This paper shows how the conclusive contribution of Christian Leclère to the working out of the concept of "restructuring" throughout his work on "standard" vs "cross" structures then his work on the "ethic dative" gave way to an analysis of the formal properties of support structures. It also shows how these works gave the author while preparing his Ph D (1979), the idea of a relation between three groupes of communication verbs and two groups of movement verbs (Gross 1975) which enabled him, as from 1994, to work out his "Defining Matrix Analysis" model.