ab st rac tContrary to what is stated in much of the literature which is based in large part on Parisian French, many dialects of French still have long and short vowels (e.g. in Switzerland and Belgium). This study had two aims. The first was to show that Swiss French speakers, as opposed to Parisian French speakers, produce long vowels with durations that are markedly different from those of short vowels. The second aim was to show that, for these two groups, vowel duration has a different impact on word recognition. A production study showed that Swiss French speakers make a clear duration difference between short and long vowels (the latter are more than twice the length of the former on average) whereas the Parisian French do not. In an identification study which used stimuli pronounced in Swiss French, it was shown that words articulated with long vowels created no recognition problem for Swiss French listeners whereas they did so for Parisian French listeners. These results are discussed in terms of models of speech perception and word recognition.
Nous proposons un nouveau modèle psycholinguistique informatique du français. Le modèle, intitulé FN5, porte sur la reconnaissance de mots parlés, présentés en isolé (déterminant, adjectif antéposé, substantif) ou en suites de deux mots (déterminant et substantif, adjectif antéposé et substantif). Grâce à un lexique de plus de 17 000 mots, à une architecture connexionniste localiste, et à des mécanismes développés spécifiquement (processeur de position, groupements de connexions, et point d'isolation), le modèle peut simuler, de manière séquentielle ou simultanée, certains effets propres au mot isolé (fréquence, longueur, homophonie) ou à une suite de mots (coupure lexicale, enchaînement, liaison, effacement du schwa). De plus, le modèle tient compte de certaines différences qui existent entre le français standard et le français de Suisse romande (nombre de voyelles, durée vocalique en fin de mot, et statut du schwa). Nous décrivons ce modèle, présentons son interface graphique, et l'illustrons avec des exemples de simulation.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.