Mulder, K. 1, 2 , Dijkstra, T. 1 & Schreuder, R. 1
1 Donders Institute for Brain, Cognition and Behaviour, Radboud University Nijmegen, The Netherlands
2 International Max Planck Research School for Language Sciences, Nijmegen, The Netherlands
Monolingual studies have shown that response latencies in visual word recognition tasks are affected by the number of morphological family members a target has, i.e. its morphological family size (MFS). For example, WORK ROOM and HOMEWORK are members of the morphological family of the word WORK. It has been found that words with larger morphological families are processed faster and than words with smaller morphological families. Since it is likely that bilinguals acquiring a second language (L2) develop morphological and semantic relationships between words from their L2, the MFS of L2 words is expected to play a role in L2 word processing. In addition, given the evidence in favor of language non-selective access, even the MFS of their first language (L1) should affect L2 word processing. The present study explores this bilingual terra incognita by using cognates. Cognates are interesting because their form and meaning overlap might strengthen MFS effects or, in contrast, cancel them in particular experimental conditions. Observing activation effects of the morphological family members of cognates in both languages would provide evidence for an integrated lexicon because cross-linguistic effects would only be observed when morphological family members of both languages interact. MFS effects in cognates were examined by testing Dutch(L1)-English(L2) bilinguals in English lexical decision and English-Dutch language decision. The results suggest that cross-linguistic MFS effects in cognates are dependent on both task requirements and the degree of form overlap in cognates (identical vs. non-identical).