In this paper we show that the Lexical Integrity Hypothesis (Chomsky, 1970) can be interpreted in two opposite ways. One of these interpretations, the most frequent one in the literature, bans the access of syntax to the internal structure of complex words. On the other hand, the second interpretation bans the syntactic relevance of the semantic information contained in the base of the derived word. The main objective of this paper is to show that this second interpretation of the Lexical Integrity Hypothesis is mistaken. In order to do so, we present three studies focused on three different Spanish word-formation processes, in which the semantic properties of the bases determine the syntactic behaviour of the complex lexical items. In the light of these studies, we claim that words derived by suffixation can be considered cases of “nuclearity paradoxes”, in the sense that in this type of complex words the categorial nucleus and the semantic one do not coincide in the same constituent.
Teoría morfológica; Integridad léxica; Palabras derivadas; Sintaxis; Palabras complejas; Chomsky, Noam