２日間にわたり、多数の皆様のご参加、誠にありがとうございました。

ドイツ・マックスプランク進化人類学研究所(Max Planck Institute for Evolutionary Anthropology; MPI-EVA)との研究協力により，日本語における動詞の自他交替と項交替に関する 国際シンポジウムを下記のように開催します。英語タイトルの発表は英語で，日本語タイトルの発表は日本語で行われますが，できるだけ多くの発表に日英語を併記したスライドを準備したいと考えています。
皆様のご来場をお待ちしています。
（近隣には食堂・レストランがありませんので，昼食は各自でご準備ください。）

• 開催日時：2012年8月4日(土)～5日（日）
  
• 開催場所：国立国語研究所 2階　講堂
  National Institute for Japanese Language and Linguistics, Tokyo, 2nd Floor, Lecture Hall
  　http://www.ninjal.ac.jp/
  

本シンポジウムは，Max Planck研究所が世界諸言語の交替現象を言語類型論的に俯瞰する”Valency Classes in the World’s Languages”という研究プロジェクトを進めているのと並行して，日本語に特化した項交替（動詞だけに限らない）の普遍性と個別性を，形態論，統語論，意味論，歴史変化，方言，習得などの観点から包括的に考察するものです。

プログラム

• 8月4日（土） Saturday, 4 August
• 受付開始 9:30-

10 : 15 ～

opening
Taro Kageyema




司会：影山太郎（NINJAL)

10 : 20 ～ 11 : 00　

“Markedness Effects in Valency Changing Operations”
Andrej Malchukov (University of Mainz / MPI-EVA)

• abstract(要旨)
The present talk, which is an outgrowth of the Leipzig Valency Classes Project, deals with the phenomenon of “voice ambivalence”, that is with the fact that some voice/valency markers perform different functions when applied to different classes of verbs (e.g., transitives vs. intransitives). While such polysemies have been noted in typological studies of individual voice categories (cf. Shibatani 1985 on passives), the general picture of this domain is still lacking. Drawing on the contributions to Comrie & Malchukov, forthcoming, the talk will provide a cross-linguistic overview of ambivalent voice categories:

- which are both valency-decreasing and valency-preserving (cf. the potential forms in Japanese)
- which are both valency-increasing and valency-preserving (for example, in Arabic, the same stem alternation produces either causative or intensive function)
- which are valency decreasing, but yield different effects when applied to different valency classes (passive/anticausative vs. antipassive polysemy)
- which are both valency increasing and valency decreasing (cf. causative-passive and applicative-antipassive polysemies)
- which are valency increasing, but yield different effects when applied to different valency classes (causative-applicative polysemy)

The explanation of the voice polysemies may partly reside in the shared syntactic features, which can be summarized in a semantic map. A more general approach would rely on the notion of local markedness or, more broadly, functional usefulness: when a certain category is extended beyond its natural domain of application, such infelicitous combination (e.g., passives of intransitives) would be either blocked or reinterpreted (cf. Malchukov 2011 on resolution of infelicitous grammeme combinations in the domain of verbal categories).

References Comrie, Bernard & Malchukov, Andrej (eds.). Valency classes: a Comparative Handbook. Forthcoming in Mouton de Gruyter. Malchukov, Andrej. 2011. Interaction of verbal categories: resolution of infelicitous grammeme combinations. Linguistics, 49–1 (2011), 229–282 Shibatani, Masayoshi. 1985. Passives and related constructions. Language 61. 4: 821-848.



司会：Wesley Jacobsen（Harvard University)

11 : 00 ～ 11 : 40

“A Typological Study of Verbal Semantic Constraints on Syntactic Alternations”
Soren Wichmann (MPI-EVA)

• abstract(要旨)
The database of the Leipzig Valency Classes Project currently contains sufficient data to study how 143 different syntactic constructions apply to a uniform set of 87 verbal concepts across 12 languages (Ainu, Arabic, Bezhta, Bora, Chintang, Hoocąk, Icelandic, Italian, Jakarta Indonesian, Mandinka, N|u, Zenzontepec Chatino). This database—possibly supplemented by additional data—will be used to illustrate difficulties in findings broad patterns when constructions are classified according to the types of verb that apply to them. On the other hand, meaningful semantic patterns are directly observed in the opposite situation, where verb types are classified according to the constructions that they appear in. The procedure confirms the assumption of Levin (1993) that verbal semantic classes underlie syntactic alternations, but takes this claim to a cross-linguistic level. When a strict procedure is applied to derive Levin’s verb classes from data on syntactic alternations in English, one finds that her verb classes do not follow in a systematic way from the alternations that she studied (they form two unconnected chapters in her book). Making a claim that highly specific semantic classes underlie alternation types is probably only possible using cross-linguistic data.
11 : 40 ～ 12 : 20

“Rigidity versus Laxity in Valency Retention: A Contrastive Study of Complex Predicates in Japanese and Indo-Aryan”
Peter Hook (University of Michigan) and Prashant Pardeshi (NINJAL)

12 : 20 ～13 : 20

昼食　Lunch


13 : 20 ～ 14 : 20

ポスター発表（POSTER PRESENTATION)

非対格自動詞のV-(s)ase 使役他動詞化の可否と出来事開始時点の可視性」
板東美智子（滋賀大学)

「日本語における述部の意味構造と使役起動交替」
日高俊夫（神戸松蔭女子学院大学，研修員）

「日本語学習者コーパスに基づくヴォイスに関する誤用の類型―中国語母語話者と英語母語話者の場合―」
望月圭子（東京外国語大学）

「事象構造による複合動詞の自他交替の分析」
史 曼（東北大学，大学院生）

“A Uniform Treatment of Nonstative Spontaneous Constructions in Japanese”
Takayuki Ikezawa (Waseda University, graduate student)

“Transitive and Intransitive Constructions in Japanese and English”
Zoe Pei-sui Luk (University of Pittsburg/The Hong Kong Institute of Education)

“Representing the Continuum between Arguments and Adjuncts within Predicate-Frames”
Pierre Marchal (Institut National des Langues et Civilisations Orientales / Waseda University, Kitakyushu Campus, graduate student)　




司会：竹沢幸一（筑波大学)

14 : 30 ～ 15 : 10

“Morphological, Phonological, and Semantic Subregularities in Causative-Inchoative Verb Pairs in Japanese: A Schema-based Account”
Yo Matsumoto (Kobe University)

15 : 10 ～ 15 : 50

“Anticausativization in the Northern Dialects of Japanese”
Kan Sasaki (Sapporo Gakuin University)

• abstract(要旨)
One of the most prominent features of the northern Japanese dialects, spoken in Tohoku and Hokkaido areas, is the productive anticausative morphology. The northern Japanese dialects employ the spontaneous suffix /-rasar/ (or /-rar/) as a morphological marker for anticausativization. The range of anticausativization in the northen dialects is wider compared not only to that in the dialects spoken in the other areas, where anticausativization is lexical, but also to that in the languages where the reflexive morpheme is employed as an anticausative marker. This presentation gives a brief sketch of anticausativization in the northern dialects and considers the crosslinguistic implications of the data from the northern dialects.
15 : 50 ～16 : 10

休憩　BREAK





司会：Prashant Pardeshi（NINJAL)

16 : 10 ～ 16 : 50

“Japanese Children's Use of Morphosyntax and Argument Structure to Infer Meaning of Novel Transitive and Intransitive Verbs”
Ayumi Matsuo (Kobe College; presenter), Letitia Naigles, Gary C. Wood, and Sotaro Kita (University of Birmingham)

• abstract(要旨)
Previous research has found that children who are acquiring argument- drop languages such as Turkish and Chinese make use of syntactic frames to extend familiar verb meanings (Go ̈ ksun, Ku ̈ ntay & Naigles, 2008 ; Lee & Naigles, 2008). This article investigates whether two-year-olds learn- ing Japanese, another argument-drop language, make use of argument number and case markings in learning novel verbs. Children watched videos of novel causative and non-causative actions via Intermodal Preferential Looking. The novel verbs were presented in transitive or intransitive frames; the NPs in the transitive frames appeared ‘bare’ or with case markers. Consistent with previous findings of Morphosyntactic Bootstrapping, children who heard the novel verbs in the transitive frame with case markers reliably assigned those verbs to the novel causative actions.
16 : 50 ～ 17 : 30

“Transitivity Alternation in Second Language Acquisition”
Yasuhiro Shirai (University of Pittsburg/NINJAL) and Zoe Luk (University of Pittsburg/The Hong Kong Institute of Education)

17 : 30 ～ 18 : 10

“Children's ‘Erroneous’ Intransitives, Transitives, and Causatives and the Implications for Syntactic Theory”
Keiko Murasugi (Nanzan University)

• abstract(要旨)
It is well-known that children erroneously produce intransitive form of (di)transitive forms and intransitive/(di)transitiveforms instead of causative forms (e.g., Daddy, I will feel you better (meaning ‘Daddy, I will make you feel better’; Fuusen fukurande (meaning ‘Please blow up the balloon for me.’). This type of “error” by children is presumably not a response to the children’s environment, and it probably cannot be explained by the limited ability of the children’s processing either. Based on the analysis of Japanese-speaking children's common errors widely observed in the previous literature, our longitudinal studies with two Japanese-speaking children, and the corpus (The Child Language Data Exchange System (CHILDES)), in this paper, we present the v-VP frame analysis for the acquisition of Japanese verbs and complex predicates. We provide a uniform account of such verbal errors following Larson’s (1988) v-VP frame of VP-Shell hypothesis: (i) the predicate-argument structures of large Vs and small v’s are acquired very early, (ii) children making such errors given above assume [±cause] v to be phonetically null, just like pass-type verbs in English, at one stage of language acquisition, and (iii) what requires time is the acquisition of the lexical form of each V and the forms in which [±cause] small v’s are realized. We suggest in this talk that the innate linguistic principles define a space of possible human language and the possible linguistic variation, and children’s linguistic errors arise because children, at an intermediate stage of language acquisition, try out the possible features in adult languages elsewhere in the world. Children erroneously producing intransitive forms instead of (di)transitive forms, and intransitive/ (di)transitive forms instead of causative forms are trying out the languages that are not their mother tongue, yet never trying out languages that violate the principles of possible grammar. Children are producing the pass-type verbs at the intermediate stage of language acquisition.
18 : 15 ～ 20 : 00

懇親会　RECEPTION
２階　講堂前スペース

• 8月5日（日） Sunday, 5 August

司会：岸本秀樹（神戸大学)

10 : 00 ～ 10 : 40

「項交替に関する理論的諸問題」
柴谷方良 Masayoshi Shibatani (ライス大学 / 国語研客員)

10 : 40 ～ 11 : 20

「平安時代日本語の感情形容詞と感情動詞―『源氏物語』『今昔物語集』のコーパス分析を通して―」
田中牧郎 Makiro Tanaka（国立国語研究所）・山元啓史 Hilofumi Yamamoto（東京工業大学）

• abstract(要旨)
現代日本語における感情形容詞や一部の感情動詞について、文のタイプによって感情主が一人称に制限されることをめぐって、多くの研究がある。一方、こうした日本語の感情形容詞や感情動詞の人称制限の歴史については、あまり研究が行われていない。本発表では、国立国語研究所で構築中の『源氏物語』（11世紀初め）と『今昔物語集』（12世紀初め）のコーパスを用いて、感情形容詞や感情動詞が取る項の分析を行い、次の三つのことを論じる。 （１）平安時代語においては、感情形容詞の感情主は一人称に制限される。 （２）平安時代語においては、感情動詞の感情主は二人称・三人称に制限される。 （３）平安時代語における感情主の人称による形容詞と動詞の使い分けのありようは、純粋の和文である『源氏物語』と、和漢混淆文である『今昔物語集』とで、異なる面がある。
11 : 20 ～ 12 : 00

“The Historical Source of the Bigrade Transitivity Alternations in Japanes”
Bjarke Frellesvig (University of Oxford / NINJAL)　and John Whitman (NINJAL)

• abstract(要旨)
A well known feature of the Japanese verbal lexicon is the existence of transitivity alternations associated with differences in stem shape. These alternations were already nonproductive in Old Japanese (8th century). Alternations between vowel (modern monograde, OJ bigrade) and consonant (OJ quadrigrade) stems are of particular interest, because the valency of the stem is not predictable by its shape. Transitive vowel stem verbs with such as tate- ‘stand it up’ may be paired with a consonant stem tat- ‘stand’, but at the same time intransitive vowel stems such as sake- ‘split’ are paired with a transitive consonant stem sak- ‘split it’. In each case the vowel stem appears derived, but which stem is transitive and which is intransitive is unpredictable. As Murasugi (this conference) shows, these alternations are mastered only at a later stage of first language acquisition, but they are maintained with minor variation across Japanese languages and dialects. In this paper we reconstruct a diachronic source for the stem shape-based transitivity alternation in Japanese. Our account is based on an insight that dates back at least 60 years among Japanese historical linguists, but it hinges on several features of reconstructed post-proto Japanese that are not universally accepted. The most important of these is that an older layer of verbal derivation in Japanese involved V-V patterns where V2 was attached directly to the basic stem (root) of V1, unlike later stages of the language where verbal derivation requires a derived stem (primarily the infinitive) of V1. We argue that the consonant-stem--vowel stem alternation derives from combination of the verb *e- ‘get’ with the underived basic stem (root) of V1. This combination results in the subregularities that survive to this day: transitive accomplishment verbs from the combination of ‘get’ with intransitive achievements, and anticausative intransitives from the combination of ‘get’ with semantically causative transitives. We also discuss the relationship of the ‘get’ derivation with other Vstem1+V2 combinations at the proto-Japanese level. Combinations of Vstem1 + ar- (probably cognate with the verb ar- ‘exist’) derive stative intransitives. Combinations of Vstem1 + *s- (possibly cognate with *sE- ‘do’) derives causative transitives. Only *s- is used to derive transitives from stative roots. *e- ‘get’ primarily derives accomplishments from achievements.
12 : 00 ～13 : 00

昼食　Lunch





司会：竝木崇康（茨城大学)

13 : 00 ～ 13 : 40

「日本語における自他交替の意味論的根拠に関する再考察」
ウェスリー・ヤコブセン Wesley Jacobsen (ハーバード大学)

• abstract(要旨)
The hypothesis advanced in Jacobsen (1985, 1992) correlating formal markedness patterns in Japanese transitive-intransitive verb pairs with semantic markedness has been shown to account for transitive-intransitive alternation patterns across a wide range of languages (Nedjalkov (1969), Haspelmath (1993)). Specifically, pairs where the intransitive form is less marked than the transitive form tend to correlate with events normally conceived of as occurring without the influence of any outside force (e.g., wak-as-u “boiltr” vs. wak-u “boilin”) whereas pairs where the transitive form is less marked tend to correlate with events normally conceived of as occurring under the influence of an outside force (e.g., kudak-u “smash” vs. kudak-e-ru “become smashed”). These markedness phenomena call into question the uniquely privileged place given in the study of transitivity to the notion of a transitive prototype, championed in its most famous form by Hopper and Thompson (1980) and representable by the semantic structure in (1) (following Dowty (1979), Kageyama (1996), and others). (1) [DO(x)]CAUSE[BECOME(STATE(y))] Soozinin ga monooki no doa o aketa.. “The janitor opened the door of the shed.” Studies such as Levin and Rappaport Hovav (1995) have viewed this semantic structure as underlying both the transitive and intransitive members of transitive alternations, so that the intransitive correlate to (1) would be as in (2), preserving the semantic structure of (1) but leaving unfilled the x slot. (2) [DO(ø)]CAUSE[BECOME(STATE(y))] Monooki no doa ga aita. “The door of the shed opened.” This presumes that transitive meaning is fundamental, and intransitive meaning derived, in such alternations. A consideration of morphological markedness patterns, however, suggests that for some (indeed, in Japanese, most) transitive alternations involving directionality, the opposite is the case. In Japanese, a full 28.2% of transitive-intransitive pairs have a less-marked intransitive member, as opposed to only 13.5% having a less-marked transitive member. This suggests the existence of an intransitive prototype that plays a role at least as fundamental, and perhaps more so, as the transitive prototype in transitive alternations. Three possibilities suggest themselves for the semantic structure of that prototype: the earlier (2) as well as (3) and (4). (3) [BECOME(STATE(y))] (spontaneous meaning) (4) [DO(y)]CAUSE[BECOME(STATE(y))] (reflexive meaning) (2) is exemplified in a much more limited range of intransitives than predicted by Levin and Rappaport, typically in intransitives with the passive-like affix –aru as in the pair kimeru “decide” vs. kimaru “be decided/settled.” The transitive-like reflexive meaning in (4) is restricted to the intentional interpretations of intransitives, less common in alternating intransitives (e.g., okiru “get up,” kakureru “hide,” etc.) and more common in the form seen in (4a) typifying activity verbs that do not alternate with transitives (e.g., aruku “walk,” oyogu “swim,” odoru “dance,” etc.). (4a) [DO(y)]CAUSE[MOVE(y)] This paper will argue that the fundamental intransitive prototype is in fact to be identified with (3), typified by a change in state, or coming into existence, of an entity apart from the influence of any force, internal or external. The prototypical status of the semantic structure in (2) is supported, first, by its relative simplicity with respect to the structures in (3) and (4), as well by the maximal difference it exhibits from the transitive prototype. Furthermore, as argued in Næss (2007), the intransitive prototype can be seen to be less marked than the transitive prototype on both formal and semantic grounds. This suggests that the intransitive prototype is in fact more cognitively “basic” than the transitive prototype, a plausible conclusion when (3) is viewed as the semantic expression of a primal impinging on the senses of an entity or some change in the entity. Such an impinging on the senses is an epistemological form of coming into existence in the field of perception, a cognitive experience that can be seen to be more direct than, and a prerequisite to, the positing of a causal connection between force and recipient underlying the causal chain embodied in the transitive prototype (Croft (1991).
13 : 40 ～ 14 : 20

「自他交替の『ゆれ』について」
辻村成津子 Natsuko Tsujimura（インディアナ大学）

• abstract(要旨)
日本語の「てある」構文は、従来自動詞化のプロセスと考えられ、「木が倒してある」や「火が消してある」などのように他動詞の目的語を主語とし、その他動詞に「ある」をつけることになっている。この構文では他動詞が不可欠な構成要素と見なされているため、「木が倒れてある」や「火が消えてある」のように自動詞を含んだ場合、非文とされる。自動詞を使ってほぼ同義の文を作るとしたら、「木が倒れている」や「火が消えている」のように、「ある」のかわりに「いる」を使わなければならない。ところが「てある」構文に自動詞が使われている例文は、インターネットをはじめ文学作品にも見られるため、「てある」構文において他動詞と自動詞の間に社会言語学的「ゆれ」が存在する可能性が考えられる。一方、この自・他の「ゆれ」はどんな自動詞でも許されるというわけではなく、ある規則性がみられる。そこで本論では、「てある」構文において、どういう要因が自・他の「ゆれ」を導き出すのかを焦点に考察していきたい。先行研究では、話者の年齢差と地域差が重要な要因だという結果が出ているが、本論ではさらに「てある」構文に使われる自動詞の意味的要素に規則性があることを述べる。
14 : 20 ～ 15 : 00

「項交替と主格制約」
岸本秀樹 Hideki Kishimoto（神戸大学）

• abstract(要旨)
いくつかの日本語の項交替現象について，節に必ず一つ主格でマークされる項がないといけないという，いわゆる主格制約に制限を受ける項交替と，一見その制約を受けない項交替がある。本発表では，この違いは，時制要素の性質に還元でき，日本語においては，時制要素が項の格を認可する場合には主格制約が課されるが，時制要素の項の認可が必要とならない環境においては，主格制約がかからない項交替が可能であることを論じる。一見，主格交替がかからないと考えられる交替においても，時制要素が格認可をする環境においては，主格制約が項交替にかかることも示す。
15 : 00 ～15 : 20

休憩　BREAK





司会：青木博史（九州大学/NINJAL)

15 : 20 ～ 15 : 55

「漢語動詞の項のあらわれ方」
小林英樹 Hideki Kobayashi（群馬大学）

• abstract(要旨)
本発表は、まず、漢語動詞の自他（「ドアが開閉した」「太郎がドアを開閉した」）に触れ、続いて、項のあらわれ方を変える複合（「所長が発電所を停止した」「＊所長が発電所を自動停止した」）を考察する。最後に、漢語動詞をいくつか取りあげながら、漢語動詞の詳細な記述の必要性について述べる。
15 : 55 ～ 16 : 30

「動詞複合による項の創出と主題役割の変更」
由本陽子 Yoko Yumoto（大阪大学）

• abstract(要旨)
英語ではLevin (1993)があげているような多様な構文交替 が動詞の形態を変えることなく起こるが、多くの先行研究が主張するように、これらの交替は動詞の意味構造の変更に伴った項の具現形式の変化として捉えられる。 系譜的には英語に近いドイツ語、ロマンス系の言語を始め多くの言語においては、これに対応する動詞の意味の変更とそれに伴う項構造の変更が接辞を付加することによって行われている。 本発表では、日本語においては「動詞(V1)＋動詞(V2)」型の複合、 すなわち右側にV2を結合する語形成がこれに対応する操作として捉えられること、 また、それによる新たな項の創出や主題役割の変更は、同じV2であってもV1との意味的関係によって異なることを示す。特に、V2が「込む」「上がる」「上げる」「かかる」「かける」「つく」「つけ る」「抜く」など移動や位置変化を表すものの場合にこのような多様性が生じることを指摘し、ドイツ語の分離動詞との平行性も指摘する。
16 : 30 ～ 17 : 00

「複合動詞の形態構造と自他交替」
影山太郎 Taro Kageyama（国立国語研究所）

• abstract(要旨)
日本語動詞の自他交替は，レキシコンで指定された単純語の和語動詞に起こるため，複合動詞は自他交替に関与しないのが原則である。たとえば，「押しつぶす」に対する「*押しつぶれる」，「切り取る」に対する「*切り取れる」などは非文法的であり，これらの非文法性は語彙的複合動詞の形成を規制する「他動性調和の原則」からも導かれる。ところが，例外的に「縫い上げる」と「縫い上がる」，「建て替える」と「建て替わる」のように自他交替が成立する場合があり，複合動詞の自他交替に関する制約はこれまで解明されていない。本発表では，まず，語彙的複合動詞と呼ばれてきたものに２種類の異なる形態構造——上掲の「押しつぶす」タイプの規範的な「動詞＋動詞」構造と上掲の「縫い上げる」タイプの「動詞＋補助動詞」構造——が区別できることを示し，次に，その構造的差異によって自他交替の可能性が決まることを明らかにする。