolingual speakers do. Such subtle differences between bilinguals and monolinguals in the meanings they assign to words have consequences for comprehension and production, and also can inform our understanding of the interconnectivity between the two languages of bilinguals. To examine such cross-language influences, the current project employs two methods that allow researchers to examine language processing as it unfolds on a very rapid timescale: eye-movement tracking and brain activity monitoring, which both provide an index of how (or whether) processing is disrupted. In one experiment participants will be presented with English sentences in which, on some trials, a word will be replaced with a word that shares its Spanish translation (for example, "John wore the clock on his wrist"). Bilinguals' and monolinguals' reading behavior of such sentences will be compared. If two words that share a translation become more similar to bilinguals, their reading behavior, as indexed by eye movements, should show reduced disruption to the anomaly compared to monolingual English speakers. In a second experiment, bilinguals will be presented with pairs of words that either share a translation or do not, while their brain responses are recorded. Increased relatedness is expected to facilitate processing of pairs that share a translation compared to pairs with different translations, for bilinguals but not monolinguals.