Top page
Japanese version
Vol. 22 No. l
(June, 1999)
Vol. 22 No. 2
(Sep., 1999)
Vol. 22 No. 3
(Dec., 1999)
Vol. 22 No. 4
(Mar., 2000)
Vol. 22 No. 5
(June, 2000)
Vol. 22 No. 6
(Sep., 2000)
Vol. 22 No. 7
(Dec., 2000)
Vol. 22 No. 8
(Mar., 2001)

Vol. 22 No. 1 (Jun. 14, 1999)

  1. Classification:
    Report
    Authors:
    NAKAO Eiko and MIYAZIMA Tatuo
    Title:
    A Study on Classifiers of Foods
    Pages:
    1-10
    Descriptors:
    classifiers; foods; norm; consciousness; usage; situation; supermarket; leaflet; cookbook

  2. Classification:
    Note
    Author:
    OGINO Tsunao
    Title:
    How Many Examples are Required in Language Research; A Proposal of 'A Mrthod to Estimate Probability of New Appearance'
    Pages:
    11-17
    Descriptors:
    number of samples; number of answers in questionnaire survey; probability of new appearance

  3. Classification:
    Note
    Author:
    HAMASAKI Nomi
    Title:
    Effects of Contexts on the Restriction of Japanese BA-Conditionals
    Pages:
    18-26
    Descriptors:
    the restriction of Japanese conditionals; ba-conditionals; tara-conditionals; hypothetical contexts; acceptability

Vol. 22 No. 2 (Sep. 13, 1999)

  1. Classification:
    Paper
    Authors:
    KANZAKI Kyoko and ISAHARA Hitoshi
    Title:
    Adverbial Semantic Relations between Adnominal Constituents and their Head Nouns
    Pages:
    51-65
    Descriptors:
    Adnominal Constituent; Adverbial Semantic Relations; Paraphrastic Relationship; Context
    Abstract:
    This paper discusses the treatment of the adverbial semantic relations in which adnominal constituents do not modify their head nouns directly, but refer to the states of being of the referents of their head nouns.

    We were able to find that in such relations between adnominal constituents and their head nouns, adnominal constituents are classified into five types based on their interaction with the state of being of their head nouns. Firstly, adnominal constituents represent how the state of being of their head noun is. Secondly, they represent the state of being of their head nouns. Thirdly, they represent the evaluation of the state that is indicated by "N + V" structure in context. And then, they represent that the action/state of the verb is in the state indicated by the adnominal constituent. Finally, they represent how concrete instances of the referents of their head nouns are selected among the concepts of their head nouns.

    Furthermore, We treat the semantic behavior of "(pure)," as an example, which modifies the state of being of the referent of the head noun. And we found that some interpretations of adverbial semantic relationships need inferences by contexts.

  2. Classification:
    Report
    Author:
    SASAHARA Hiroyuki
    Title:
    Frequency of Exposure of University Students to Kanji Variants
    Pages:
    66-79
    Descriptors:
    Frequency; Jitai; Kanji Variants; Traditional forms of kanji; Revised forms of kanji; Preference; Familiarity; Word processor
    Abstract:
    In this investigation, some 200 students from women's universities in Tokyo were asked to record each time they saw kanji belonging to one of three pairs of variant forms. In all 1120 sightings were reported, in books and a variety of other media. The results show that of the pairs (gull, used in the name of the Meiji era novelist Mori Ogai) and ϡ (grab), exposure to the traditional forms predominated, while for the pair ͡ (noodle), exposure to the simplified form predominated.

    The frequency of exposure measured in this investigation may be understood as an important element in the degree of familiarity with and preferences for kanji variants.

  3. Classification:
    Note
    Author:
    TSURUOKA Akio
    Title:
    Syntactical Units Of Japanese Based On W-UNIT
    Pages:
    80-87
    Descriptors:
    Basic syntax of Japanese; word; W-UNIT

  4. Classification:
    Miscellaneous
    Title:
    Descriptors and Abstracts
    Pages:
    90-91

Vol. 22 No. 3 (Dec. 13, 1999)

  1. Classification:
    Note
    Authors:
    KAGEURA Kyo and SEKINE Satoshi
    Title:
    On Ogino's "Method to Estimate Probability of New Appearance"
    Pages:
    93-103
    Descriptors:
    unseen events; urn model; binominal distribution; hypergeometric distribution; reverse problem; random variable
    Abstract:

  2. Classification:
    Note
    Author:
    YAMAUCHI Kana
    Title:
    An Analysis of Essay Examination Data Using a Many-facet Rasch Model
    Pages:
    104-115
    Descriptors:
    essay test data; Many-facet Rasch model; rating objectivity
    Abstract:

  3. Classification:
    Miscellaneous
    Title:
    Proceedings of the 43rd Annual Meeting
    Pages:
    116-126

  4. Classification:
    Miscellaneous
    Title:
    Descriptors and Abstracts
    Pages:
    128

Vol. 22 No. 4 (Mar. 13, 2000)

  1. Classification:
    Paper
    Author:
    TAKEDA Kanji
    Title:
    On Sore and Kore which Refer to the Context in the Sentence
    Pages:
    129 - 146
    Descriptors:
    demonstrative; sore; kore; reference to the context in the sentence; antecedent context; sore sentence; kore sentence; semantic relation between linguistic units
    Abstract:
    In this paper we investigate sore and kore which refer to the context in the sentence, judging from the semantic relation between the antecedent context and the sore/kore sentence.

    We classified the relation between the antecedent context and the sore/kore sentence into the following six types, by the presence of the conjunctive expressions or the linguistic forms.

    1. When the antecedent sentence is a subordinate clause.
    2. When the antecedent sentence and the sore/kore sentence have the relation between cause and effect.
    3. When the sore/kore denies or opposes the content of the antecedent context.
    4. When in the sore/kore sentence a content is added to the antecedent contest.
    5. When the antecedent context is a question.
    6. When the antecedent context and the sore/kore sentence are a subject-predicate relation. When in the sore/kore sentence a new subject is introduced into the context. When there are no conjunctive expressions between the antecedent sentence and the sore/kore sentence.
    We counted the sore/kore sentences in materials. As a result, we found that the sore sentence was used in A -- E, and that the kore sentence was used in F.

  2. Classification:
    Report
    Authors:
    BABA Yuka, MIYAKE Kiyoshi, BABA Mitsuru and YOSHIDA Norio
    Title:
    A Statistical Quantitative Analysis on the Construction of `zuwa' in GENJI MONOGATARI
    Pages:
    147 - 156
    Descriptors:
    zuwa; subjunctive condition; modification of declinable words; Spearman's rank correlation coefficient; concord rate; Genji Monogatari
    Abstract:
    Zuwa, which grammatically consists of zu of the negative auxiliary verb and wa of the postpositional particle, is a subject of continuing debate among Japanese classical grammarians whether it were used as subjunctive condition or whether it modified declinable words in Genji Monogatari. We examined the above issue by the statistical quantitative analysis of the concord rate between zuwa and the specified expressions, which include mu, beshi and ji of the auxiliary verb. As a result, it has been suggested that zuwa in Genji MOnogatari was generally used as subjunctive condition. This result in the main agreed with the results of the previous work by the viewpoint of interpretation, and it seems to show the validity of the analytical result and method in this paper.

  3. Classification:
    Note
    Author:
    KUSANAGI Yutaka
    Title:
    Joyo Kanji and Place Names
    Pages:
    157-161
    Descriptors:
    Kanji in common use; municipality names; Kanji for given names; JIS Kanji; total numbers of Kanji; numbers of variety of Kanji

  4. Classification:
    Miscellaneous
    Author:
    TSUCHIYA Shin-ichi
    Title:
    In Memory of Mr. Hiroshi Nakano and His Works
    Pages:
    162 - 165

Vol. 22 No. 5 (Jun. 12, 2000)

  1. Classification:
    paper
    Author:
    NODA Harumi (Kobe Gakuin University)
    Title:
    Co-occurrence of the Adverb zenzen with Affirmative Predicates
    Pages:
    169-182
    Descriptors:
    zenzen; tigau; absolute standard; context; variations
    Abstract:
    The adverb zenzen can co-occur with affirmative predicates under certain conditions. This study aims to grasp the conditions for the co-occurrence of zenzen with affirmative predicates. The conclusions are as follows:
    1. The words tigau and dame were found to frequently co-occur with the adverb zenzen. THe words tigau and dame have two features in common. Firstly, these two words both indicate that a state is distinctly differ from an absolute standard, and if zenzen co-occur with these words, zenzen means that the distance from the standard is great. Secondly, the individual meanings of these two words aren't strong.

      The words which lack such features cannot co-occur with zenzen without certain contexts.

    2. Zenzen can co-occur with affirmative predicates in the case when speakers deny hearers' opinions and also in the case that speakers wish to express that one thing is much better than something else. Especially, the younger generation accepts this usage. This usage originated from zenzen's meaning mentioned in 1.
    3. The younger generation also accepts the co-occurrence of zenzen together with ii/heiki/daizyoobu when compared to the older speakers. Furthermore, they consider zenzen oisii as more acceptable than zenzen mazui. These two results mean that "zenzen + affirmative" is tending to express preferable state.

  2. Classification:
    paper
    Author:
    OH Hyunjung (Osaka University, Graduate school Faculty of Letters)
    Title:
    Word Order of Lexical Semantic: Focus on Verbs
    Pages:
    183 - 200
    Descriptors:
    lexical semanttics; word oerder; verbs; modifying nouns; adjectives
    Abstract:
    This paper examines word-order tendency of verbs (V) and other modifying elements (nouns (N), adjectives (A)) in modern Japanese and reveals that word order is connected to the meaning of the verb to which it modifies. Specifically, the finding show that; when a verb and noun are in the modifier, (1) when a verb expresses the action of the head noun that it modifies word order tendency is V+N1+N2; (2) when a verb expresses the nature or attribute of the head noun that it modifes (noun phrases expressing situation, noun phrases expressing possession), in word order tendency is N1+V+N2, however, when the noun phrases expressing attribution, the word order tendency is V+N1+V2; (3) when a verb expresses resultant state of the head noun that it modifes, then both the situation in (1) and (2) are deserved. On the other hadn, (4) when a verb and adjective are in the modifier, the word order tendency is V+A+N. While it is usually thought that word order of modern Japanese has a loose rule system, the results of the present study, show that word order in the modifier adheres to some basic principles.

Vol. 22 No. 6 (Sep. 12, 2000)

  1. Classification:
    Paper
    Authors:
    MATSUURA Tsukasa (The University of Tokyo), and KANADA Yasumasa (The University of Tokyo)
    Title:
    Authorship Attribution in Japanese Modern Sentences via N-Gram Distribution
    Pages:
    225-238
    Descriptors:
    authorship attribution; Japanese modern sentence; n-gram distribution
    Abstract:
    We propose a method for authorship attribution via distribution of n-grams of characters. The proposed method uses a function named dissim which calculates dissimilarity between two sets of text data. It is applicable to all the machine-readable texts because (1) it does not require any additional information on texts, i.e. any pre-analyses such as tagging or morphological analyses, (2) it does not presuppose any manners of specific languages or compositions. It is language independent and can be adapted to new sources without modification. Since there is no need for recognition of boundaries between words, languages in which separators between words are often ambiguous, such as Japanese or Chinese, could be analyzed easily via dissim. In experiments with 92 works in total written by 8 Japanese modern authors, we had cases where dissim achieved the accuracy of 100% in authorship attribution. When we evaluate effectiveness of the proposed method, Tankard's method and divergence (Kullback-Leibler information) are taken as competitors.

  2. Classification:
    Paper
    Author:
    YOSHIOKA Yasuo (The National Language Research Institute)
    Title:
    Social and Regional Variation of the Politeness Strategies and the Norm Consciousness in Japanese
    Pages:
    239-251
    Descriptors:
    honorific behaviour; respect expression; modesty expression; glorified usage; error; public opinion survey; survey of major Kyushu cities
    Abstract:
    This research makes an analysis of the glorified usage of the modesty expression ageru(to give) and honorific expressions of common language and addresses the social and regional variations in politeness strategies and norm consciousness in Japanese. The data analysed was from a public opinion survey of Japanese language and a survey of major Kyushu cities.

    The relationship between the survey results and the social attributes of informants was analysed. The following tendencies became clear regarding politeness strategies and norm consciousness.

    1. Glorified usage of the modesty expression "ageru"(to give)
      There is a high usage rate of the glorified form "ageru" (to give) in metropolitan areas and among young women, however, the usage rate is low in areas of varied honorific dialects. In area with varied honorific dialects, the rate of inappropriate usage of glorified forms is around 90 %, which is extremely high. In older age groups, who favour the Tokyo dialect, however, the rate of inappropriate usage is relatively low.
    2. Norm consciousness of honorific expressions
      70 % of people in areas of varied honorific dialects answered that the act of not using honorific expressions towards hearers (face to face) "gets on their nerves". In contrast, 60 % of people in simple honorific dialectal areas answered that the speakers' act of not using honorific expressions towards hearers (face to face) "doesn't get on their nerves".

      In area with varied honorific dialects, 80~90 % of people thought it was inappropriate not to use honorific expressions towards the act of third party is, which is extremely high. Conversely, among junior high school and high school students, it was 50~70 %, which is low when compared to that of other age groups.

    3. Politeness strategies when referencing the hearers' act In simple honorific dialectal areas, the rate of not using honorific expressions was twice as high as that in varied honorific dialectal areas.

      In metropolitan areas, there was a high usage rate of high-levelled honorific verbs like "nasaru" and "o ~ ninaru" forms. The rate of using correct honorific expressions was high and about to that in varied honorific dialectal areas. On the other hand however, overuse errors and the use of redundant double honorifics ("mosareru") were high as well.

      In varied honorific dialectal areas, honorific auxiliary verbs ("reru, rareru") forms were highly used. The rate of using correct honorific expression was high, equal to that in metropolitan areas. The rate of not using honorific expressions was the lowest.

    In this study, it was revealed that the language situation of honorific acquisition (i.e. whether the language repertoire of local society where informants live has varieties of honorific expressions or not) has a strong influence on honorific behaviour and norm consciousness in common language.

  3. Classification:
    Note
    Author:
    Mizutani Sizuo
    Title:
    Rhetorical Questions with Conditional Clauses;From Meaning and Logic to Syntactics
    Pages:
    252-256
    Descriptors:
    rhetorical question; conditional clause; negation; ku-hukugoo (roughly, compound sentence); scope of predication; meaning; pragmatic equivalence; proof for verval reasoning

  4. Classification:
    Note
    Authors:
    SEKINE Satoshi (New York University), and KAGEURA Kyo (National Institute of Informatics)
    Title:
    Further Notes on Ogino's "Method to Estimate Probalibity of New Appearance"
    Pages:
    257-261
    Descriptors:
    probability of new appearance; Good-Turing estimate

Vol. 22 No. 7 (Dec. 12, 2000)

  1. Classification:
    Paper
    Author:
    HIRAI Yoko (Tokyo Metropolitan University)
    Title:
    On the Effects of the Word Processor upon the Ability to Write Kanji and the Errors
    Pages:
    267-280
    Descriptors:
    Kanji writing; word processor; error pattern; vocabulary; logistic analysis
    Abstract:
    The effect of depending on a word processor upon the Kanji ability of writing in a short time was examined. Forty writing tasks including 60 Kanji characters were administered. The response time was limited to 15 seconds for each task. The error pattern of a subject's responses was classified in two ways; one has 4 categories indicating the degree of correctness of the response as a character, and the other has 5 categories indicating the frequent type of his/her errors relative to other sbjects. Logistic regression models were used to describe the relationships between these classifications and the time amount of the word processor use and/or the time spent in manual writing in a day. The effect of lowering the ability to write Kanji was identified in both of the long use of the word processor and the shortness of the manual writing. The former had a larger effect than the latter. It was also found that the subjects with a small large vocabulary tended to write correct Kanji whereas the subjects with a small vocabulary tended to write wrong Kanji with the same sound and different meaning, or, did not write at all. This means that a large vocabulary also has the effect to alleviate the decreasing ability caused by the word processor.

  2. Classification:
    Paper
    Author:
    KAGEURA Kyo (National Institute of Informatics)
    Title:
    The Dynamics of Phenomena and the Dynamics of Data --- On the Relationship between Events and Structures in Terminology ---
    Pages:
    281-302
    Descriptors:
    terminology; lexical growth; dynamic/static description; event; structure; binomial interpolation/extrapolation
    Abstract:
    In this paper I examine the relationship between actual growth of terminology and a mathematical model of terminological growth based on binomial interpolation/extrapolation which reflects the "incompleteness" of the terminological data. In order for that, I will introduce the basic mathematical model, and define the structural framework of tracing the terminological growth. Then, I will examine the basic conditions under which the structural models can be regarded as the description of actual terminological growth.

  3. Classification:
    Note
    Author:
    TSUCHIYA Shin'ichi (Kyoritsu Women's University)
    Title:
    New Trends in Use of Chinese Characters
    Pages:
    303-305
    Descriptors:
    rate of Kanji using word; native Japanese words; word processor

Vol. 22 No. 8 (Mar. 13, 2001)

  1. Classification:
    Paper
    Authors:
    TAKEUCHI Kazuhiro, and MATSUMOTO Yuji (Nara Institute of Science and Technology)
    Title:
    Relation between Text Structure and Linguistic Clues: An Investigation on Text Structure of Newspaper Articles
    Pages:
    319-334
    Descriptors:
    Text Structure; Rhetorical Structure Theory; Reliability of Coding Scheme; Decision Tree Learning
    Abstract:
    In this paper, we investigate the relation between linguistic clues and the relevance of sentences in a text. In the area of linguistic, many frameworks have been proposed to represent the relevance in terms of text structure. However, the result of the text structure analysis which humans do have ambiguity. We perform an experiment to know the agreement among 5 human participants in the text structure analysis of 32 newspaper articles. As a result, we find out that there is a difference in the difficulty to analyze the text structure. The judgments of relevance between the sentences that are not adjacent in the text are difficult even for human participants. On the other hand, the judgments of relevance of the adjcent sentences have comparatively higher agreemnet rate. Furhtermore, from an experiment using decision-tree-learing method, we show that superficial linguistic clues are useful to predict the relevance of a pair of adjacent sentences.

  2. Classification:
    Note
    Author:
    ITO Masamitsu (National Language Research Institute)
    Title:
    Full-Automatic Text Composing System "Fujimura" (ver 3.0)
    Pages:
    335-354
    Descriptors:
    random generation; second-order word approximation; 2-grams; automatic pastiche; artificial intelligence for the humanities; song-writer android; full-automatic composition; popular songs; Matsutoya Yumi

  3. Classification:
    Book Review
    Author:
    NIWA Yoshiki (Hitachi Central Research Laboratory)
    Title:
    "Keiryo Johogaku" (Informetrics) by Kageura Kyo, Maruzen (2000)
    Pages:
    355-357

  4. Classification:
    Review
    Author:
    MIYAZIMA Tatuo (Kyoto Tachibana Women's University)
    Title:
    Samples of Statistical Studies
    Pages:
    358-365

Back to Homepage