Top page
Japanese version
Vol. 17 No. l
(June, 1989)
Vol. 17 No. 2
(Sep., 1989)
Vol. 17 No. 3
(Dec., 1989)
Vol. 17 No. 4
(Mar., 1990)
Vol. 17 No. 5
(June, 1990)
Vol. 17 No. 6
(Sep., 1990)
Vol. 17 No. 7
(Dec., 1990)
Vol. 17 No. 8
(Mar., 1991)

Vol. 17 No. 1 (Jun. 17, 1989)

  1. Classification:
    paper
    Authors:
    HIGUCHI Kazue and TODA Masanao
    Title:
    A Structured System of Informational Features to Analyze Conversational Utterances
    Pages:
    1-24
    Descriptors:
    Japanese conversational sentences; natural language processing; feature system, semantic cases
    Abstract:
    This study is a part of the long range project initiated by Toda of creating a "chatting" machine which is supposed to be capable of engaging in daily conversation with human partners. One of the major requirements for such a machine is its ability to correctly discern numerous informational contents contained in each sequence of conversational utterances. The plan for the structural system of informational features described in this Paper is a part of our tentative answer to this challenge. Note that actual conversational utterances are made to serve multiple purposes, since the speaker of an utterance is, as a rule, a social game player as well as an information transmitter. Therefore, not only a far more elaborate system of informational categories like our system of informational features is needed than the one traditional semantic-case approaches customarily provide, in order to capture diverse aspects of propositional information, but also more is needed to handle extra-propositional information contents representing mainly the intentions of the speaker as a game-player. Thus we obtained 63 informational categories altogether, as introduced in this paper with respective accounts for their natures and functions.

  2. Classification:
    report
    Author:
    YIM Young Cheoul
    Title:
    The Language Life of Koreans in The United States: Results from A Questionnaire Survey
    Pages:
    25-33
    Descriptors:
    consciousness structure; language life; tradition-consciousness; tradition-consciousness with respect to language selection; language selection according to setting and interlocutor
    Abstract:

    The author conducted a questionnaire survey of 426 persons consisting of both Korean-Americans and Korean nationals presently resident in the United States in order to shed some light on the condition of their "language life and to identify factors that which relate to it.
    The findings of the survey can be summarized as follows:
    (1) Korean-Americans have a lot of contact with Korea and with fellow ex-patriots and while there is a gap between generations, they have a strong ``Korean consciousness.''
    (2) There is a great deal of contact with the Korean language in everyday life, Korean is used very much in the home and a great many households use Korean as the dominant language in a bilingual life-style.
    (3) There is a strong tradition-consciousness with the Korean language, especially in the lower (as compared to the upper) social classes, and in those born in Korea (as opposed to the United States).
    (4) In every setting, Korean is used more than English, but Korean is more prominent in Private, informal settings than in public, formal ones.
    (5) There is a tendency for Korean to be used more to people older than the speaker. In other words, the age of the listener plays a large part in the choice of whether to use Korean or English.
    (6) A strong link exists between Korean tradition-consciousness and actual usage of Korean language.

Vol. 17 No. 2 (Sep. 16, 1989)

  1. Classification:
    paper
    Authors:
    INOUE Fumio and OTA Utako
    Title:
    Multi-Variate Analysis of Geographical and Individual Differences of Dialect
    Pages:
    49-63
    Descriptors:
    dialect division; geographical differences; Gifu prefecture; factor analysis; cluster analysis
    Abstract:

    This study was done to measure degrees of differences of language among localities and among informants of a locality The data for this study is the one showings degrees of mutual communality of answers in a dialect survey of 32 informants of 8 localities in a monotonous area in central Japan The data was put into computer and two multi-variate analyses were applied Factor analysis showed similarities and differences of dialects of 8 localities clearly Cluster analysis was also applied to the data Degrees of similarities of 8 localities were shown more clearly, and communality of speech of informants from adjacent localities was also shown clearly This weans that language differences between neighboring localities are not always decisive, and that answers of informants in researches of Iinguistic geography should not be taken absolutely.

  2. Classification:
    report
    Author:
    MIZUTANI Sizuo
    Title:
    Word Construction Taken Account of High-Order Combinaitions
    Pages:
    64-70
    Descriptors:
    word construction; compound word of high-order; tree; branching; distribution of patterns; fluctuation of values

  3. Classification:
    report
    Author:
    NAKANO Hiroshi
    Title:
    Writing-Form Variation of Words in Romanized Japanese
    Pages:
    71-84
    Descriptors:
    romanized Japanese; writing system; spelling; writing-form variation

  4. Classification:
    news
    Author:
    ITO Masamitsu
    Title:
    Micro-OCP Workshop
    Pages:
    85-87
    Descriptors:
    OCP(Oxford Concordance Program); Micro-OCP

Vol. 17 No. 3 (Dec. 16, 1989)

  1. Classification:
    paper
    Author:
    MIZUTANI Sizuo
    Title:
    Japanese Imperatives in Which Their Subjects Occur
    Pages:
    91-111
    Descriptors:
    subject of imperatives; Japanese; occurrence rate; ``Onihei Hankatyoo''; vocative; case-marker ``ga''; kakari-zyosi (thematic postposition); ``na'' for prohibition; Polya-Eggenberger distribution
    Abstract:

    It is not so rare in Japanese imperatives that their subjects occur explicitly To certify this fact, the occurrence rate of such type to the whole imperatives was inferred by ratio-estimage, and usage of subjects in imperatives was examined.

    As materials, 122 novels of series ``Onihei Hankatyoo'' by Ikenami Syootaroo, in Bunsyun Libray (18 vols. published already) were taken, since a lot in the Edo Period. The result obtained from a random sample of 24 works amounted to 3.518+1.504% with confidence probability 0.95, therefore it can not be said that occurrences of the subjects are negligible as exceptional events. Numbers of such imperatives per work follows approximately to Polya-Eggenberger distribution. The 122 examples of such kind found in the whole 122 works are listed up at the end of this paper. If the notion of ``zero case-marker'' was accepted, then some ten examples could be added besides, though they were omitted in this survey for the sake of clearness. Notice that 35 examples have their subjects of the third person.

    In addition, vocatives appended to imperatives are not to be interpreted as semantic substitutes for subjects. An evidence of inadequacy of such interpretaion is given from the statistical point of view.

    Examples are classified by their grammatical forms into three groups: The subject is represented by (1) [noun phrase] ga (22 examples), (2) [noun phrase] wa (73 examples), or (3) [noun phrase] mo (27 examples).

    As a rule, the first form is exerted in the case to choose the person(s) to obey the order among people possible to do so. The second form is exerted in the case to specify the very person referred to by the subject to do a certain act, mostly in contrast with (an)other person(s) to be requested some different act(s), especially to assign him/her/them some task. The third form is exerted in the case that, at the stage someone does, did or is going to do an act, the addressee(s) is/are also to be asked to do the same (sort of) act.

  2. Classification:
    report
    Authors:
    TSUCHIYA SIn-ichi and MATSUBARA Kyoko
    Title:
    The Use of Euphonical Change Form in Higuchi Ichiyo's Novel: On Making Her Style Halfway Between the Classical and the Common
    Pages:
    112-122
    Descriptors:
    style; written language; style halfway between the classical and the common; recitative; euphonical change form

  3. Classification:
    report
    Author:
    SANADA Haruko
    Title:
    Automatic Sentence Style Conversion in Japanese: Conversion to DEARU-Style
    Pages:
    123-133
    Descriptors:
    style conversion; style; common style; polite style; Desu/Masu-style; Dearu-style; conjugation of verb
    Abstract:

    This paper deals with an automatic conversion system which converts any sentence style to the dearu-style in Japanese. The followings are the characteristics of the system:
    (1) It is available on personal computers.
    (2) It converts clause-final and sentence-final predicates in all sentences except conversational sentences.
    (3) Verb-conversion incorporates a probability logic according to circumstances.

    The system was developed with the following considerations in mind;
    (1) Conversion of clause-final predicates
    (2) Conversion of polite verbs

    Two problems still remain. One is the conversion of conjunctions such as desukara, desuga, etc. The other concerns by examining accompanying homonymous verb forms; this may be facilitated by examining accompanying postpositions.


Vol 17 No. 4 (Mar. 17, 1990)

  1. Classification:
    paper
    Author:
    UMEDA Michio
    Title:
    Quantitative Analysis of Characters in A Word Based on The Similarity of Pattern Shapes
    Pages:
    147-168
    Descriptors:
    Japanese word dictionary; word determination; similar characters in shape; character recognition
    Abstract:

    This paper discusses some characteristics of a character as an element of Japanese words. A word determination rate is defined to quantitatively evaluate the probability that an uncertain character in a word is determined by using a word dictionary. The rate calcurated for characters in all words is 27.5%. That is, about 1/4 of characters are determined as a word from the knowledge of before and behind characters. The rate is also calcurated for all characters when candidate categories of these characters are selected at random. Next. similar character categories for each character are choosed by using peripheral directional shape features and the rate is calcurated in this case. Furthermore. characteristics of a character as a word are discussed for two different category sets selected by the matching order and the matching distance in a feature matching operation. It is shown that the matching distance is much effective in a character recognition system.

  2. Classification:
    paper
    Author:
    MARUYAMA Naoko
    Title:
    Case Particles, Cases, and Valency
    Pages:
    169-192
    Descriptors:
    Mizutani grammar; case representation; adverbialization core case; oblique case; case continuity
    Abstract:

    The case of a noun is mainly represented by a kakujoshi(case particle) in a Japanese sentence. This paper describes an attempt to clarify the function of Japanese case particles by examining their relationships with representing cases and the valency of governing predicates. Functions of 9 case particles are categorized into 28 cases, and their relationships with 1,749 valency patterns are identified. Particles to, ni, and de are sometimes better to be considered as adverbialization depending on the situation. The continuity among core cases, oblique cases, and adverbialization is also discussed.

  3. Classification:
    report
    Author:
    ISII Hisao
    Title:
    Length of Words Used in A Magazine
    Pages:
    193-206
    Descriptors:
    word; length; morph; writing; sound; word class by origin; part of speech; correlation; regression

Vol. 17 No. 5 (Jun. 16, 1990)

  1. Classification:
    paper
    Authors:
    TANIHARA Kimio, YEN Jui Chen and Debbie LEE
    Title:
    A Study on Japanese Classifiers: Men, Mai, Hon, Ko, and Tsu
    Pages:
    209-226
    Descriptors:
    Japanese classifiers; prototypes; criterial properties; concrete vs. abstract nouns; mental images
    Abstract:

    This study is to determine the criterial properties of Japanese classifiers, men, mai, hon, and tsu in relation to the features of nouns. A questionnaire survey of 425 Japanese native speakers was conducted.
    Results of the survey are summarized as follows:
    (1)
    men for a flat and relatively large surface
    mai for a flat, thin object
    hon for a long, this object, or an object of enormous scale inside
    which there is no human activity
    ko for a solid object with a fixed shape and certain weight capable
    of existing on its own
    tsu for both concrete and abstract nouns without any specific criterial
    properties,
    capable of replacing ko in many cases,
    unnatural for prototypes of mai and hon
    (2) Japanese native speakers tend to choose classifiers for an object depending on the images they seize upon their encounter with it in their daily life in spite of the actual shape or attributes of that very object.

  2. Classification:
    paper
    Authors:
    ISHIZAWA Hidemi and FURUGORI Teiji
    Title:
    Reading Comprehension: Processes and An Approach to Computational Model
    Pages:
    227-240
    Descriptors:
    reading; comprehension; newspaper column; survey; computational model
    Abstract:

    Reading is an act of understanding and comprehending something. However, how and what we understand will depend on the materials we are to read. In this paper, we consider a way of reading newspaper columns that will reach what we may call a critical and/or explorative comprehension: we Present a survey result on what reading comprehension means and give some thoughts on a computaional model for the processes of comprehension.

  3. Classification:
    note
    Author:
    MIZUTANI Sizuo
    Title:
    Indices on Similarity Between Word-Forms
    Descriptors:
    divergence index; dissimilarity index; similarity index; non-symmetricity; distance; mora; parody; fluctuataion of word-forms
    Pages:
    241-244

Vol 17 No. 6 (Sep. 14, 1990)

  1. Classification:
    paper
    Authors:
    MINAGAWA Naohiro and KASHU Kan
    Title:
    A Relationship Between The Component Word Relatedness in A Haiku and The Degree of Liking for A Haiku
    Pages:
    265-272
    Descriptors:
    word relatedness in a sentence; linguistic intuition; sympathetic comprehension of a literal work
    Abstract:

    In this study, we tried to apply a method used to analyze sentence construction to Haikus. A Haiku is a type of poem with a fixed form that is unique to Japanese, and has two rules. First, it is composed of 17 characters. Second, it has to include one Kigo, the word that indicates the season. Therefore to understand a Haiku, the relationship between a Kigo and the other words must be investigated.

    As a preparatory analysis, we had 74 readers choose Haikus which each reader felt were agreeable from a collection of 232 Haikus and defined the number of readers who chose each Haiku as the degree of liking. According to this analysis, we defined l 6 Haikus which more than 9 readers selected as Haikus with high liking, and 24 Haikus which none of the readers selected as Haikus with no liking. We chose 8 Haikus respectively from these two groups of Haikus as the stimuli of the experiment. We divided each Haiku stimulus into its component words. Forty students rated the degree of relatedness between the words of each Haiku. Half of the subjects rated the 8 Haikus with high liking, and the other half rated the 8 Haikus with no liking.

    We averaged relatedness scores for each Haiku between subjects after averaging the averaged relatedness scores for each component word. There was no difference between Haikus with high and no liking. This result showed that the degree of liking of a Haiku was not determined by the relatedness score as a whole. However, we also compared the score of word relatedness for a Kigo with a whole relatedness score. This analysis showed that all components of a Haiku were related to a Kigo, and Haikus in which the components were strongly related to a Kigo were liked by many readers.

  2. Classification:
    report
    Author:
    SHIMAMURA Naomi
    Title:
    Acquisition Rates of Kanji by School Children: The Difference Due to Kanji Assigned to Each Grade
    Pages:
    273-279
    Descriptors:
    Kanji; acquisition rates; school children
    Abstract:

    This paper presents acquisition rates of Kanji by school children. The data used here are from the National Language Research Institute Research Report 95, `Acquisition of Joyo Kanji by School Children'.

  3. Classification:
    report
    Author:
    OGINO Tsunao
    Title:
    A Method to Evaluate Validity of Semantic Description of Japanese Dictionaries
    Pages:
    280-286
    Descriptors:
    Japanese dictionary; semantic description; introspective research method

  4. Classification:
    report
    Author:
    MIYAZIMA Tatuo
    Title:
    The Relations Among The Length, Oldness and Frequency of Classical Japanese Words
    Pages:
    287-300
    Descriptors:
    word-length; oldness of words; frequency; frequency dictionary of classical Japanese; Zipf

  5. Classification:
    report
    Authors:
    YOSHIDA Masakazu, SEKI Yoko and KOSHIBA Ryosuke
    Title:
    Measuring Method of Autographic Alignemnt: Measuring of Handwriting with a Computer 1
    Pages:
    301-308
    Descriptors:
    handwriting; handwriting identification; measurement of character; autograph

  6. Classification:
    note
    Author:
    MIZUTANI Sizuo
    Title:
    A Line of Discontinuity Across Japanese Language Processing
    Pages:
    309-310

  7. Classification:
    note
    Author:
    NAKANO Hiroshi
    Title:
    CCL of UMIST: Letter from England
    Pages:
    311-313

Vol. 17 No. 7 (Dec. 15, 1990)

  1. Classification:
    paper
    Authors:
    SIBATA Takesi and SHIBATA Ritei
    Title:
    Is Word-Accent Significant in Differentiating Homonyms in Japanese, English and Chinese?
    Pages:
    317-327
    Descriptors:
    accent; homonyms; probabilistic model; binominal model; Bernoulli trial
    Abstract:

    To answer the question if word-accent is significant for differentiating homonyms, a probabilistic model is constructed and applied for real data. It is a binominal model for the number of key words in a set of homonyms, each of which generates a group of words with the same accent. The common probability of a keyword occurrence indicates how stronly homonyms are differentiated with word-accent, although the model fitting is good for Japanese but not so for other English or Chinese. The estimated probabilities are 0.1357, 0.0047 and 0.71 respectively for Japanese, English and Chinese, which are significantly different.

  2. Classification:
    paper
    Author:
    ISII Hisao
    Title:
    Observational Error of Frequency in Language: A Case of Personal Equation
    Pages:
    328-353
    Descriptors:
    frequency; hiragana; Japanese medieval tale; observational error; normal distribution; regression curve
    Abstract:

    Discussed here is a distribution model of errors in the observation which the author made of the hiragana frequency in one of Japanese medieval tales. The set of his observation contains 98 values, for which the true values are inferred from the observation by another person, and the sum of which is more than 12,500. The distribution of the observational errors in case of this personal equation is characterized as follows:
    (1) Many of the observation values have no error.
    (2) The absolute or relative errors of the observational values may present a normal distribution partially or a complex distribution.
    Besides probability functions, regression curves are assumed.
    Some kinds of mistakes apt to mingle with observational error, such as failure of photocopying of material, are also discussed.

Vol. 17 No. 8 (Mar. 16, 1991)

  1. Classification:
    paper
    Author:
    TAKEUCHI Haruhiko
    Title:
    Measuring The Meaning of Linguistic Hedges by Fuzzy Rating Scale Method
    Pages:
    365-376
    Descriptors:
    fuzzy rating scale method; linguistic hedges; words on evaluation; meaning; frequency of personal use
    Abstract:

    This paper presents the psychological measurement of 80 linguistic hedges by fuzzy rating scale method. The merit and demerit of fuzzy rating scale method is briefly described. Frequency distributions of linguistic hedges are divided into 5 groups and figures of typical distributions are shown.

  2. Classification:
    paper
    Author:
    KUSANAGI Yutaka
    Title:
    Recursive Transitional Parser Using Phrase Structure Grammar
    Pages:
    377-392
    Descriptors:
    top-down; depth-first search; independent rule component; backtracking; indeterministic
    Abstract:

    The purpose of this paper is to consider an algorithm of a parser with the phrase structure rule component independent from the parsing unit, and to implement it by a Lisp program. This system is a top-down depth-first parsing with a backtracking mechanism, but it parses all structures of an syntactically ambiguous sentences. This algorithm has the following features:
    1) It can be used to parse sentences of any language;
    2) It is easy to strengthen only by means of revision of the rule component by linguists unfamiliar with parsing;
    3) The first output of this system is a sentence structure in labeled brackets in the form of a LisP list, so that this output is easy to use as the input for the semantic analysis; and
    4) The second output is a tree-diagram bY which linguists can easily examine the feasibility of rules.
    Since the algorithm is explained in accordance with linguistic analyses, linguists unfamiliar with computer programming are supposed to be able to understand a parsing mechanism. It is hoped that this paper will contribute to increase the number of linguists interested in the natural language processing.

  3. Classification:
    report
    Authors:
    IZUMI Asako and MIZUTANI Sizuo
    Title:
    On Sound-Distribution in Tutui Yasutaka's ``ZANZOO NI KUTIBENI WO''
    Pages:
    393-402
    Descriptors:
    speech sound; letter; vanishing of a sound; relative frequency; word processor; entropy
    Abstract:

    ``Zanzoo ni Kutibeni wo'' is an experimental story that speech sounds vanish one by one and objects referred to used by such sounds have also vanished in the world. This paper considers Tutui's plan of the order of vanishing sounds by means of statistics of sound-distributions, and transition of entropy on each chapter.

  4. Classification:
    comment
    Authors:
    ITOU Katunobu, OKUMURA Manabu and TANAKA Hozumi
    Title:
    On Efficient Retrieval Methods from Large-Scale Databases: Hash and Trie
    Pages:
    393-402
    Descriptors:
    machine readable dictionary; corpus; data retrieval; hash; trie; sort; variation; LR-parsing; dynamic hash

Back to Homepage