多くの言語において，文否定要素が引き起こす干渉効果が観察される。本稿は，英語における数量詞の相対的作用域解釈，フランス語の単一wh疑問文における元位置のwhの作用域解釈，更にはドイツ語の作用域標示構文におけるwhの作用域解釈等について考察し，Chomsky（2000）以降の枠組みを採用することにより，これらの現象が統語的に説明されることを論ずる。干渉効果を捉えるため，本論考は，文否定辞を主要部とする否定辞句がフェイズの一つと認定されることを提案する。不可視移動をスペル・アウト後の演算と認めた上で，数量詞繰り上げ及び数量詞繰り下げも他の演算と同様，フェイズ不可侵性条件に従うものとする。これにより，数量詞の相対的作用域解釈における干渉効果が導き出される。また，whはCを占める作用域標識と局所的に一致が必要とすることにより，フランス語の単一wh疑問文やドイツ語の作用域標示構文におけるwhの作用域が説明される。
本論文はアイヌ語の南北海道方言における引用という文法範疇を統語的，談話機能的，歴史的，類型論的視点から体系的に記述する。先行研究においては発話の報告は単に「引用」として，あるいは直接・間接話法の対立として分析されてきた。本論文では引用の位置付けを決定する中核的原理としての人称直示形式によって区別されるアイヌ語の引用構文には3つの主要なタイプがあることを示す。
（1）直接話法：元の発話者の視点；（2）間接話法：引用者の視点；（3）準直接話法：元の発話者と引用者の視点の組み合わせ。
準直接話法においては，二・三人称の元発話者を指示する際にいわゆる「不定人称」が使われる点で引用者の視点への代名詞の調整がおこっているにもかかわらず他の指示対象は元の発話で使用されたはずの形式がそのまま現われる。このような「不定人称」の用法（田村2000: 74の「引用の一人称」）はロゴフォリックなもの（つまり，ある種の引用構文において引用者を指示する特別の形式）であると提案する。これはHagege（1974: 287）がアフリカの諸言語の例によってはじめて示したロゴフォリック性の定義に適合する。
重要なのは，準直接話法はさまざまな点で直接話法の一種と分析できないということであり，上の3つの構文は文体，あるいは談話構成上の都合で選ばれていると考えられる。ただし，民話のジャンルによっては主要なスタイルとして慣用化されているものもある。
本研究では，日本語の使役形に現れる言語変異である「さ入れ言葉」（例，やらさせる，入らさせる）に関して，数量的分析を試みる。さ入れ言葉については，これまで言語内的要因に関しては詳細な分析が行われているが，言語外的要因との関係についてはあまり分析がない。しかしながら，言語変異において性差，スタイル差などの言語外的要因がその分布に大きな影響を与えていることが様々な研究において観察されている。さ入れ言葉の性質の解明のためには，言語外的要因に関して，更に各要因間の関係に関しても分析を行う必要がある。そこで，本研究では『日本語話し言葉コーパス』を用い，言語外的要因がさ入れ言葉の分布に与える影響，各言語内的・言語外的要因間の関係を明らかにすることを目的とする。分析の結果，各要因のさ入れ言葉の分布への影響，それらの影響の大きさ，各要因間の関係が新たに明らかになった。また，変異理論全体への示唆が得られた。
ドイツ語の[f]または[x]を含む無意味語を用いて日本語母語話者を対象に知覚実験を行い，摩擦の持続時間を測定した。[f]と[x]の間に持続時間の差が見られなかったが，日本語に実在する借用語に見られるものと同様に，[x]をもつ語の方が[f]をもつ語より促音挿入が起こりやすいことがわかった。さらにCVC，CCVC，CVCən（Cは無声摩擦音のものである）の語型と促音化率の相関関係について実在語に見られる傾向と比較して論じる。
本稿はワロゴ語（オーストラリア）の母音間の子音連続のデータを提示し，それをもとにして，音素の聞こえ度階層を提案する。この階層は，従来提案されている聞こえ度階層と異なる点がある。半母音と流音の相対的な位置などである。更に鼻音の階層と閉鎖音の階層も考察する。
Intervention effects induced by sentence negation can be found in many languages. In this paper, we discuss the scope interpretation of quantifiers in English, and in-situ wh-phrases in French single questions, as well as those in German scope-marking constructions. Our claim is that intervention effects should be explained syntactically. In order to capture the relevant phenomena, the framework of Chomsky (2000) and subsequent work is adopted. This paper proposes that NegP, headed by sentence negation, should be identified as a phase. It is further assumed that quantifier raising (QR) and quantifier lowering (QL), whether applied covertly or overtly, are subject to the Phase Impenetrability Condition just like other operations. Under these assumptions, intervention effects in quantifier scope phenomena can be obtained. We also argue that local agreement must hold between a wh-phrase and a scope marker in C in French single in-situ questions, as well as German scope-marking constructions. The locally limited scope of wh-phrases in these constructions follows from phase theory.*
This paper attempts to provide a systematic description of reported discourse (RD) categories in Southern Hokkaido dialects of Ainu from syntactic, discourse-functional, historical and typological perspectives. While RD has been analysed in the previous literature either as “quotations” or as direct vs. indirect discourse, I argue, based on a continuum approach, that Ainu has three main RD-constructions distinguished by person deixis as a core principle in determining the orientation of RD: (i) direct discourse (DRD) with the author’s perspective, (ii) indirect discourse (IRD) with the reporter’s perspective, and (iii) semi-direct discourse (SRD) with the combined perspective of the author and reporter. In the case of SRD, use of the so-called “indefinite person” in the case of coreference of the third/second person author with a participant in the quote indicates a consistent reporter-oriented shift in pronominal reference, while the other references occur as they would have been used in the original discourse. I propose to identify Tamura’s (2000: 74) ‘first person in quotes’ usage of the indefinite person as logophoric in the sense of Hagège (1974). In this analysis SRD, DRD, and IRD are regarded as stylistic and discourse-organizing options, though some of them have been conventionalized as primary styles in certain folklore genres.*
This research statistically investigates the recent synchronic variation in Japanese causative forms involving a process called Sa-insertion, which adds an extra -as- to the causative morpheme, as in yar-asase-ru vis-à-vis the standard yar-ase-ru ‘let someone do’ and hair-asase-ru vis-à-vis the standard hair-ase-ru ‘let someone enter’. Although previous studies have elucidated the language-internal properties of this phenomenon especially with respect to the length of stem verbs and the types of the following elements (Sano 2008, among others), language-external factors such as gender and style, as well as the interaction of those language-internal factors, remain to be explored.
The present paper aims to verify the claims of the previous studies and reveal the language-external factors as well as the effects of the interaction of major factors governing the distribution of Sa-insertion. For this purpose, Corpus of Spontaneous Japanese (CSJ) is employed which has rich annotations concerning language-external factors. An exhaustive examination of CSJ brought forth a total of 42 causative forms with Sa-insertion, as opposed to a total of 1,498 standard causative forms; thus, the rate of Sa-insertion amounts to 2.73%. The data are then subjected to a factor by factor analysis, followed by a multivariate analysis.
The result shows that the distribution of Sa-insertion is significantly affected by language-external factors in addition to language-internal factors, and that the factors that govern the distribution interact with each other in an intricate way.
Nonsense words involving German voiceless fricatives [f] or [x] (or [ç]) were used to test Japanese subjects’ perception of gemination, and the friction durations were measured. Results showed that, similar to existing loanwords in Japanese, subjects were overall more likely to perceive gemination for stimuli involving [x] (or [ç]) than for those involving [f], despite the fact that no significant difference between the durations of these voiceless fricatives was found. In addition, the relationship between fricative environment (i.e. CVC, CCVC, CVCən, where C is the voiceless fricative) and rate of gemination was examined. Similar to existing loans, results indicated that fricatives were more likely in the environment CCVC than CVC to be adapted into Japanese as geminates. Opposite of the existing loanword trend, however, fricatives were less likely to be geminated if followed by a syllabic nasal (i.e. CVC vs. CVCən). Comparing friction duration across the different word environments, [f] and [x] (or [ç]) were found to have a shorter duration word-medially (CVCən) than word-finally (CVC, CCVC). Thus, a shorter difference in frication duration, controlled by word environment, was correlated with lower perception of gemination in the test words. Concerning the quality of the preceding short vowel (e.g. [lax] vs.[lox]), no influence on fricative geminate perception was observed.*
This paper provides data on the intervocalic consonant clusters of Warrongo (Australia), and offers the following observations: (a) The liquids /|, l/ should be considered more sonorous than the semivowel /w/; (b) The liquid /|/ should be considered more sonorous than the semivowel /j/; (c) Among the nasals, the alveolar /n/ should be considered more sonorous than the bilabial /m/ and the velar /N/; (d) Among the nasals, the palatal /−/ should be considered more sonorous than the velar /N/; (e) Among the stops, the alveolar /d/ should be considered more sonorous than the velar /g/; and (f) Among the nasals and stops, the alveolars may possibly be the most sonorous, and the velars possibly the least sonorous. In terms of (a) and (b), Warrongo diverges from most of the languages discussed in the literature, where semivowels are considered more sonorous than liquids.*