Acesso lexical e produção de fala em bilíngues português-espanhol e espanhol-português (original) (raw)

This study aimed at analyzing the mechanisms of lexical access and speech production in a group of bilingual Portuguese-Spanish and Spanish-Portuguese speakers. More specifically, our goal was to observe the performance of these bilinguals in L2 picture naming tasks in order to: (a) identify the effects of interference (semantic, identity and phonological/orthographic facilitation), (b) investigate if the status of cognate words could affect the lexicalization process, (c) verify the existence of interaction between languages, and (d) determine whether there was a correlation between reaction time, accuracy and the degree of similarity between languages. Two tasks, in which the nature of the relationship between pictures and words and the types of words involved were controlled for, were developed within the picture-word interference paradigm. The sample was composed of 23 bilinguals, who were asked to name 90 pictures that appeared twice in the tests (once in the related condition and once in the unrelated condition). To assess each interference effect (semantics, identity and phonological/orthographic facilitation), 30 pictures, which were subdivided into three groups, with 10 pictures each, depending on the type of word (cognate, non-cognates and false cognates) were used. Thus, our experiment had an orthogonal configuration. Overall, the results show the specificity of selection mechanisms, since the nominations, as predicted, were faster in the related condition, in the effects of cross-linguistic identity and of phonological/orthographic facilitation; in addition, participants were slower in the same related condition, for semantic interference. These results seem to give support for the hypothesis that predicts that lexical selection is language specific. Another important result concerns the role played by the cognate status of words in the lexicalization process, since statistical significance was found either alone or in comparisons and interactions with other effects. These results show that, when assessing the types of interference effects, one must consider the types of words involved. In addition, important indications that the principle of interactivity is functional among languages were found, since false cognates showed faster reaction times in the context that favored sublexical activation (phonological facilitation effect). We argue that this effect is due to the occurrence of bidirectional activation, which comes from the phonological segments activated by the L1 translation (that does not correspond to the same concept in L2), which in turn increased the activation of the target lexical node in the L2. Another relevant finding is related to the linguistic similarity = ease of acquisition of L2 formula, which appears to be inconsistent due to an attentional overload caused by the similarity between the languages. This assumption is supported by the reaction time and accuracy results that were found between the cognates, which were not correlated, since there were contexts in which naming scores were faster but accuracy was lower. This decrease of correct responses may be a result of attentional overload, caused by the similarity of words. However, further investigation is needed to assess these effects.