Back to list

Detail of contribution

Auteur: Toshio MATSUURA

Titre:
Presence or absence of accent shift in Nagasaki Japanese


Abstract/Résumé: This study examines the pitch accentual characteristics in Nagasaki Japanese (henceforth, NJ), which is a dialect spoken in the central part of Nagasaki Prefecture. NJ has a contrast of presence or absence of the pitch accent, which is traditionally treated as a contrast of word tones: HL-tone vs. LH-tone (Hirayama 1951). Standard Japanese (henceforth, SJ), which also has a contrast of presence or absence of the pitch accent, exhibits an accent shift rule which shifts the pitch accent on the second mora of a long syllable to the preceding mora (McCawley 1968). According to perceptual impressionistic description on NJ accent (Hirayama 1951, Sakaguchi 2001, and others), NJ does not have an accent shift rule, so the pitch accent can be located on the second mora of a long syllable even if the mora consists of only a consonant. This study confirms these descriptions by using acoustic measurements. Test words are designed to investigate whether the pitch accent on the second mora of a long syllable shifts in NJ. Long syllables of Japanese are classified into three types: a syllable with a long vowel (VV), with a moraic nasal (VN), and with a moraic obstruent (VC). All test words contained a long syllable initially, such as booto 'boat', panda 'panda', and matto 'mat', and embedded in short sentences (e.g., booto-ni notta '(He) boarded a boat'). Two native speakers of NJ and one native speaker of SJ participated in the recording. The subjects read 15 test sentences three times. First, we compared the rate at which the peak pitch appeared on the vowel vs. on the following moraic nasal. For the NJ speakers, the peak pitch was almost always on the moraic nasal, while it appeared on the moraic nasal only half of the time for the NJ speakers. Second, we compared the normalized-time of a peak pitch against a whole syllable length of long vowels and moraic nasals. The peak pitch for NJ was located later than for SJ in all test words. Finally, we compared the highest pitch before a moraic obstruent with the highest pitch after a moraic obstruent since moraic obstruents do not have fundamental frequency in most cases. Whereas the highest pitch after moraic obstruents is lower than that before moraic obstruents in SJ, the highest pitch after moraic obstuents is higher than that before moraic obstruents in NJ. The results of the above measurements reflect the presence or absence of accent shift. In every case, the peak pitch in NJ is located later than that in SJ. This means that although the pitch accent shifts to the syllable nucleus in SJ, it does not in the case of NJ.