[View this page in Japanese] [View this page in Korean]

Learner's Conversion Dictionaries for Japanese Input Methods

Kazumi Hatasa (khatasa@purdue.edu) & Wonchang Chung

Department of Foreign Languages and Literatures, Purdue University

Due the increasing populatrity of the Internet in foreign language instruction, opportunities for learners of Japanese to input Japanese at the keyboard are on the rise. However, a piece of software which handles Japanese input, called an input method editor or IME, is designed for native speakrs of Japanese. When non-native learners of Japanese inputs Japanese using the same IME, various problems become apparen such as misspelling and misselection. (Tsuchiya 2000, Kubota 1999, Hatasa 2001) One method to solve these problems is to develop user dictionaries which contain common errorneous entries and incoorperate them into an IME. (Tsuchiya 2000) In this project, we have developed such dictionaries for intermediate learners of Japanese. We are rerealsing β-version for feedback.

There are six dictionaries altogether. The 1000 series contain words which are consist of most frequent 1000 kanji (approximately, kyoiku kanji) while the 2000 series montain words which are consist of the most frequent 2000 kanji. (joyo kanji) In eitehr series, Minus-Dict contains entries which have missing kana. (e.g. "ぎんこ" is registered as a reading of 銀行 or "おな" is registered as a reading of 女) Voicing-Dict contains errors on voicing. (e.g. "ちが" is registered as a reading of 地下) Plus-Dict contains words which have extraneous long vowels. (e.g. "じしょう" is registered as a reading of 辞書)


A user needs to open necessary dictionaries to handle errorneous inputs. This is done in 環境設定パレット in kotoeri, and in Micorsoft IME-2000 property page of Windows 98/2000.


Kotoeri setting (環境設定パレット)

MS-IME property

The following illustrates how conversion possibilities are presented on the screen when "かくせい" is typed.


Example conversion screen with ことえり and MS-IME


ATOK users

ATOK is a commercial IME developed and sold by Just System. It is available for both Windows and Macintosh. We have packaged a set for ATOK13 for Windows.

ATOK allows users to generate their own main dictionaries. Two main dictionaries for 1000-kanji level and 2000-kanji level have been generated. (i.e. 1000-Main-Dict and 2000-Main-Dict) For example, when you use 1000-Main-Dict, instead of the default main dictionary, no kanji beyond 1000-level will ever appear in the conversion process. Hence, the number of conversion possibilities becomes much smaller and manageable.

ATOK13 Property

Example converstio screen with ATOK13 (input: かくせい)

Since there are different versions of ATOK is in use both on Windows and Macintosh, you may have to recompile dictionaries using Dictionary Utility Program which comes with the version of ATOK you are using.


Learner's conversion dictionaries for ATOK13 (Windows) 5 MB



土屋順一 (2000) 「外国人学習者の日本語ワープロ誤入力の分析と外国人用漢字変換辞書の開発」文部省科学研究報告書 課題番号09680296

Hatasa, K. (2001) "Study of Japanese Input Method for Learners of Japanese" 35th Annual Convention of ACTFL. Washington D.C.

Kubota, R. (1999) "Word Processing and WWW Projcts in a College Japanese Language Class." Foreign Language Annals, 32(2), 205-218.

Last updated: 9/13/2002 by Kazumi Hatasa (khatasa@purdue.edu)

This project is suppported by Technology-Enhanced Language Learning Center of Department of Foreign Languages and Literatures, Purdue University.