In standard Japanese, high vowels between voiceless consonants are regularly devoiced. In order to investigate various factors influencing vowel devoicing by Taiwanese learners of Japanese, a production experiment was carried involving isolated words and words in context with speakers differing in language proficiency. Comparison was done of the rate of devoicing in Taiwanese and Japanese speakers reading a fixed Japanese text. The native Japanese participants were grouped into speakers of Tokyo and Kinki dialects, respectively. The Taiwanese speakers were grouped according to their proficiency of Japanese: advanced and elementary Japanese. The results indicated that devoicing rates of advanced Taiwanese learners showed a pattern similar to the Kinki dialect speakers. Moreover, the devoicing rate for elementary learners was lower than that for advanced learners. The results suggest that vowel devoicing skills can be acquired as proficiency in the second language increases.

抄録全体を表示

This study examines factors which affect the recognition of the intervocalic tense fricative by the groups of native Korean speakers and of Japanese learners of Korean language. The latter consists of the subgroups of the beginners and the advanced level learners. The stimuli consist of the variety of sounds transformed from either of a Korean tense word and of a non-tense word. The length of the pair of fricatives and of preceding vowels were gradually extended and contracted. Participants were requested to listen to the stimuli and judge whether they recognized the sounds as tense consonants. Results showed that the length of the preceding vowel had no significant influence on the recognition either of native Korean speakers or of Japanese learners of Korean language. Regarding the length of fricative, however, the recognition rate of the tense consonant was found to be increased in all groups when the length of the fricative section was prolonged. While native Korean speakers showed the higher recognition rates in stimuli produced from the tense word than those produced from the non-tense word, Japanese learners, on the other hand, showed slight difference between the two kinds of stimuli regardless of their Korean language proficiency level. These facts suggest that while Korean native speakers take factors other than the length of fricative in their judgment of the tense consonant into consideration, Japanese learners tend to rely solely on the length of fricative.

抄録全体を表示

This study examines the influence of word accent on high vowel devoicing in the Keihanshin area. Thirty-one native Japanese speakers (of old, middle-aged, and young generations) who had grown up in Kyoto, Osaka, and Kobe participated in this experiment. The 117 words with four types of accents (HH, LL, HL, and LH) were analyzed. These four types are the pitch configurations of the first two morae in a word with two, three, or four morae. The high vowel of the first mora has a possibility of devoicing. Two main findings emerge from the study. First, differences among the four types of accents do not significantly affect the devoicing rate of the speakers who have traditional Keihan-shiki accents. Second, the devoicing rate of the older generation is significantly lower than that of the middle-aged or younger generation.

抄録全体を表示

In this paper, a panoramic view is given on the relationship between three different laryngeal features found in Eastern Asian languages. The features under investigation consist of manner of articulation of the consonant, phonation type and pitch, all of which are controlled by larynx. Close correlation between these features is observed, especially among tone languages distributed across the center of the East Asian region. Moreover, examples from surrounding languages (i.e. Japanese and Korean, as well as Austronesian and Austroasiatic languages) are also surveyed.

抄録全体を表示

This paper explores multiple modes of articulation such as preaspirations, pitch tones, creaky and breathy voices and phonation types (registers) attested in the Tibetan dialects. From the historical viewpoint through the comparison with the Written Tibetan (WrT) form, most of these characteristics except for the distinction of phonation types are related to an ancient segmental features such as differences of voicedness and consonant clusters, and they have developed through an independent sound change in each dialect (group). The origin of phonation types attested in Shar Tibetan has little relation with WrT forms and requires further investigation.

抄録全体を表示

The Northern Vietnamese dialect is known to involve laryngeal configuration as an integral part of its tonal system. This work analyses the relationship between pitch and voice quality in another variety of Vietnamese spoken in the North-Central region, named Lam River Speech. Acoustic data of fundamental frequency (F0) and duration in the citation form of their unstopped tones is presented in the form of mean F0 time-course; and voice quality is determined by perceptual observation of ten female speakers. Normalised mean F0 and various voice qualities are discussed from viewpoints of both between-speaker and within-speaker differences. The within-speaker variation points to the fact that the degree of laryngealisation is correlated with relative pitch of a speaker. Intrinsic vowel F0 data supports this fact, where laryngealisation is likely to occur in low vowels that carry lower F0. It is concluded that pitch is a primary cue in the LRS tonal system, whereas laryngealisation is a secondary one. Ultimately, this usage of laryngealisation is different from the well-studied Northern Vietnamese dialect.

抄録全体を表示

本研究はタイ標準語の声調生成に関わる喉頭筋の活動について,EMG(筋電図)を用いた先行研究(Erickson 1976, 1993)を基礎としながら,とりわけ下降調生成のメカニズムについて音声生理学的な観点から再検討を加えたものである。タイ語の各声調は,輪状甲状筋(CT)および胸骨舌骨筋(CT),胸骨甲状筋(ST),甲状舌骨筋(TH)等外喉頭筋の運動に関して,それぞれ異なったパターンを示す。CTの活性化は基本周波数(F0)の上昇をもたらし,その不活性化はF0の下降をもたらす。下降調生成においては,CTの不活性化によってF0が中音域まで下げられ,外喉頭筋の活性化によって中音域以下への下降がもたらされると考えられる。喉頭筋の活動パターンに基づき,声調生成に関わる喉頭調節の様々な方策について検討した。

抄録全体を表示

歌声を形成する言語情報(歌詞)と非言語情報(音高・音長情報を含む旋律及び声質)のうち非言語情報である声質に焦点を当て,電気声門図及び音声信号を基にチベット声明の発声における特徴の解明を試みた。その結果, 声明の音源パラメータは低声門開放率及び高声門開閉速度率の特徴をもち,それらに対応する音響特徴としてH1-H2及びH1-A3が共に低い数値を示した。これらは声明におけるりきみ発声の特徴を表している。次に,電気声門図波形及びスペクトル解析により,単なるりきみだけでなく,声門上構造物の振動によると推測されるザラザラ感(harsh)のある発声も確認された。ここではその周波数は声帯振動と同じF0であった。声明における音域は通常発話時の半分の音域に値する2半音であり,音高も発話時より2半音低めであった。このような特徴からチベット声明は喉詰型発声の伝統を汲むことがわかった。

抄録全体を表示