Development of multimedia collocation information search system: the use of graphic/video information for foreign language learning

Lee Sangmok
Tohoku University

This paper adopts a broader definition of collocation and reports on an attempt of developing a multimedia collocation retrieval system for the learners of Japanese as a foreign language, based on the mass supply of multimedia contents currently available on the web. Our definition of multimedia collocation is based on the text data combined with speech and video information recorded synchronous to the text data; namely, collocation in the multimedia context. We argue that this broader definition entails regarding as collocation the entire setting for the use of a particular expression that not only includes lexical and semantic usage but also covers the context where it is actually uttered. This kind of setting is arguably where an expression actually occurs in real-time with audio-visual confirmation of articulation, intonation, facial expression, and other paralinguistic information.

It is obvious and well understood among language teachers that the use of a large scale corpus specifically facilitates learners' effort to understand the collocational meaning of expressions that are more than the combination of each of the consisting words. Definition of collocation in a narrower sense is g a word or phrase which is frequently used with another word or phrase, in a way that sounds correct to people who have spoken the language all their lives, but might not be expected from the meaning h. Learning collocation entails not only acquisition of eright/wrong' judgment but also of naturalness for a given phrase as well. There is a two-sided difficulty for learning collocation and acquiring naturalness of the native speaker. The native speaker does have intuition for his/her judgment but, as regards systematic description for a given expression, he/she may not be persuasive to the learner. On the other hand, due to the limitation of time it is totally impossible for the learner to be exposed to the entire body of the language data. In our efforts in making use of the present-day internet multimedia environment in language education we developed an on-demand web-based dictation system for learners of Japanese mainly targeted to Korean students.

we developed a prototype of multimedia collocation information retrieval system with a limited vocabulary but with the following potential advantages for learners and teachers of Japanese:

•  Presentation of collocation is available in the speech-video context.

•  Access to multimedia collocation data activates learners' memory and facilitates learning process.

•  Relational database employed to renew and maintain the data is easy to use and reliable for the teacher and/or system administrator.

This paper starts with the extended concept of multimedia collocation, and reports on the development and evaluation of collocation retrieval system for learners of Japanese in an on-demand real-time server-client environment. We confirmed that the use of multimedia collocation database system promotes understanding and memorization of words for learners of Japanese, and that the present retrieval system enabling users to search the video-speech-text multi-channel collocation database will be a powerful tool for language learning.