This paper addresses, in French linguistics, three complementary aspects of “grammatical tradition”, as a concept, as a term, and as a corpus of texts aiming at describing objective properties of languages. What does mean “tradition” exactly ? Is the term of “grammatical tradition” used to refer to a scientifically deficient approach to grammar, as exemplified by the term of “traditional grammar” ? What is a “tradition” in the scientific field ? What are the criteria (historical, epistemological) which allow us to consider this corpus homogeneous ? How is this corpus constitued ? Could “grammatical traditions” be subsumed into a generic category (the “grammatical tradition”) ?