CIPsi, School of Psychology, University of Minho
How the bilingual mind encodes letter order during reading is a question that is receiving increasing attention in the field of word recognition. Yet, the existing studies have not led to conclusive results (e.g., Comesaña, Coelho, Oliveira, & Soares, 2018; Font, 2001; Li, Baugert, & Schwartz, 2015; Witzel, Qiao, & Forster, 2011). For instance, Font (2001) found that the recognition of cognate words (translation equivalents that share their form) is affected by the position of the deviant letter - e.g., the final letter in the English-Portuguese pair paper-papel. These results question the validity of the input-coding scheme proposed in the Multilink model (the most relevant computational model of bilingual word recognition, Dijkstra et al., 2018); as this model assumes that letter position is perfectly encoded. Nonetheless, using a carefully controlled design, Comesaña et al. (2018) failed to replicate such modulations. The aim of this study was to test the Multilink model using three standard tasks (masked priming lexical decision, two-alternative forced-choice, and same-identity) with different bilingual populations (Portuguese-English, Catalan-Spanish, French-English, and English-French) and their respective controls. The experimental stimuli consist of a set of cognate words that vary in the position of the deviant letter. The deviant-letter position of cognates was manipulated while several linguistic variables affecting word recognition were controlled within and across-languages. The results revealed that bilinguals exhibit an advantage for first letter position, an effect that is well-known in the monolingual literature, although this effect is modulated by the language considered (first or second).