Acceso usuarios
Víctor Manuel Longa Martínez 
In this article I argue against Matthews (1981) claim about the impossibility of deriving dependencies from constitution. Based on his claim, Matthews defends that there is no strong equivalence between the corresponding theoretical frames. Contrary to that opinion, I start defending the evident connection between weak and strong equivalencies, and weak and strong generative capacities, and I point out the error that drives Matthews to his wrong hypothesis: he correctly defines the notions of weak and strong equivalence, but he applies the latter imperfectly. If strong equivalence consists not only in the generation of sequences, but also in the attribution of an structural description to each sequence, such descriptions give a knowledge of dependencial properties (at least, the most important ones) mechanically, with the only exception of the descriptions referred to sentences. Summing up, I do not explicitly affirm the existence of a strong equivalence between constitutional and derivational frameworks, but I demonstrate that in order to deny such an equivalence arguments other than those of Matthews' are required.
