A cognitive-conceptual dictionary, and its application in computer analyses of content

Abstract

This is a report on an operational technique for the content analysis of free flowing text, employing a conceptual dictionary. The method has been applied in numerous studies, and while it is undergoing constant revision, currently stands as a feasible approach to automatic content analysis. The dictionary is like an interlingual dictionary, translating words into a cognitive-conceptual language by substituting concepts for the words. The cognitiveconceptual language consists of 114 content categories empirically derived by grouping words related in meaning. About 25,000 words of English are so categorized. A text may be analyzed for total content or for selected categories. A recent advance in programming, which we call a "fall back" routine, permits the program to process a virtually unedited text and to match text words with dictionary words even where they differ by suffix. The procedure does this by taking the dictionary word which matches most closely to the text word.