証拠性と感嘆性―ブルガリア語のl-分詞の意味的・機能的側面

本論文は，ブルガリア語のl分詞を伴う述語の諸形式の意味・機能について論じながら，l分詞の共通意味特徴の再定義を試みるものである．l分詞の働きの中には，「証拠性」と「感嘆性」という異質な働きがある．一見，関係がないかのように見えるこれらの働きは，共通の意味特徴によって関連付けることができる．本論文では，この意味特徴を「発話時における話者の意識的関与」，またはより広い意味で「知識に関する話者の態度」と定義付け，この概念は，出来事または発言に対する話者の現在（発話時）の意識的関与の状態という意味要素（「現在性」）を含んでいると同時に，発話時以前の何らかの意識的状態（「過去性」）を想定し，それを現在の意識的状態と対比させるという意味要素まで含意することを示した．また，この概念に含まれている意味要素が，l分詞の原型である「現在完了形」の基本的意味特徴に含まれている同様の意味要素と巧く合致することを論じた．

本論文のもう一つの論点となっているのは，分詞のタイプ，すなわち定過去分詞または半過去分詞が機能の分布を如何に特徴付けるかということである．定過去分詞が現れる形式には，非モーダル及びモーダルな働きがある．一方，半過去分詞が現れる形式はモーダルな働きのみを担っている．

Three Stages of Diffusion of Standard Japanese Forms: Railway Distance and Percentage of Standard Usage by 4 Clusters

In this paper a new technique for representing dialectal differences will be introduced. Two techniques of simplification are applied to represent geographical distribution patterns of standard Japanese. The first one is a representation of two-dimensional geographical distribution patterns by one dimension. The second one is plotting the geographical locations making use of railway distances from cultural centers. By simplifying the two-dimensional geographical distribution into one by railway distance, another dimension can show the average percentage of usage of standard Japanese forms for each prefecture. In order to see the patterns of distribution of the Japanese standard forms, "Kasai data", or numerical data of the "Linguistic Atlas of Japan", was utilized.

By comparing railway distances from Tokyo and from Kyoto in the form of scattergrams, the degree of linguistic influence of each of the capitals of Japan was analyzed. The overall pattern showed that railway distances surely correspond to standardization of dialects. The standardization of language in Japan proceeded in two historical stages. The first stage was standardization from Kyoto. The second stage appeared later from Tokyo. These two stages were ascertained by classifying 82 words into 4 clusters. West cluster words showed Kyoto-centered diffusion and east cluster words showed Tokyo-centered diffusion.

An analysis of younger Japanese people through a countrywide postal survey showed that standardization has proceeded with greater speed in recent years. Most of the words of the Kasai data approached the end of diffusion. This third stage of standardization seems to have been influenced by the mass media.

The Biblical Hebrew Suffix Conjugation in the Light of Canaano-Akkadian

The Biblical Hebrew (BH) verbal system has traditionally been understood as follows:

This phenomenon is generally explained as follows: when several verbs occur consecutively, the tense, mood and aspect (TMA) of the first verb can determine that of the following verbs.

In 1986, A. F. Rainey demonstrated in the light of Canaano-Akkadian that the BH wayyiķţol is a survival of the older Northwest Semitic yaqtul preterite, that the BH yiķţol is a reflex of the yaqtulu imperfect, and that the BH jussive and cohortative originate from yaqtul and yaqtula. This radically alters the two-by-two framework above, since the wayyiķţol can now be regarded as a preterite in its own right rather than as a converted ķɔţal.

In this paper, I focus on the BH suffix conjugation (ķɔţal) and demonstrate that the suffix conjugation should be characterized as a form which cannot be marked for TMA. The resulting framework is:

I illustrate how the TMA of the suffix conjugation are construed from the Aktionsart of the verb and the context using two text specimens, Genesis 2: 4-8 and 1 Samuel 1: 1-6.

On the Choice between wh-relatives when/where and a Preposition+which in Time and Place Relative Clauses

In this paper, we are concerned with establishing the principles governing the choice between a wh-relative and a prep(osition)+which in time and place relative clauses (=RCs), such as (i) the day [{when/on which} ...[e]...] and (ii) the street [{where/on which} ...[e]...]. We argue, among other things, that the cases where a wh-relative is used, which we call "hamidashi" (or "pressed-out") RCs, and those where a prep+which is used are two different types of RC. In "pressed-out" RCs the RC defines or characterizes the attribute of its antecedent and restricts the range of the attribute of the antecedent. This function differentiates "pressed-out" RCs from the "prep+which"-type (=restrictive RC). We claim that these semantic differences between the two types of RC can be attributed to the nature of the "gap" (indicated by "[e]" in (i) and (ii)) inside of the RC.

We also suggest that there is a parallelism between pre-modification and post-modification, more specifically, that "pressed-out" RCs (post-modifiers) are comparable with classifying adjectives (pre-modifiers) in their relevant respects.

Finally, we provide syntactic evidence that suggests that "pressed-out" RCs are not complements: rather, both "pressed-out" RCs and ordinary restrictive RCs (including the "prep+which"-type) could be on the same level (both outside complement elements) in the X-bar structure. To reconcile such a situation with the claim that the "pressed-out" RC and the "prep+which"-type are two different types of RC, we propose a hypothesis to the effect that types of RC can be determined according to the nature of the gap regardless of their phrase-structural position.