東京アクセントの特徴再考 : 語頭の上昇の扱いについて  [in Japanese] A New Analysis of Accent in Tokyo Japanese  [in Japanese]

東京方言において, ソバヤ(そば屋)やサ<カナ>^^^-(魚)などの単語単独発話に現れる第1モーラから第2モーラへの上昇については, ア<ノソバ>^^^-ヤ, ウマ^^-イソバヤ__-, コ<ノアカイサカナ>^^^-, ニガ^^-シタサカナのように文中ではなくなるという観察から, 上昇はアクセントの特徴ではなく, 「句」の特徴であるという上野善道氏, 川上蓁氏らの考え方が現在有力である。本稿では話者10名の読み上げ資料の音響分析により, ニガシタサカナに類する「南禅寺のみやげ」の「みやげ」などには上昇作用が安定的に存在していることを示し, これは上昇が「みやげ」の韻律的特徴としてもともと備わっているためと考えるべきであること, すなわち上昇はアクセント単位が持つ特徴の一部, すなわち本稿で言う「アクセント」の一部であり, 「句」そのものの特徴ではないと考える方が音声的実態に即していることを示した。ただし, この上昇はアクセント(型)の区別には役立たず, 環境によっては顕在化しないと考えるべきである。その環境とは, 前文節から意味的な限定を受ける場合, しかも前文節が無核か, 核位置が文節末に近い場合, あるいは前文節以前にフォーカスがある場合である。

In Tokyo Japanese, a word has an initial audible pitch rise when uttered in isolation. This rise, though, often seems to disappear when the word is uttered in the medial or final position of a sentence. In the present study, in order to determine whether or not the rise really disappears, a careful examination of pitch movement was made in sentences in which an initial noun with an accent kernel ('), followed by the genitive case marker no, syntactically modifies the next noun without an accent kernel as in Na'nzenzi-no (place name [gen.]) miyage-o (souvenir [acc.]) minna'-ni kuba'tta (gave to everyone), where the initial rise in miyage is expected to be lost. The examination revealed that, in all of the ten speakers examined, the initial rise in miyage-even though it was, on average, as small as 1.7 semitones-was always present in Na'nzenzi-no miyage-o. Based on this fact, the present paper proposes a new analysis whereby the rise is an intrinsic part of the word, contrary to the prevailing view that it signals the beginning of a prosodic (accentual or intonational) phrase. In a sentence which begins with Na'ra-no (place name [gen.]) miyage-o, however, some speakers uttered miyage without rise. They are the same speakers who showed a smaller rise than the other speakers in Na' nzenzi-no miyage-o. The disappearance of the rise in Na'ra-no miyage-o can be accounted for by the too short stretch from the accent kernel of the preceding word to the beginning of the following word. There is not enough time for the kernel, which triggers a sharp pitch fall, to fully realize the expected amount of the fall, and the remaining amount of the fall counterbalances or masks the initial rise of the following word because the amount of rise in these speakers is idiosyncratically small. A theoretical consequence of this new analysis is that the word's initial rise is an inseparable counterpart of the sharp pitch fall at the accent kernel of the word. Therefore, the use of the term "accent, " if it refers only to the sharp fall, is not appropriate for the phonological description of this variety of Japanese.