back to research page

back to HOME

* Lexical Research on Animal Group Nouns in English

English Language and Literature (Chuo University; Tokyo) 45 (2005/February) pp. 233-57

This paper tries to list every animal group noun in the English language such as "a pride of lions", "a tower of giraffes", "a shiver of snakes", "a parliament of owls", "an army of ants". Through extensive lexical research, five main types of animal group nouns (semantic categories) have been suggested:

1. nouns to describe a prominent feature of a certain animal when group together
2. nouns to describe an image associated with groups of a certain animal
3. nouns which are based on the sound an animal makes
4. nouns which reflect how people commonly feel about a certain group, such as joy or fear
5. nouns used as units of livestock breeding

Finding about 250 kinds of animal group nouns through my research, I have made an exhaustive list of animal group nouns, which will enable Japanese learners of English to find appropriate or possible animal group nouns from a listing animal names in Japanese. Non-native speakers may find some poetic flavor of English from it while discovering the rich variety of expressions in this language.

back to research page

back to HOME

* Learning Collocation: An Introduction to Online Corpora for College English Classes

English Language and Literature (Chuo University; Tokyo) 44 (2004/February) pp. 175-194

A corpus is a collection of language material, and corpus linguistics is the study of language as expressed in samples (corpora) or texts in the real world. These collections are recently used by teachers and learners to aid language learning (i.e. a form of CALL). By introducing the case study of a semantic analysis of the numerative phrase "a piece of", this paper attempts to demonstrate college teachers of English how they can manage an English class using an online corpus.

To use a corpus, we have to access it by using a search engine (a program which generally runs through the text). Although there are numerous types of corpus systems available on the Internet, we choose the use of the free sample access to the COBUILD Bank of English (BOE) corpus named "The Collins WordbanksOnline English" boasting a rich source of data. It enables learners to get information on words collocating with the targeted word or phrase with easy handling on a web browser.

The majority of Japanese college students find it difficult to measure out uncountable nouns using appropriate numerative phrases, such as "a piece of paper", "a chunk of chocolate", "a block of ice", as they have no distinction between countable and uncountable nouns in their native language. Especially, they cannot figure out what kind of nouns can be counted by the phrase "a piece of". To let them know its appropriate collocation, we suggest using an online corpus in classroom.

Once you access to The Collins WordbanksOnline English and search the phrase, you can instantly get more than 800 hits. Having a careful analysis of them, students may find that the usage of this phrase can be organized into four types: counting a part of something that has been separated, measuring uncountable abstract nouns, counting a single thins that is part of a set of things, and counting or numbering artwork and academic work.

Online corpus systems are a good means to carry out projects in language description in college English classes. It gives non-native students an opportunity to learn the flavour of the language by searching words and phrases among the huge collection of real texts in English.

back to research page

back to HOME

* Improving English Skills and Information Search Techniques in the Internet Age

English Language and Literature (Chuo University; Tokyo) 43 (2003/February) pp. 165-188

Computers are becoming increasingly available especially in Japanese private universities, enabling the students to activate the Internet for their further studies. At the same time, new demands are being made on English education not only to read online passages and to write e-mail messages in English, but also to search and integrate proper pieces of online information. As it is reported that more than 70% of the communication via the Internet takes part in English, it makes college teachers of English intrigue to discover how to master the Internet for the benefit of themselves and their students.

By introducing a case study of "Internet English" given to sophomores in regular English classes at Chuo University from 2001 to 2002, this paper makes its attempt to encourage college teaches of English who are interested in the Internet but do not have enough confidence to manage an English class with computer equipments.

The class of "Internet English" was designed to make students understand what is the Internet, and be familiar with browsing international Web pages on the Net. In the first half of each class, students are required to enrich their basic knowledge about the Net by reading several offline materials, such as the system of the online community, the history of its development, the languages used on the Internet and Netiquette. Holding a class in a room fully equipped with personal lap-top computers, then, an online task is announced on the Web page in the latter half of the class every week. Students spend the rest of their time working on it. They complete it by submitting a brief report on the searching results to the instructor by e-mail.

These activities give students an opportunity to activate their English expressions and searching skills with having a great excitement to discover new pieces of information of the unknown world. Consequently, most of the students of the class have learned the basic skills to use the Internet in English and the proper techniques to search and summarize information on the Web in English.

back to research page

back to HOME

* Composing E-mail in English: A Case Study of Classes given to First and Second-year University Students

English Language and Literature (Chuo University; Tokyo) 42 (2002/February) pp. 251-270

Research published in 2000 shows that about 380 million people are using the Internet and more than 80% of the communication via the Internet is in English. While many Japanese university students send and receive e-mail in Japanese, it was found that less than 10% of the freshmen in this sample had exchanged English e-mail on the Net. Most Japanese students are missing out on most of what the Net has to offer ─ in order for them to become full-members of the on-line community they must learn to use English to communicate by e-mail.

This paper reports our class activities in 1st and 2nd year regular English classes during the spring semester, 2001. The classes were designed to make students familiar with the writing style of e-mail in English and to express their own ideas in e-mail.
In the first week of each unit, students were given a sample e-mail or web page based on a topic, carefully selected for undergraduate students. Students learnt useful expressions from these samples. Then, in the second week of each unit, students were asked to write replies (or ask for more information) on sheets which were marked and returned the following week.
It was found that studying Internet English not only helps prepare students for communicating on-line, but it is also an effective means helping students to accustom themselves to expressing their ideas in English.

back to research page

back to HOME

* Semantic Structure of Japanese Classifiers for Counting Animate/Inanimate Entities: how do we classify "robot dogs"?　

TULIP(Tokyo University; Tokyo)20(2001/August)

This paper intends to clarify the semantic structure of the classifiers "-nin(人)", "-wa(羽)", "-tou(頭)" and "-hiki(匹)" in counting animate beings in Japanese. We have examined the sentences cited from a database (January, 1990-March, 2001) of four major Japanese newspapers and several popular TV programs, in which these classifiers for counting animate beings, as well as other classifiers for counting inanimate objects such as "-tai(体)", "-dai(台)" and "-ko(個)" are found.

First, the semantic features of each classifier for counting animate beings are shown. It is to be noted that the classifier "-hiki" has two levels of usage: "-hiki1" for counting unspecified non-human beings, and "-hiki2" for counting non-human, non-avian beings without salient features in physique, social function or significance.

Next, we discuss on some recent examples which come between animate/inanimate categories, such as "robot dogs (AIBO)". Robot dogs, first marketed in June 1999, have been used with various types of classifiers. In the first stage, they were counted with "-dai", which is for classifying mechanic objects. Then gradually, some articles started adopting animate classifiers "-hiki" and "-tou" to emphasize the high degree of animacy of robot dogs. In the latest stage, they are counted with the classifier "-ko", one of the most unmarked classifiers for counting objects.

Finally, this paper claims that the concept of animacy and its degree take precedence of biological classification, and they play the most important role when a speaker decides which classifier to apply in counting an entity. If one feels a high degree of animacy in a certain object, one can count it with animate classifiers; on the other hand, when one does not find animacy in what one is counting, there is no way to use animate classifiers.

back to research page

back to HOME

* Classifiers for Counting Actions and Events: Comparison between "kai" and "do"
　

TULIP(Tokyo University; Tokyo)18 (1999/August)

Japanese has two major classifiers for counting actions and events: "kai" and "do". These two classifiers are interchangeable in most cases, but not always. In this paper, we compare the similarities and differences of the use of these classifiers, with reference to examples and figures found in our database. This paper has proved that Japanese classifiers are not only for classifying nouns with concrete referents, but also for classifying abstract entities. To support our idea, we have seen that the major classifiers for counting actions and events in Japanese, "kai" and "do", reflect the classification of the types of action and events being counted; whether by they are held regularly, could be repeated, could be carried on successively, how frequent it takes place in a certain period of time, and whether it is possible to tell the next occurrence of them.

back to research page

back to HOME

* Modification of Numeral Quantifiers: "Feed three cats a fish" and "Feed a cat three fish"

Colloquia (Keio University; Tokyo) (1998/October)

This study explores the priority of the semantic agreement between NQs and counted NPs in quantifier floating systems over the syntactic rules which are claimed in previous studies. The Japanese sentence "Taro-wa san-biki neko-ni sakana-o yatta" could be interpreted as "Taro fed a cat three fish." and "Taro fed three cats a fish.", and the reason of this ambiguity cannot be explained with the previous analysis of QF(quantifier floating). Referring to our questionnaire collected among 46 native speakers of Japanese, this paper shows that the syntactic conditions play an important role when the NQ agrees with a single NP in the sentence. We claim that the system of QF in Japanese cannot be explained solely from some aspects of sentence structure, rather, it is necessary to incorporate the idea of semantic agreement between NQs and NPs. Accordingly, an alternative explanation must be proposed. The examples analyzed in this paper suggest that semantic agreement between the classifier of the NQ and the NP being counted is prior to the syntactic rule QF. When a speaker encounters a sentence with an NQ, he or she may refer to the agreement between the classifier and the NPs as candidates in the first stage, and then, if he or she is sure that the only NP is focused as a candidate for agreement, may apply the syntactic rule for QF. But the sentence given may not always have a single NP that agrees with the NQ. In this case, we may say that the NP which directly follows the NQ tends to be modified if it agrees with that NQ.

back to research page

back to HOME

* Features and Functions of "-tsu" in Japanese: comparison with the major classifiers "-ko" and "-hon"　

This paper intends to clarify the semantic features and functions of the morpheme "-tsu" used in counting things in Japanese.
"-Tsu" is attached only to numerals one to nine as in "hito (ONE)-tsu", "futa (TWO)-tsu", "mit (THREE)-tsu", and so on. In some studies, it is considered as a neutral classifier having no positive semantic feature for classification, and in others, as a mere default classifier lacking a semantic core, used only in the absence of proper classifiers. First, sentences which contain "-tsu" and the major classifiers "-ko" and "-hon" are collected from a database of major Japanese newspapers (1990-1998). Then, minimal sentence pairs are arranged to see in what case "-ko" and "-hon" are replaceable by "-tsu". The fact is found that all the abstract nouns and nouns in metaphoric use are counted with "-tsu", but not with "-ko" or "-hon". In other words, "-tsu" has the function of classifying specific types of nouns. In conclusion, this paper claims that "-tsu" is not a default or neutral classifier, but that it should be treated as one of those special classifiers which have a restriction on the numerals to which they are attached.

back to research page

back to HOME

* Semantic Structures of Classifiers: The Meaning Chain of The Japanese Classifier hon'

This paper has considered the following issue: whether or not a single classifier depicts a single prominent property as a meaning core shared with all the whole entities being counted with hon.

There had been a sort of tacit agreement in the synchronical studies of classifiers that the entities classified with a single classifier share a common semantic property either directly or extensively. However, we show that this idea of 'every entity classified with hon has something to do with the property of length' can not cover the wide range of variety of the "hon-entities". The semantic properties which decide the modern use of the classifier hon are in a form of meaning chains, not in a concentric circle. Classifiers are not a mere morpheme or a marker to add or emphasize the prominent characteristics of an entity being counted, but form their category in a form of meaning chains, as well as other parts of speech in natural languages.

back to research page

back to HOME

* Prototypes of Full-Defined Categories

Colloquia (Keio University; Tokyo) (1997/October)

This paper gives an account of how prototypes constitute a full-defined category, and how a man makes use of it to expedite his categorization. We clarify these points referring to the problem of whether a category is formed with the notion of resemblance with a concrete prototypical member of a category or extracted with an abstract schema.

We discuss the prototype theory of mental categories in light of our experimental findings of full-defined categories in three types.

The well-defined categories of infinite members such as "odd number" and "even number" are constructed based on their definition, and the members involved in this definitions easily become prototypes. To define "odd number" category, it is necessary to adopt the number 2, which is also a prototype of this category.

In the categories which consist of circulation-free, finite members, such as "blood group" and "professional baseball team", all of the category members are involved in the definition of them. Therefore, prototypes are not always necessary to form categories in these cases. Moreover, there is room to yield hierarchy of membership by means of speakers' subjectivity like 'my blood group' and 'my favorite baseball team'.　

Lastly, in categories with circulation, finite members, such as "days of the week" and "months of the year", one often conceptualize categories as linear structure of ordered members with having a clue in the starting point and in the endpoint of the circulation. All of the members are involved in the definition of the category. In this case, the members at the starting point and at the endpoint often become prototypes of that category.

Consequently, prototypes are not always functioning as basic cognitive cores in constructing categories. Especially in full-defined categories, it was clarified that prototypes do not play central roles to form the categories. Full-defined categories with finite members are defined with giving the holistic lists of the members, while categories with numerous or infinite members are defined with the categorical definition, not with the lists of category members.

back to research page

back to HOME

* Aspect and Classifiers: A Study of Japanese Classifiers for Counting Correspondence

Colloquia (Keio University; Tokyo) (1996/17)

This paper discusses the use of the Japanese classifiers, mainly on hon, kai, ken and tsuu, for daily mutual correspondence referring to the idea of three categories of aspect proposed in Vendler(1967). Based on the idea that categorization and classification are in the different systems, we examine five measures of modern correspondence; the telephone, telephone answering machine, facsimile, the e-mails and the snail-mails. The result leads us to the conclusion that aspect of verbs and selection of classifiers are closely related. When a sender tries to correspond with a recipient, that activity is classified by kai both in the telephone and in the mail correspondence. The classifier hon for counting long objects is applied to the accomplishment of that activity. We concern that this is because one conceptualizes the telephone being connected or the letters being delivered are on the line of tidings. This use of hon could be sometimes found in the mail correspondence, but tsuu is usually adopted. Finally, when a sender has satisfied with telling one's messages or done one's businesses to the recipient, the focus is given on the achievement of that action, which leads to the use of the classifier ken.

back to research page

back to HOME

* Classification and Categorization: Semantic Properties of Japanese Classifier "hon"

Tokyo University Linguistics Papers (TULIP) 1996/15 pp.113-41.

This paper examines the use of the Japanese numeral classifier "hon", which is prototypically for counting long objects.
Referring to the results of our experiment, we have found that the fundamental semantic property on which all of the members of the category "hon" are based is LONG, ONE-DIMENSIONAL, and there are some other properties being shared with the members. To extract them, we analyze many fuzzy NPs from multiple aspects, and suggest that there are at least three properties which help to give decision on the application of "hon" to fuzzy NPs.

(i) inanimacy: The classifier "hon" cannot be attached to animate entities though they are long enough to be counted with "hon". Thus, entities talked about should be inanimate, or should not show a clear vestige of animacy. The acceptability of "hon" is weakened in proportion to the strength of the property of animacy.

(ii) one-dimensionality: A container which is relatively long could be counted with "hon" if it is sealed or has a trace of being sealed regardless of fluidity or amount of contents. A seal of a container gives an image that the container is an one-dimensional object, and make a speaker to use "hon" as the most favorable classifier.

(iii) constancy: When one counts objects with rolled parts functionally working, the classifier "hon" is attached only to the objects whose rolled parts do not accompany any consumption. This means that "hon" requires rolled objects to be constant in length.
We may, therefore, conclude that the fuzzy NPs should satisfy the following semantic properties to be referred to with the classifier "hon": INANIMATE, ONE-DIMENSIONAL, LONG and CONSTANT.