researchmap

researchmap

Grant amount：\44590000 （ Direct Cost: \34300000 、 Indirect Cost：\10290000 ）

We improved the system of the Corpus of Historical Japanese at NINJAL, to handle the original text significantly different from the transliteration text, and to annotate multiple readings such as kakekotoba, paronomasia and temporary furigana. Then, in cooperation with NINJAL "Diachronic corpus" project, "Manyoshu" with original text, Christian materials written in both Roman characters and Japanese, Sharebon and Ninjobon with multiple readings, "Hachidai-shu" with kakekotoba, corpus of "Toyo Gakugei Zasshi" and colloquial materials in the early Meiji period were prepared and made available to the public through "Chunagon" on the Web.
In addition, we developed studies of Japanese language history using this corpus, and in particular, we published the research in the fields of grammar of Old and Early Middle and Early Modern Japanese, and vocabulary of Modern Japanese.

researchmap

Grant amount：\2730000 （ Direct Cost: \2100000 、 Indirect Cost：\630000 ）

The purpose of this research is to extend the conventional version of the Hachidaishu thesaurus (ca.905-1205) to include the Nijuichidaishu (ca.905-1429). At the same time, through its development, we attempt to develop a theory of the term space analysis (vocabulary system change) encompassing 534 years of ancient words based on classical Japanese poetry. Thesaurus data input was completed for all of the Nijuichidaishu, and the thesaurus code system and its vocabulary index system were completed. As a result of checking whether the description of the word in the thesaurus system is unified or not, we found that the system of the thesaurus itself is consistent with modern language. However, there exist contradictions when utilizing the distance calculation between ancient words.

researchmap

Grant amount：\4940000 （ Direct Cost: \3800000 、 Indirect Cost：\1140000 ）

In order to promote the use of thesaurus in the Japanese linguistics, we annotated the useful information to the "Bunrui Goihyo (Word List by Semantic Principle, revised and enlarged edition)" which is the most popular thesaurus among Japanese linguists. Based on our original criteria, we chose the "representative meaning" of the 27,171 polysemic words which appears more than one semantic cluster in Bunrui Goihyo.The annotated thesaurus is due to be released in July 2015 on the web. Using this thesaurus, it will be possible to improve the precision of semantic analysis and make appropriate comparison between different analyses. We also examined the differences between old and new versions of Bunrui Goihyo. The result was reported in the Journal of the Study of Vocabulary Vol.12.

researchmap

Grant amount：\2990000 （ Direct Cost: \2300000 、 Indirect Cost：\690000 ）

The principal investigator has previously developed a tool for the morphological processing of waka poems in 2007. However, its range of applicability was limited to the Hachidaishu. The goal of the present research is to automatically segment and annotate part-of-speech tags for the Nijuichidaishu using the previously annotated segmentation data and token adjacency probabilities of the Hachidaishu. Using the KyTea (Kyoto Text Analysis Toolkit) morpheme segmentation toolkit, with its default L2 regularized SVM learning algorithm, model learning took less than a minute. This model also achieved a high segmentation accuracy of around 96% on the Nijuichidaishu. While there is some remaining work to be done concerning the addition of unknown tokens and the learning of adjacency probabilities around unknown words, the development of a dictionary that can segment the Nijuichidaishu with a high accuracy can be considered complete.

researchmap

Grant amount：\2820000 （ Direct Cost: \2400000 、 Indirect Cost：\420000 ）

The goal of this research project was to construct Natsume a Japanese language writing support system. Firstly we proposed an effective method based on collocation search. A linguistic analysis essential to international students was made based on Japanese language corpora. Additionally, we constructed Natsume equipped with automatically correct learners writing system by using a combination of the Natane learner corpus and native Japanese language corpora.

researchmap

Grant amount：\5200000 （ Direct Cost: \5200000 ）

本研究は,直接的に古典テキストをデータベース化したり,分析することではなく,そのための基礎的なソフトウェアの開発,Web公開のための手法開発である。すなわち,古典研究者が古典テキストに対して情報処理を試みるのに必要な基本的ソフトウェアの整備を目指したものである。すなわち,古典テキストのデジタル化,データベース化を支援するためのソフトウェアの開発,及びデータベースをWeb上で公開するための新しい手法「BB-DB」を開発し,その実現を図った。「BB-DB」はBare Born Databaseの略で,従来のDBMS(Database Management System)をまったく利用しないで,同等の機能,性能を実現したものである。具体的には以下のような項目について研究開発を進めた。
1)古典テキスト原典のPDF化に関するフィージビリティ研究
2)データベース構築・利用支援ソフトウェアの開発
3)日本語・ギリシャ語混在表記ツールの開発
4)BB-DB手法の開発及びその実現
5)古今集のパラレル・テキストデータベースの作成

researchmap

Grant amount：\8500000 （ Direct Cost: \8500000 ）

The purpose of this research is developing and distribution of general teaching materials parts for supporting Japanese language education. This research advances in the enforcement plan for four years.
The summary of our research project are as follows.
1. Our research project developed the model of general teaching materials parts based on the database for teaching and learning Japanese. We used some basic electric dictionaries, dictionary database of kanji, hitsujun database of kanji, basip word database, usage database, Japanese-English dictionary, academic term database phonetic database of basic Japanese and visual database of basic Japanese, for example.
2. Our research project developed and evaluated several application tools which process general teaching materials parts. At first we developed multimedia tools, input/output tool, retrieval tool and link tool, for example. Next we developed multilingual tools which treat Japanese, English, Spanish, Portuguese, German and Italian.
3. To improve and evaluate the database, our research project invited the foreignresearchers and we went to overseas to meet the foreign researchers.

researchmap

Grant amount：\8000000 （ Direct Cost: \8000000 ）

The purpose of this research is deveioping the CASTEL/J database for Japanese language education into foreign country languages. This research advances in the enforcement plan for four years. and this year is the last year. Our CASTEL/J project had developed the database for teaching and learning Japanese applied multimedia technology. We developed (1) Dictionary database of Kanjj, (2) Hitsujun database of Kanjj, (3) Basic word database, (4) Usage database, (5) Academic term database, (6) Japanese-English dictionary, (7)Text database for readings, (8)Phonetic data of basic Japanese and (9) Visual data of basic Japanese. Dividing the above database included in th~~ CASTEL/J CD-ROM. We were able to develop the basic word dictionary into many languages by this research. Specifically, we improved the CASTEL/J database which consists of Japanese and English into Spanish. Portuguese, German, and Italian. Finishing this research, we are sure that the CASTEL/J database can be used now for more Japanese teachers and students. But, If we evaluate the developed system, how to spell a basic word cannot call it completeness yet at present. This is a future subject.

researchmap

Grant amount：\9800000 （ Direct Cost: \9800000 ）

Recently, both in quality and in quantity of the increase of the Japanese learners are very remarkable. Therefore, many tests have been done for finding the solution of each of the problems that follows, and our researching group also have been developing the Japanese educational supporting system by using personal computer.
Our project had developed the database for teaching and learning Japanese applied multimedia technology for two years.
We developed (1) Dictionary database of Kanji, (2) Hitsujun database of Kanji, (3) Basic word database, (4) Usage database, (5) Academic term database, (6) Japanese-English dictionary, (7) Text database for readings, (8) Phonetic data of basic Japanese and (9) Visual data of basic Japanese.
Dividing the above database included in the CD-ROM. The copyrights of these databases are cleared for the limited usage in education and in research, and are planned to be distributed as much as possible.
Several members of this project had visited the countries to investigate their needs for Japanese education and abilities to use this database system.

researchmap

Grant amount：\2000000 （ Direct Cost: \2000000 ）

本研究では、特に漢字読解教育に焦点を絞り、CAIを開発するために必要な日本語教材データを開発することである。漢字・読解教材をデータベース化するだけでなく、それを利用する学生の学力の測定や診断を行い、適切な教材提供までを考慮するシステムのためのデータベース開発も視野にいれて、将来的には統合化を目指せるようなデータベース開発を行った。具体的には、中級段階の個別教材が容易に作成できるようにする、興味に応じて、個別教材を選べるようにする、学力を診断した上で、適切な教材を選べるようにする、などの教育活動を支援するのに十分な情報のデータベース化が今までにない、実践的な教育システムのためのデータベースであろうと考える。
本研究においては、19種のデータベースを開発し、教材開発支援、評価支援、教育支援に役立つデータを管理できるようになった。今後の課題として、データは大規模であってもその構造に一貫性を持ち、拡張性、汎用性において頑健であること、数学的に和および差が意味をなす集合として解釈できること、人間が行う実務に即していること、を追求し、データ構造に反映する作業が不可欠となる。
日本語教育においては、1.言語そのものが持つ性質のデータ:文法、漢字、表現など、2.教育そのものが持つ性質のデータ:指導順序、教材、テストなど、3.学生そのものが持つ性質のデータ:難易、習得順序など、の領域固有のデータを取り扱う必要があろうが、現段階では、上記1.のみの研究に終っており、2.の一部はテストデータベースの形で実現しているが、それ以外は、試みが個々の発表論文の形式で終っている。これら論文で提案されている方法をデータベースとして何を用意すればよいかを検討し、実務に移す必要があろう。

researchmap

Grant amount：\1900000 （ Direct Cost: \1900000 ）

本研究は,研究代表者を始め国立教育研究所が中心になって,全体の企画調整と現有データベースの機能拡充を行った。そして,筑波大学・関西外国語大学・名古屋大学の研究分担者が海外関係機関を担当し,東北大学・国際基督教大学の研究分担者が国内関係機関を担当し,国際研究集会の企画調査を行った。具体的には,(1)CASTEL/Jに興味関心のある研究者・教育者を把握し,(2)新規開発すべき電子化教材データベースの要望の取りまとめと,(3)現有データベースの改良すべき点を明らかにした。なお,鳴門教育大学の研究分担者は,国際研究集会においてCASTEL/Jを紹介するためのプレゼンテーション用のデモンストレーションソフトウェアの開発を行った。
本研究を遂行するために,約1カ月に1回,研究連絡会議を開催し,第2回「日本語教育とコンピュータ」国際会議の準備を行った。また,日本語教育とコンピュータに携わる研究者・教育者のネットワークを作るために,ニューズレターの発行を行った。そして,研究代表者はイタリアと台湾に出張し,海外の日本語教育研究者と今後の研究打ち合わせを行ってきた。
今後,第2回「日本語教育とコンピュータ」国際会議を計画通りに進め,実施することが責務であるが,まだ,(1)世界各国から参加者人数の把握や,(2)会議開催のための諸費用の捻出等の課題が残されている。

researchmap

Grant amount：\2300000 （ Direct Cost: \2300000 ）

日本語の漢字・読解教育をカリキュラムとして有機的な関係として捉え、さらに文法的情報とも関連づけて、学習を円滑に進めるための支援となる教育データベースのデータ構造を検討した。漢字診断テストのデータベース、漢字学習情報データベース、簡易文法テスト、文法問題データベース、読解練習データベースという5つのデータベースを統合し、関係づけるための相互関係テーブルを作成し、学習者が自分の漢字学習の進度に合わせて読解教材を選んで学習したり、また学習しようとする読解教材において使用頻度の高い語彙や漢字を順に学習したりできるように、教育データ同士を円滑にやりとりできるデータ構造について試案を作成した。それぞれのデータベースを充実させるとともに、実際に学習者に試用してもらう際のフィードバックの形式についても日本語教育機関から情報を集めている。

researchmap

Grant amount：\5200000 （ Direct Cost: \5200000 ）

SPOT (Simple Performance-Oriented Test) has been originally devised by Noriko Kobayashi and Junko Ford-Niwa as a new device that allows JSL(Japanese as a second language) learners at various levels to be placed simply, swiftly and accurately. The results of the research suggest that SPOT is a sufficiently accurate means to place students.
SPOT consists of an audio taped recording of a native speaker reading a series of complete sentences, to which the testee listens while reading the same sentences in Japanese script on the answer sheet. Each sentence on the answer sheet contains a blank which the testee must fill in based on the tape. Each blank consists of a single hiragana character which represents all or part of a grammatical item.
This research project has elucidated the following.
1 )The results of trials at several institutions under different circumstances have been analyzed and the results show the score of the SPOT to be highly reliable. (The numbers of the co-operative institutions are more than 50, including not only those in Japan but also in USA, Australia, Korea, Indonesia, etc.)
2) SPOT does not seem suitable for testing each discrete point of language knowledge, or the proficiency of each language skill such as reading, listening, writing and speaking, but is effective for testing integrated proficiency or processing ability in natural speed.
3)The phonetic variations (duration, pitch contour, devoicing or weakening of vowels or consonants) of the sentences read in the SPOT have a strong relation with the score.
4)The research suggests that SPOT ranks Japanese learners by measuring their automatic processing ability.
5) SPOT would be a good instrument to investigate the process and outcome of grammar acquisition by Japanese Language learners.

researchmap

Grant amount：\6800000 （ Direct Cost: \6800000 ）

The current Project follows on from a previous project, '"Development of a Learning Support System for Measuring Foreign Students' Kanji Learning Stategies" (Grant-in-aid for scientific research (C), 1994-1995, Chief Investigator : Stefan Kaiser, Project Number : 06808026.
In the current Project, about 400 new kanji were added to the Kanji Information Database, making it a total of about 900 kanji. New fields for selection by the user include information on frequency, based on Newspaper frequency counts, str_ans (information on kanji structure), ready-made story (a story or combination of keywords to be used by the used as mnemonics), and animated sequence of stroke order. English is used as an indermediating language. The Kanji Information Database emplys a faily large number of fields which the user can choose to provide information assisting the learning of kanji by clicking, providing in return a record of how the user approached the learning of a kanji.
The user interface for using the Kanji Information Database on the internet has been designed, and as soon as the Database has undergone final editing, it will be made available on the World Wide Web. Used feedback will be utilized to improve the system to a more user-friendly one in the future.

researchmap

Grant amount：\1500000 （ Direct Cost: \1500000 ）

1「非漢字圏学習者のための漢字学習情報データベース」非漢字圏日本語学習者支援システムを開発する目的で、このデータベースは開発された。最終的な目的は、システムを使用する学習者がどういう機能をどの程度利用するか記録し、データを収集し、そのデータの分析に基づいて漢字学習に有効な学習・指導方法があるかを検討することである。学習者によってはシステムで用意されているデータは不十分であったり、デザイン上漏れてしまったり、他の機能が欲しかったりする可能性があり、そのためにはユーザ側のコメントをできる限りとりこむようなデータのフィールドをあらかじめ用意しておいた。
2論文読解のための表現文型データベース
科学技術の読解支援システムに利用するためのデータベースを開発した。本研究ではこのデータベースを利用して、電子化された専門文献から問題点を指摘する,問題を考察する観点を示す,問題を後回しにする,問題の考察を始める,つなぐ,反対の趣旨に導く,などの機能を含む日本語文を自動抽出し、単文、複文、名詞修飾節を持つ文、またその複文のように文の複雑さ、文の長さでソーテイングを行い、単文読解訓練用の文章を作成するシステムを作った。
3漢字学習方法処方のための漢字力診断問題データベース
本研究におけるデータベースは漢字の診断に用いるのに役立つ問題のデータベースとその構造の開発である。データの形式は従来は1問題1レコードの形式を採用していたが、SGML化をおこない、テストシステムとして用いる時には、HTMLにコンバートし、プリントテストとして用いる時にはLaTeXとしてきれいに清書されて出力できるようにした。現在はNetscape2.0などのWWWのプラウザでアクセスできるようになっている。

researchmap

Grant amount：\1100000 （ Direct Cost: \1100000 ）

筑波大学,九州大学,東京学芸大学等において初級・中級の漢字教育,漢字テストに関する調査を行い,外国人学習者の学習項目,学習方法,評価項目などについて情報を収集した。また,一般に広く普及しているテストとして日本語能力試験の文字問題,外国人のための漢字教材および日本人の子ども向けの漢字教材のいくつかを検討し,合わせて,テスト項目の分類・整理を行った。
本年度は初級漢字教科書『Basic Kanji Book』Vol.1,2のテストデータの入力を始めるに当たり,テスト項目の立て方,分類方法,データ形式に関する一応のフォーマットを作成し,45課分の小テストのデータおよびまとめのテスト4回分のデータを入力した。さらに,全国の日本語教育機関に向けて,実施している漢字テストの内容,評価項目に関するアンケート調査および聞き取り調査を行い,テストのサンプルデータを収集するための準備作業も進行中である。
今後は,テストデータの入力を増やし,中・上級レベルのテスト項目についても分類・整理を進めながら,本格的アクティブ・テスト・データベース化に向けてデータ形式,ラベル付けの方法およびデータ構造,プログラム構造のあり方を検討する必要がある。

researchmap

Grant amount：\7800000 （ Direct Cost: \7800000 ）

We have completed a database for Japanese conversion using video and sound. The number of fields in the main database is five, From this year, the format for the relational video and sound database is the widely-used MPEG ; the database comprises a total of 3269 conversational records, of which 1097 new records have been added, and is searchable. The video, which is in the form of computer files, have also in their entirety been recorded onto video. Separetely, we have also developed a stand-alone system for video CD, and have carried out a comparative survey of both systems.
In order to measure the educational effect of the database, from September 1997 we have started using the picture materials utilizing the database on the Intensive Program and Japanese Training Program in the International Student Center, University of Tsukuba, and conducted a survey by questionnaire in order to assess the users' evalutation of the video materials. Regarding the time of use of the video materials, there many users who replied "for revision" and "for preparation and revision". This show that one of the strengths of a database using video materials is that it can provide comprehensible input. From now on we need to show the learners the methodology of how to prepare for class using video materials. Regarding the purpose of use, the majority replied "for comprehension only", "for repeating" and "for dictation", and regarding the users' preference for practice materials to be provided, many replied "Pronunciation Exercises", Comprehension Quiz" and "Dictatieparation, and devise a system where the user can "repeat" as revision practice and have his input checked.

researchmap

Grant amount：\14600000 （ Direct Cost: \14600000 ）

The purpose of the present research was to create a network system for educational information to support overseas institutions of Japanese language education and conduct a survey of its use.
Some of the major problem points in Japanese langugage education are 1., the lack of information on educational manpower and materials, such as teachers, teaching materials and practical the fact that despte a strong demand for Japanese language education in the world sufficient educational provision can not be provided, and even though in many places practical examples exist, they can not be utilized elsewhere for developing the .theory of education further.
To solve these problems, we accomplished the following in the present research. 1., We made a database of Japanese. language teaching textbooks tests developed at the University of Tsukuba. 2., Created a system enabling the user to generate word indexes and kanji indexes via e-mail. 3., Made it possible to carry out tests and the collection of test data on the Web. 4., Developed video materials and made them available over the internet.
Having completed the above, we have begun to provide a service that makes it possible to use educational materials and tests from around the world over the internet ; however, the survey of how these a used in educational institutions around the world still incomplete.

researchmap

Grant amount：\8700000 （ Direct Cost: \8700000 ）

When organizing a language course, it is necessary to determine the general language proficiency level of the students and place them in the appropriate class. We have invented SPOT (Simple Performance-Oriented Test) as a new device that allows learners at various levels to be placed simply, swiftly and accurately. The results of the research suggest that SPOT is a sufficiently accurate means to place students.
SPOT consists of an audio taped recording of a native speaker reading a series of complete sentences, to which the testee listens while reading the same sentences in Japanese script on the answer sheet. Each sentence on the answer sheet contains a blank which the testee must fill in based on the tape. Each blank consists of a single hiragana character which represents all or part of a grammatical item.
This research project has elucidated the follwings.
1) The results of trials at several institutions (both in Japan and foreign countries) under different circumstances has been analyzed and the results show the score of the SPOT is highly reliable.
2) The SPOT does not seem suitable for testing each discrete point of language knowledge, but is effective for testing intergrated proficiency.
3) The phonetic variations (duration, pitch contour, devoicing or weakening of vowels or consonants) of the sentences read in the SPOT have strong relation with the score.
4) The research suggests that SPOT ranks Japanese learners by measuring their automatic processing ability.
5) The SPOT would be a good instrument to investigate the process and outcome of grammar acquisition by Japanese Language lerarners.

researchmap

Grant amount：\2000000 （ Direct Cost: \2000000 ）

1「非漢字圏学習者のための漢字学習情報データベース」非漢字圏日本語学習者の漢字学習支援システムを開発する目的で、このデータベースは開発された。最終的な目的は、システムを使用する学習者がどういう機能をどの程度利用するか記録し、データを収集し、そのデータの分析に基づいて漢字学習に有効な学習・指導方法があるかを検討することである。学習者によってはシステムで用意されているデータは不十分であり、デザイン上漏れてしまった、他の機能が欲しかったりする可能性があり、そのためにはユーザ側のコメントをできる限りとりこむようなデータのフィールドをあらかじめ用意しておいた。
2論文読解のための表現文型データベース
科学技術の読解支援システムに利用するためのデータベースを開発した。本研究ではこのデータベースを利用して、電子化された専門文献から問題点を指摘する,問題を考察する観点を示す,問題を後回しにする,問題の考察を始める,つなぐ,反対の趣旨に導く,などの機能を含む日本語文を自動抽出し、単文、複文、名詞修飾節を持つ文、またその複文のように文の複雑さ、文の長さでソーティングを行い、単文読解訓練用の文章を作成するシステムを作った。
3漢字学習方法処方のための漢字力診断問題データベース
本研究におけるデータベースは漢字の診断に用いるのに役立つ問題のデータベースとその構造の開発である。データの形式は従来は1問題1レコードの形式を採用していたが、SGML化をおこない、テストシステムとして用いる時には、HTMLにコンバートし、プリントテストとして用いる時にはLaTeXとしてきれいに清書されて出力できるようにした。現在はNetscape2.0などのWWWブラウザでアクセスできるようになっている。

researchmap

Grant amount：\900000 （ Direct Cost: \900000 ）

科学技術の読解支援システムを開発、とくに、本研究では、科学技術文献における重要表現の訓練システムの開発を行った。
開発にはサンマイクロシステム社のワークステーションSun SPARCstation 5を用いた。開発には、C、AWK、SEDなどのプログラミング言語、スクリプト言語などを使った。日本語の科学・技術論文・雑誌の記事を収集し、それぞれ機能別に例文と文型を整理し、機能、日本語、英語、翻訳のインデックスをつけたデータベースを作成した。たとえば、問題の核心や要点を述べる表現を日本語でなんというかを調べたい時はkey,glimpse,core,heartなどの用語で検索し、“これは問題の核心にふれる/これは問題を解く鍵である"という日本文を提示してくれる。これは作文にも使えるデータベースであるが、本研究ではこのデータベースを利用して、電子化された専門文献から問題点を指摘する,問題を考察する観点を示す,問題を後回しにする,問題の考察を始める,問題の核心・要点を述べる,例をあげる,理由や根拠を述べる要約する,由来を示す,本論へ導入する,分類する,付け加える,表現法を限定する表現される,反論する,反対の趣旨に導く,対照的なことへつなぐ,反対の趣旨に導く,などの機能を含む日本語文を自動抽出し、単文、複文、名詞修飾節を持つ文、またその複文のように文の複雑さ、文の長さでソーティングを行い、単文読解訓練用の文章を作成するシステムを作った。今後の課題は、このシステムが出力する単文に対する回答方法を受理し、その正誤の判定をする部分を開発すること、システムの評価があげられる。

researchmap

Grant amount：\10200000 （ Direct Cost: \10200000 ）

From 7th May to 9th May 1996, Associate Professor Atsushi Fukada from Purdue University attended the M.I.T.Summer Intersive Techinical Japanese Course and installed the CATERS (Computer-Assisted TEchnical Reading System) in the computer service system of M.I.T.The orientation concerning the use of CATERS for the students of the summer course was carried out by the director of the course, Prof.David O.Mills, at the beginning of June.
During the period from 22nd May to 20th June 1996, Prof.Fukada visited the International Student Center of the University of Tsukuba several times to attend meetings concerning the improvement of CATERS.Based on the discussions held at Tsukuba, Prof.Fukada next worked on improving the software of CATERS at Nagoya University.
During July 1996, Assistant Professor Hilofumi Yamamoto and Associate Professor Chieko Kano from the University of Tsukuba visited M.I.T.to evaluate the system. Questionnaires were conducted to assess the general impression of CATERS as well as the learning characteristics of students using the system. Instructors and students were interviewed regarding the effectiveness of CATERS and students were further given a reading test. Finally, reading classes were observed for a week in order to analyze the reading comprehension learning process of students.
In October 1996, Prof.Michio Tsutsui from the University of Washington, Prof.Keiko Unedaya form the University of Michigan, Prof.Kazue Yamamoto Zon from Johns Hopkins University and Prof.Richard Harrison form the University of Melbourne were invited to the University of Tsukuba. They and all the members of this project in Japan met together at the "Teaching Technical Japanese : Reading Instruction of Technical Japanese Documents" on the 26th of October. A meeting concerning the final report of this project was also held.
From 17th December 1996 to 12th January 1997, Prof.Kazumi Hatasa visited the University of Tsukuba and attended another meeting concerning the final report. He wrote about the improvement of the system by consulting with Prof.Fukada by E-mail.

researchmap

Grant amount：\1700000 （ Direct Cost: \1700000 ）

本研究では、「基本漢字500」の漢字すべてについて漢字情報データベース(媒介語、英語)を作成した。学習者の選択により複数のルートで学習が可能になっており、今後はカリキュラムの中で学習者に使わせることによって、学習者がどの機能をどの程度利用するか記録しデータを収集するし、データの分析に基づいて漢字学習に有効な学習・指導方法を設立する予定である。
漢字情報データベースは1字ごとまずその漢字についての基礎知識(基本漢字500番号・漢字の音・訓・中心義)が用意されている。また、いろいろな学習学習ルートがとれるように、学習者が自分で考える覚え話しが入力できるスペースを設ける。学習のためのルートとしては他には次のような項目が用意してある。
同音符同音字・同音符類音字・同音符異音字・類似形態音符・類似形態漢字・同部首字・類義字・単字例文・熟語例(単独)・熟語例(文脈中)。
さらには、学習者がコメントやメモを書くフィードバック欄が設けてある。
今後の課題として残っているのは、漢字の構造を当てる機能を盛り込むこと(新しいパターン分けはすでに行っており、形成字15パターン・会意字8パターンを考えている)である。それから、筆順情報、グラフィックス、既成の覚え話しを考える予定でいる。
さらには、漢字情報データベースは本研究では学習・指導方法設立のために使用するが、漢字学習ソフトとして十分実用化できるものと考える。

researchmap

Grant amount：\1600000 （ Direct Cost: \1600000 ）

This study attempts to analyze the difficulty of JFL listening comprehension through learners' dication errors. Dictation does not necessarily reflect aural comprehension as it is, i.e.wrong spelling does not necessarily mean misunderstanding of contents and vice versa. However, dictation written by learners is valuable as it shows some listening difficulties in a visible shape.
Six kinds of audio resources of about 4-5minutes length were given to learners so that they could write them down listening them as many times as they wanted. The number of total dication data is 173. A relational data-base of these dictation errors is expected to be useful for Japanese language teaching, though this has not been explored in this study. Instead, data is printed out so that the trends of errors can be seen.
The following issues are discussed.
1) Dictation errors spread over meaning units, or grammatical units, making database complex.
2) From the phonetic standpoint, errors are found in the combination of vowels, devoiced sounds, special mora, and r/t/k sounds in particular.
3) From the grammatical standpoint, errors are found in the particles, and other aural expressions such as toritate-joshi, sentence-final expressions that indicate aspect and mood, and modifying expressions using 'toiu', 'yoona' etc.

researchmap

Grant amount：\6000000 （ Direct Cost: \6000000 ）

The cultural distance between Japan and foreign countries is considerable, and consequently problems arise due to insufficient understanding of areas such as students' cognitive and emotional processes during Japanese language classroom instruction. A better understanding of students' cognition and teachers' decision making process are needed.
The purpose of this research is to clarify the cultural factors and to describe the model of Japanese teaching and learning process. The methodologies adopted this research is stimulate-recall procedure and questionnaires.
The following results are known in consequence of classifying and comparing student' strategies during Japanese language classroom instruction.
(1) Students with high academic achievement use many strategies to learn Japanese, but students with low academic achievement use few ones.
(2) The learning strategies which students used during classroom instruction were complicated and continuous.
(3) The cultural difference is shown using of learning strategies.
(4) The cues gathered from the students were useful in planning future lessons and improving the lesson plan.
(5) For non-Kanji group learners do not have optimistic views on their potential in reading and writing Japanese. However, many of the learners have self-confidence in studying Kanji by themselves.

researchmap

Grant amount：\800000 （ Direct Cost: \800000 ）

本研究では心理学実験による検討とSignalingのデータベース開発を行った。Signalingの有無、差異によって説明文と専門文献の構造の把握と理解の度合についてどのように異なるか、実験を行った結果、外国人の説明文読解にはSignalingそのものの効果は見られなかったが、専門文献において専門語の知識の他にSignlaing効果が有意な差で見られた。このSignalingをカテゴリ分けし、論文に使われるSignalingデータベースを開発した。科学技術Signalingのデータベースを利用し、論文のテキストを自動解析し、論文カテゴリを抽出する装置の開発の可能性が生まれた。科学技術論文のプレーンテキストを入力することによって自動的に科学技術文献Signalingデータベースからの知識を用いて文章を解析するシステムである。これを元に次の研究において専門文献読解支援システムの開発を目指すことができよう。

researchmap

Grant amount：\1100000 （ Direct Cost: \1100000 ）

In order to make a diagnosis of foreign learners' Kanji ability including their working knowledge of Kanji usage and also to use the result for making teaching and learning Kanji more efficient, we developed a Kanji test using a personal computer. We also designed a radar chart to show the result visually for learners.
This computer assisted testing system (CAT) works on Macintosh computers. Advantages of using this CAT system are as follows ; 1)learners can take the test individually, 2)the result and the feedback can be given immediately, . Taking into consideration the difference between CAT program and CAI program, we carried out a questionnaire survey for both learners and teachers and a simple experiment to improve its interface and design of the testing screen.
At the beginning of this project we planned to ask our students to use the CAT system and analyze the result to improve the cotents of the test and its format, but due to limit of time could not finish the final stage. We are now planning to use the CAT system in the Japanese language course at the International Student Center, University of Tsukuba, and use this test for more effective ways of teaching and learning Kanji to Kanji users and non-Kanji users.

researchmap

Grant amount：\8500000 （ Direct Cost: \8500000 ）

From April 26 to May 15, 1992, Associate Professor Kazumi Hatasa from Purdue University visited the University of Tsukuba and Nagoya University and completed the computer-assisted reading comprehension system called CATERS (Computer Assisted TEchnicalReading System) with Associate Professor Atsushi Fukada of Nagoya University. Hatasa took CATERS back to the United States and sent it to M.I.T. He gave an orientation concerning the use of CATERS for the students at the beginning of the M.I.T. Summer Intensive Technical Japanese Course.
From July 11, 1992, Fukada was sent to M.I.T. and took charge of the reading classes in the course. He maintained the CATERS system and sometimes sent information about the course to University of Tsukuba by E-mail. Then from July 18, 1992, Research Assistant Hirofumi Yamamoto from the University of Tsukuba was sent to M.I.T. to carry our questionnaire survey concerning the evaluation of the system and the student learning style. From July 22, Assistant Professor Chieko Kano from the University of Tsukuba was sent to M.I.T. and carried out interviews of both the students and the teachers. Reading classes were Observed for 8 days in order to analyze the student learning style. Date on the histories of study of the users were sent to the University of Tsukuba after the conclusion of the course.
On October 4 and 5, 1992, meetings were held at Nagoya University at which reports were presented on the test use of CATERS, on the results of the questionnaires and the interviews, and on the evaluation of the system.
From December 27, 1992 to January 7, 1993, Professor Otusbo of Tohoku University was sent to the United States. He had meetings with Prof. Michio Tsutsui at the University of Wshington, Prof. Noto at the University of Chicago, and Prof. Mills at the University of Pittsburgh concerning the system evaluation of CATERS.

researchmap

Grant type：Competitive

The mathematical analysis of classical Japanese vocabulary using co-occurrence patterns which are seen in the Hachidaishu (ca. 905-1205). The patterns are visualized in network/graph models.

researchmap

Grant type：Competitive

八代集に見られる用語を計量的に分析する。任意２語１組のパターンを用いて、グラフモデルを描き、可視化する。

researchmap