Effects of consonantal environment and speech rate on vowel devoicing are examined from glottal opening patterns of an Osaka subject. Results demonstrated that for the initial consonant of /CVC/, the glottis widely opened both for stops and fricatives, which presumably facilitates devoicing. For the second consonant, the glottis widely opened for fricatives but not for stops. However, vocal fold vibration of the preceding vowel continued into the consonants only for fricatives. Thus, fricatives are likely to suppress devoicing in intervocalic position. For the consonant /h/, continuation of vocal fold vibration lasts until the end of the consonant, which will strongly suppress devoicing. A sequence of fricatives suppresses devoicing presumably due to faster glottal adduction for fricatives than stops. The distance of glottal openings between preceding and following consonants becomes shorter in fast speech, which may facilitate devoicing.

抄録全体を表示

英語の2音節名詞は第1音節に強勢を持つ語が大多数であり,2音節動詞は第2音節に強勢を持つ語が多い。このような傾向は語強勢の典型性と呼ばれる。先行研究ではゲーティング課題により,英語母語話者及び学習者が,典型的な強勢型の語を非典型的な強勢型の語よりも短い提示時間で認知することが示された。しかし,学習者の聴解力の影響など未検証の課題も残された。そこで本研究では,学習者の第一言語を日本語に統一し,聴解力上位群と下位群の比較を含めた再検証を行った。その結果,日本人英語学習者はやはり音声単語認知の際に語強勢を手掛かりとして利用しており,典型的な強勢型の語を非典型的な語よりも短い提示時間で認知することがわかった。また聴解力は典型性効果の度合いに影響を及ぼすことが明らかになった。

抄録全体を表示

This paper describes the conditions of how the merger in Class 4 and 5 bimoraic nouns is developing around the Keihan area. This occurrence of the merger originated from Osaka and has clearly spread around the Kinki and Shikoku areas. However, the diffusion of the merger has not affected the Tokyo-type accent. It can be said that the change of Class 4 from LLH to LHL causes the disappearance tendency of the low-beginning unaccented category and also has a disproportionate effect on Keihan accent system due to this change.

抄録全体を表示

This paper is a descriptive study of the accent of two-mora + two-mora native Japanese compound nouns in the Kyoto dialect. Some characteristics of the accent are the same as those of long compound nouns, such as three-mora + two-mora words, while the others are different from them. These similarities and differences show the continuity and discontinuity between the accents of short compound nouns and long compound nouns.

抄録全体を表示

In the Keihan-type accent system, which includes the Osaka dialect, the lexically specified pitch patterns (i.e., accents) of words are classified into the following two categories: high-beginning and low-beginning. During the pronunciation of isolated words, the low-beginning type has two characteristics, which are an initial low pitch, and a gradual pitch rise from the beginning to the accent nucleus. A number of researchers claim that the latter characteristic is more consistent and therefore more pertinent. This paper provides a detailed phonetic description of the behavior of this type of accent in various sentence conditions and argues the relative importance of the two characteristics. An acoustic analysis performed on 15 sets of sentences uttered by a total of 26 speakers revealed the following: An initial low pitch is a firm characteristic that does not easily disappear in a sentence context; A gradual rise may be acoustically flat or even slightly descendent in some conditions, but it is also a firm characteristic that distinguishes low-beginning words from high-beginning words. These results suggest that it is not legitimate to conclude that one of the two characteristics is more important.

抄録全体を表示

In Japanese accentology, it has been under dispute whether kakoo-shiki, or falling pitch register, should be recognized as a phonological category, independent of so-called kooki-shiki, or high-beginning pitch register. A production study was carried out on 7 dialect groups with 3 different shiki systems, where cross-dialectal comparisons are made on the properties of pitch (fundamental frequency) contour patterns of different shiki types. The results reveal that the kakoo-shiki can be characterized as having (i) steeper pitch fall toward the end of the word and (ii) later onset of the pitch fall, as compared with kooki-shiki. Both of these tendencies were confirmed with the second production study where the mora length of the experimental words was varied. This set of findings suggests that it is possible to define kako-shiki by examining the acoustic (pitch contour) properties of the relevant shiki types.

抄録全体を表示