Abstract
How can speakers master an infinite number of possible linguistic expressions ? This question has ancient roots, but was only precisely addressed in the second half of the 20th century, with the approaches of formal linguistics. Knowledge of language can be modeled as the possession of a system of recursive rules, capable of reapplying themselves indefinitely on their own result (Chomsky), thus generating a potential infinity of structures. Several decades of discussion on the nature of linguistic rules have led to the hypothesis that the system can be reduced to a single, very simple and general recursive rule, called " Merge " (assembly) in the minimalist program. This extremely simple and elegant syntactic calculation system produces tree-like representations, the syntactic trees, which must be " readable " by sensory systems (on the side of visible sounds/gestures) and by thinking systems (the systems of concepts and intentions).
