Robert J. Hartsuiker is professor of Cognitive Psychology at the Department of Experimental Psychology, Ghent University. He is the President-Elect of the European Society for Cognitive Psychology. He was the editor in chief of the Journal of Cognitive Psychology and an associate editor of Psychological Science. His research focuses on language production and comprehension, control and monitoring of speech, and multilingualism. He has published more than 100 journal papers on these topics in a wide array of journals, including Psychological Bulletin, Psychological Science, Cognition, and Cognitive Psychology.
When are syntactic representations shared across languages?
Robert J. Hartsuiker
Department of Experimental Psychology, Ghent University, Belgium
Cross-linguistic structural priming effects suggest that proficient late bilinguals share syntactic representations across their languages (e.g., Hartsuiker, Pickering, & Veltkamp, 2004). Such priming can occur even when the syntactic structures are not fully identical in the two languages. The shared representations must therefore abstract across such differences. What then determines whether representations are or are not shared across languages?
In my keynote I will discuss results from cross-linguistic structural priming studies (sometimes combined with artificial language learning) with respect to two factors: second language (L2) proficiency and word order overlap. Studies that considered L2 proficiency showed stronger priming stronger in more proficient participants, suggesting that learners start out with separate representations and start sharing them as they go along. I then turn to word order, and argue that the field sometimes incorrectly assumes that structures can only be shared across languages (and hence prime each across languages) if they have the same word order in each language.
In fact, there are several counterexamples to that generalization in the literature. Instead, I will propose the syntactic preemption hypothesis, according to which a structure S in one language and a counterpart S’ in another language can only have a shared representation if neither language distinguishes between S and S’. Thus, priming can occur between a passive with SVO structure in one language and one with SOV structure in another language, but not if either language allows both word orders. In that case, the need to distinguish between different word orders preempts the formation of an abstract, word-order independent representation for the passive.
I will discuss structural priming studies using picture description, translation, and artificial language learning tasks that are consistent with the syntactic preemption hypothesis. I will conclude with suggestions to further test this hypothesis.