French Basic Vocabulary Database, Its Microstructure and Information Description

Abstract

This paper shows the results of our work on two issues that are necessarily raised when making a French basic vocabulary database: first how to select basic vocabulary, and second how to represent grammatical informations in the database. For the first issue, we adopted the A-level vocabulary proposed by the Référentiels, but in order to revise it, we conducted frequency-based adequacy tests. 430 words are excluded because they are rarely used and thus inappropriate as basic vocabulary. Instead, we added 200 very high-frequency words that we believe should be learned at the beginner level, and we propose a basic vocabulary of 1,300 words. For the second issue, we developed a representation model with reference to the lexicon-grammar and the theory of "classe d'objet". It follows their basic framework, but it has different microstructure and representation methods so that the usage of predicate words can be understood in an integrated way.