人間が音声言語を理解する仕組みは,脳や聴覚などの観点から考えると基本的には普遍的なものと考えられなくもない。しかしながら,各言語の音韻構造が言語によって異なることから,この違いが音声言語の処理過程に反映することが十分有り得ると考えられる。この問題を理解するには,単一の言語の話者による研究もさることながら複数の言語の話者に対して同一の条件による対照研究の手法の方がより有効であることを,(1)連続音声における分節の単位,(2)語彙認識における超分節素の役割,(3)音素の処理過程,(4)心内語彙表示の音韻単位など,これまで実施した研究に基づいて論じる。

抄録全体を表示

Electromyographic data on Thai and Suzhou Chinese are presented and physiological mechanism of the pitch control is discussed. Activation of the extrinsic muscle, sternohyoid (SH), contributes to low (or non-high) pitch initiation of a syllable. It also contributes to a relatively slow pitch fall of the falling tones, particularly in the non-high pitch range. Relaxation of the muscles are highly relevant in the initiation of pitch fall and rise: a quick fall from the high pitch range is initiated by suppressing the activity of cricothyroid (CT), and a pitch rise from the low pitch range is initiated by suppressing the activity of SH. The vocalis muscle (VOC) functions reciprocally with CT in the maintenance of the high pitch as well as in the vocal terminationof a syllable.

抄録全体を表示

In languages such as Japanese, hearers often encounter a sequence of more than two identical vowels. We investigated how, and to what extent a hearer can successfully segment each mora in such consecutive vowel series. In the following experimental reports, we particularly discuss effects of prosodic information such as pitch pattern and rhythm of speech. We argue that various kinds of information (pitch, rhythm, duration, lexical information) contribute to the segmentation of three or more consecutive vowels.

抄録全体を表示

The present paper compares the phonology of two typologically-unrelated languages, Japanese and Winnebago, and characterizes their 'accent' in terms of phonetic interpretation, basic tone melody, tonal rules, sonority hierarchy, prosodic structure (mora, syllable, foot, etc.), and word minimality. I will show that in spite of the existence of their apparently distinct typological and structural properties, these languages nonetheless exhibit striking similarities when seen from accentual typology in the generative paradigm, or universal grammar. My arguments will lead to the following specific claims: 1) they both are pitch-accent languages although they are different in some tonal rules; 2) they both are mora-counting, syllable-accenting languages and constitute a mirror-image pattern with distinct moraic structures; 3) in theoretical terms, they have moraic feet and foot extrametricality in common, which, however, are assigned from the opposite directions; and 4) they each have a distinct minimal word requirement, which is evidenced by the minimal size of a word and the existence of unaccented words. I will also demonstrate that the apparent diversities of their accentual characters directly follow from a few differences in rules and parameter values for tonal and prosodic structures.

抄録全体を表示

The distribution of sentence accents is shown to be regulated by cognitive-semantic factors. In particular, a model of sentence accent assignment for English is proposed which is comprised of two levels of derivation: one for pitch accent assignment as a semantic-phonological mapping process having access to the specific/nonspecific and action/nonaction distinctions, and the other for nucleus assignment as a genuine phonological process having access to the phonological phrase in the sense of prosodic phonology. The validity of the model is further shown by accentual facts in German, where the mapping process for pitch accent assignment in English is also operative, and in Danish, where the two semantic distinctions play slightly different roles from those which they play in English.

抄録全体を表示

英語母語話者と日本人英語学習者による冠詞の知覚パターンを比較,検討した。先行子音の無音区間の持続時間と冠詞の母音持続時間の二要素が冠詞の聞き取りにどのように関与しているかを無音区間加工聴取実験と冠詞持続時間加工聴取実験により観測した。実験の結果より,日本人と米国人でこれらの音響要因の冠詞知覚に与える影響の違いが明らかにされ,日本人の場合,時間制御のパターンに過度に頼っていることが示された。

抄録全体を表示

D. Jones's English Pronouncing Dictionary has been totally revised for the 15th edition after an interval of twenty years. The writer would like to show some characteristics of the new edition by comparing it with the 14th edition. Of the new characteristics and devices, the most remarkable one is, above all, the introduction of American pronunciation, as well. The type of English described in the 15th edition is not RP but BBC English and 'network English' (American English). It can be said that the 15th edition holds a position as a pronouncing dictionary of world-wide English.

抄録全体を表示

This book is extensively revised edition of the author's 1962 textbook. The significant feature of the new edition can be found in the chapters 9-11, where basic notions and analytical methods of the digital speech processing are explained. Especially, the easy-to-understand explanations of Discrete Fourier Transform and Linear Predictor Coefficients are highly recommendable for all the linguistics and phoneticians who are interested in the computerized speech analysis but lacking the basics of linear mathematics and computer science.

抄録全体を表示