Comesaña, M. 1 , Soares, A. 1 , Sánchez-Casas, R. 2 , Frade, S. 1 , Rauber, A. 3 , Pinheiro, A. P. 1 & Fraga, I. 4
1 University of Minho (Portugal)
2 University of Rovira i Virgili (Spain)
3 Catholic University of Pelotas (Brazil)
4 University of Santiago de Compostela (Spain)
Two major positions have been proposed in order to explain the differential processing of cognate vs. non cognate words: a lexical-morphological hypothesis (Davis et al., 2010), according to which the differential processing observed in cognate words is due to their special morphological representation in bilingual memory; and a symbolic, localist connectionist framework (Dijkstra et al., 2010) that emphasizes the cross-linguistic similarity of cognate words. In order to contrast these hypotheses, we examined the role of phonological and orthographic similarity in the processing of cognate words, as well as the duration of SOA, by recording electrophysiological (event-related potentials -ERP) and behavioral data. One-hundred and ninety-two words (96 cognate vs. 96 non-cognate words) were selected based on to their orthographic -O- and phonological -P- overlap and matched on frequency, length, grammatical category, thematic structure, and orthographic and phonological neighborhood. Forty-eight proficient European Portuguese-English bilinguals performed a silent reading task combined with a masked priming paradigm. The results showed that the processing of cognate words was modulated by both phonological and orthographic overlap, although the orthography effect was more pronounced for the longer SOA.