文字と音韻論
平安時代の仮名表記――書き分けない音韻を中心に――
平安時代,平仮名・片仮名による日本語表記が定着した後,そのシステムでは表現できない要素(撥音・促音のような新音韻や,外来語音)を,どのように組み込もうとしたか,特に「書き分けない」という選択をした事例に着目して考察した。
清濁を書き分けないことについては,「にごり」が持っているプロソディ的な性質と結びつける解釈が行われてきた。本稿では,促音・量的撥音便の撥音が零表記であったのも,音長というプロソディの範疇の性質を持つためと考え,さらに,守護国界主陀羅尼経長保頃点の,開拗音・合拗音全般を直音表記する方式も,「拗」の要素をプロソディ的に把握したものであった可能性を指摘した。つまり「書き分けない」要素は,いずれもプロソディに類する性質を帯びたものであって,たとえ外来音・新音韻など,既存の仮名の外側にある要素であっても,プロソディ的な把握に馴染まないものは,表記を工夫して書き分けることを志向すると推定した*。
キーワード:平安時代,仮名,撥音,促音,拗音
日本語における漢字列
中国語をあらわすための文字である漢字は,中国語に対しては「表語文字」として機能している。しかし日本語の文字化に際しては,漢字は「表意文字」として機能することが多い。本稿では,そのことを確認した上で,12世紀半ば頃に成立したと目されている3 巻本『色葉字類抄』,室町期に成ったと目されている『節用集』,江戸時代に出版された『書言字考節用集』を具体的に観察し,同一の漢字列が和語も漢語も文字化しているという状況を整理しながら示した。こうしたことをふまえて,明治期に整版本の草双紙として出版されている『賞集花之庭木戸』と,同一のタイトルで,「ボール表紙本」として活字で印刷されて出版されているテキストとを対照し,漢字列を軸として,連合関係が形成されていることを指摘した。そうであれば,漢字列は非音声的に,連合関係を形成していることになり,そのことは日本語における特徴といってよいと考える。
キーワード:表意文字,漢字列,連合関係
西夏文字における「点」の出現環境と機能
本稿では,西夏文字の「点」はどの位置に,何のために付加されるのかを考察した。本稿の議論に関係する,西夏文字の構造,筆画について術語とともに述べたのち,考察対象となる筆画「点」を定義づけた。筆者は,西夏文字全字形を網羅的に掲載する発音字典『同音』により,点を持つ西夏文字の原文を調査した。結果,西夏文字の「点」について,字形全体ではなく,要素部品の右に付され,文字要素を構成するものと分かった。また点の出現環境を精査すると,点が付される要素部品は,「A 類:くノ型を持つ」と「B 類:匕型を持つ」に二分されることが分かった。A, B 類ともに類似する要素部品(ノメ,ヒ)がある。点の機能は,西夏文字創製時には,類似する部分(ノメとくノ,ヒと匕)の差異を際立たせる「強調符号」だったが,後に点の有無のみが弁別要素と誤認されることになったと筆者は推測する*。
キーワード:文字学,西夏文字,文字要素,筆画,点
ヒッタイト語楔形文字粘土板から導き出される新たな音韻特徴――破擦音の長短の対立――
古代アナトリアにおいて楔形文字粘土板に記録されたヒッタイト語の音韻特徴として,子音の長さの対立は閉鎖音,摩擦音,共鳴音,喉音においてみられるが,破擦音については対立がないと従来考えられてきた。しかしながら,文献学の立場から古期,中期,後期ヒッタイト語という厳密な時代区分を行ったうえで分析すると,古期ヒッタイト語において母音間のシングルの-z-がアナトリア祖語の*d に遡る例があることが分かった。他方,シングルの-z-が*t を反映している例はない。この知見は,アナトリア祖語の*ti は*tsi になる一方,*di は*dzi になるという音変化を示している。そして無声の破擦音はダブルの-zz- で書かれ,有声の破擦音は-z- で書かれる。子音の有声無声という特徴が閉鎖時間の長さと相関することはよく知られている。したがって,古期ヒッタイト語では破擦音にみられる子音の長さの対立がなお存続していたことが分かる†。
キーワード: 文献学,比較言語学,ヒッタイト語楔形文字粘土板,アナトリア諸語,破擦音
論文
英語史における受動虚辞構文の発達について
本論文では,英語における受動分詞を伴うthere 構文,すなわち受動虚辞構文の歴史的発達の全体像を明らかにするとともに,特に関連要素の分布の変化について,VP 内基底語順の変化と機能範疇Pred(ication) の出現に関連付けて説明することを目的とする。その説明に際して,Tanaka and Yokogoshi(2010)による小節の統語分析を拡張することにより,受動虚辞構文におけるbe の小節補部が機能範疇を持たない構造からPred を主要部とする構造へと変化したことを提案する。Pred を主要部とする構造は14 世紀に出現し18 世紀に確立されたが,PredP の出現および確立により,後期中英語期に関連要素が受動分詞に先行する語順が優勢になり,関連要素が受動分詞の直後に来る語順が18 世紀中に消失したという事実が正しく説明されると主張する。
フォーラム
時制交替の一般化再訪――阿久澤・窪田論文に答う――
先行研究において,時制交替の一般化(Tense Alternation Generalization,以下TAG)は日本語における時制節を含んだ繰り上げあるいはコントロールとされる構文を説明するために用いられてきた(Uchibori 2000, Fujii 2006)。Akuzawa and Kubota(2020, 2021),Kubota and Akuzawa(2020)はこの一般化を批判的に検討し,問題があると論じた。本研究は両氏のTAG 批判のうち,(i)TAG に定位した日本語ヨウニナル構文の繰り上げ分析を支持するとされているデータは説得力に欠けているとの批判,および(ii)TAG の接近法が,両氏の提案する意味論的接近法と異なり当該一般化が説明を欠く措定的規定であるという弱点を孕んでいるとの批判に答えることを試みる。まず(i)を解決するために容認度判断実験を行い,被験者の判断がFujii(2006)の報告に沿っていることを示す。またTAG が措定的であるという(ii)の批判については,批判は正しいが,両氏の提案するコントロールの意味論的分析はTAG による接近法が措定的規定に頼って説明した事実を説明しないことを示す。
Graphology, Grammatology, Orthography, and Phonology
Phonemes that were not Written Distinctively in the Heian Period
This study examines how linguistic elements that could not be expressed in hiragana and katakana were written in the Heian period, focusing on cases where people chose not to indicate certain distinctions in writing. Given that sei-daku consonants were not written distinctively, that geminate consonants and a part of moraic nasals were not notated, and that the notation of certain materials did not reflect kaiyō-on (CjV) or gōyō-on (CwV), it is presumed that these sound variants were perceived as prosodically conditioned.
Strings of Sinographs in the Japanese Language
Sinographs (漢字 Ch. hanzi, Jp. kanji) largely function as logograms representing the spoken word when writing the Chinese language. When Sinographs are used to write the Japanese language, however, they tend to function ideographically. That is, historically, Sinographs serve as somewhat abstract markers of meaning whose meaning and/or pronunciation is made explicit through intertextual notation (e.g., furigana or okurigana) that relates the denotation (語義) of a Sinograph to a corresponding Japanese term. After arguing the above point, I analyze three texts, the mid-12th century Iroha jiruishō, the Muromachi Era Setsuyōshū, and the Edo Period Shogen jikō setsuyōshū, and demonstrate how the same string of Sinographs are used to represent both wago, vernacular Japanese lexical items, and kango, words originating from Chinese. Furthermore, I compare the Meiji Period Shōshū hana no niwakido, an illustrated storybook reproduced from an original plate, to a text of the same name printed in movable type and published in the so-called cardboard-cover format (ボール表紙本). By focusing my analysis on strings of Sinographs, I demonstrate the associated relationship that exists between them and Japanese lexical items, and show how such relationships are constructed not only in these two texts but also within the larger ecosystem of written Japanese. This suggests that a string of Sinographs can establish a non-phonetic associative relationship between Sinographs and Japanese lexical items, one characteristic of the Japanese language.
On the Environment in which a “Dot” Appears in Tangut Characters and on its Function
Tokyo University of Foreign Studies)
In this paper, we consider where and for what purpose the “dots” of Tangut characters are added. After describing the structure and strokes of Tangut characters, together with the relevant terminology, we define the Tangut “dots” in question. The author investigates the original texts of Tangut characters with dots by checking the original shapes in the Tangut dictionary “Tongyin,” which exhaustively lists all the Tangut characters. As a result, it was found that the “dots” of the Tangut characters were attached to the right of the element and constituted the “part,” not the whole character. When the environment of the appearance of dots was scrutinized, it was found that the component parts to which dots were attached could be divided into two parts: Class A: a “くノ”-type and Class B: a “匕”-type. Both classes A and B have similar component parts (ノメ,ヒ). At the time of the creation of the Tangut characters, the dots functioned as an emphatic sign that distinguished the differences between similar parts (ノメvs くノ,ヒvs 匕), but the author surmises that later the mere presence or absence of dots was misidentified as a distinguishing element.
A New Phonological Feature Obtained from Hittite Cuneiform Clay Tablets: The Linguistic Contrast between Single and Double Affricates
There is no positive evidence that Hittite scribes employed the signs including voiced and voiceless consonants in a contrastive manner to distinguish between voiced and voiceless values. What the Hittite scribes tried to show by orthography was a linguistic contrast between single and double consonants in intervocalic position. Single and double consonants indicate lenited and unlenited (or short and long) qualities, respectively. The Proto-Indo- European contrast of voiced/voiceless consonants, inherited also in Proto-Anatolian, was reinterpreted as one of lenited/unlenited (or short/long) in Hittite. This contrast is observed in stops, fricatives, laryngeals and sonorants. It is, however, traditionally assumed that the same contrast is not observed in affricates. We discuss this problem in this paper by analyzing the distribution of the 3 sg. present active endings, -Vzzi and -Vzi.
There are a small number of verbs which show the single -z- intervocalically in Old Hittite, as represented by ú-e-mi-zi (< *au-ém-̯e-di < *au-ém-̯e-ti) ‘finds’ and arnuzi (< *h1 ̥r-né-u-di < *h1 ̥r-né-u-ti) ‘(re)moves’. The relevant verbs all go back to the Proto-Anatolian forms characterized by the ending with voiced *d created by lenition rules. An obvious inference to be drawn from this fact is that the single -z- in these verbs reflects the lenited quality produced from *d by affrication before *i in its prehistory. Although the lenited *-dzi was replaced by the corresponding unlenited *-tsi (< *ti) to a large extent, the Old Hittite verbs in -z- preserve a notably archaic feature speaking for the linguistic contrast between the lenited and unlenited affricates. This claim is supported by some further pieces of evidence.
In the case of the historical study of dead languages, written documents are virtually the only sources of linguistic information. Furthermore, written documents are recorded by means of letters, which are nothing more than tools for indirectly representing phonetic information. In spite of these inherent difficulties, the techniques of historical linguistics enable us to obtain evidence useful for inferring language change. In this paper we have discussed a case in which markedly improved philological analyses contribute to inferring linguistic changes that occurred in the internal history of the language as well as in its prehistory.
Articles
On the Development of Passive Expletive Constructions in the History of English
This paper aims to account for the development of passive expletive constructions (PECs) in the history of English, especially the changing distribution of their associates, in terms of the rise of a functional category Pred(ication), as well as the change of the underlying word order within VP. By extending Tanaka and Yokogoshi’s (2010) analysis of small clauses, it is argued that there was a structural change of the small clause complement to be in PECs: from the structure lacking a functional category to the structure headed by Pred, which appeared in the fourteenth century and was established in the eighteenth century. The rise and establishment of the PredP structure are shown to be responsible for the fact that the order in which the associate precedes the passive participle became predominant after Late Middle English, finally replacing in the eighteenth century the order in which the associate immediately follows the passive participle, because the latter order can be derived only from the structure without Pred.*
Key words: associate, passive expletive construction (PEC), Pred(ication), small clause, underlying word order
Forum
Tense Alternation Generalization Revisited: A Reply to Akuzawa and Kubota
Tense Alternation Generalization (TAG) has been used in previous studies to account for some distinctive features of the constructions that are conceived as raising and control across tensed CP in Japanese (Uchibori 2000, Fujii 2006). Akuzawa and Kubota (2021), as well as Akuzawa and Kubota (2020) and Kubota and Akuzawa (2020), critically assessed the generalization, arguing that it is problematic. The current study attempts to respond to two of their criticisms: (i) the data used in Fujii (2006) to support the TAG-based argument for the raising analysis of the yooni naru construction are uncompelling; and (ii) the TAG approach suffers from the generalization being unexplained and stipulated, unlike the semantic approach that the two-author team proposes. To settle the first issue, we conducted an acceptability judgment experiment, whose results revealed that our participants’ judgments were in line with those reported in Fujii. As for the second criticism, the current study holds that while the criticism is right in that TAG is stipulated, their semantic analysis of control does not achieve what the TAG approach achieves by relying on the stipulation.*
Key words: control, raising, nominative-genitive conversion, tense, yooni naru