Abstract
Although reputation analyses have been developed by utilizing NLP technology, such studies focus on texts on the Web. Critiques of humanities are not seen as main target of those analyses. The purpose of this research is to found a basis of profound semantic analysis of critique. It will be realized by applying numerical analysis to objects of critiques of humanities. This paper focused on names of person and titles. Categorization and co-occurrence analysis were executed about that "who" and "what title" are frequently mentioned in critiques of movie and theater. As a result, critiques of theater are concentric but those of movie are individual. Also as for critiques of theater, the consciousness of art is strong.
