Abstract
The purpose of the present study is to compare the cultures and social systems behind languages, based on probabilistic knowledge structures constructed from large scale language data in Japanese and Chinese. The present study applies the probabilistic algorithm developed by Kameya & Sato, more strict than the existing method like LSA (Latent Semantic Analysis). Practically, we compares interesting classes which have same names between Japanese and Chinese as the result of interpretation, selected from the latent classes estimated using the above algorithm. It is an important future plan to construct the same knowledge structures using other foreign languages like English.
