Retour vers liste

Détail de la contribution

Auteur: Seiya FUNATSU

Co-Auteur(s): Masako FUJIMOTO, The National Institute for Japanese Language and Linguistics

Is the mechanism of vowel epenthesis language specific or language universal?: an EMA study

Abstract/Résumé: Vowel epenthesis in consonant clusters appears in many utterances of non-native speakers in foreign languages. For instance, English speakers pronounce “zd” and “zb” in Czech as /z(schwa)d/ and /z(schwa)b/, respectively. Japanese speakers often insert vowels in consonant clusters (e. g. English word “try” may be pronounced as /torai/), possibly because Japanese does not allow the consonant clusters due to its phonotactics. Previous studies (Davidson, 2005, 2006) showed that vowel epenthesis for native English speakers would arise from mis-timing in C1C2 sequence or the insertion of a lexical vowel. Whereas, vowel epenthesis in consonant clusters in Japanese originated from the perception of an “illusory vowel” (Dupoux, 1999, 2001). Consequently, vowel epenthesis in English speakers is caused by production, whereas for Japanese speakers it is caused by perception. Is the mechanism of vowel epenthesis language specific? Or, rather, is the mechanism of vowel epenthesis language universal? In an attempt to answer this question, we investigated the pronunciation of consonant clusters in native Japanese speakers. The present study aims to elucidate the production mechanisms of vowel epenthesis in consonant clusters using an electromagnetic articulograph (EMA). The target languages were Japanese and German. Japanese does not allow consonant clusters, while German does. Two Japanese speakers and two German speakers participated in this experiment. Speech samples were 4 nonsense words (bnaht, pnaht, gnaht, knaht) in both Japanese and German. These words (X) were embedded in German sentences “Sage X.” (“Say X.”). These sentences were uttered 8 times in Japanese and 10 times for German speakers. Tongue tip displacements and moving times from first consonants (/b/, /p/, /g/, /k/) to second consonant /n/ were measured. For Japanese speakers, tongue tip displacements from the first consonants to the second consonant in clusters were significantly larger than those of German speakers (p<0.001). Also, the moving times of tongue tip for Japanese speakers were significantly longer than those for German speakers (p<0.001). These results suggested that the coarticulation between first and second consonant for Japanese speakers would be weaker than that for German speakers. Accordingly, we inferred that the production of consonant clusters was accompanied by vowel epenthesis in consonant clusters for native Japanese speakers. (This study was supported by a Grant-in-Aid for Scientific Research (No. 24520434, 23520539) from the Japan Society for the Promotion of Science.)