Abstract
Evidence is increasing that the first stages of word processing in bilinguals are language-independent. Early findings that seemed to point to the existence of independent lexicons and a language selective access mechanism, can be explained cither because participants were able to use loss-level orthographic and phonotactic cues about the language of the stimulus, or because the dependent variable was not sensitive enough to reveal the interactions. Recent research based on lexical decision and masked priming indicates that, in the very first stages of word recognition, the words of the different languages mastered by a bilingual, behave as if they were words of the same language.