Acceso usuarios
Julie Franck, Michel Hupet
This paper surveys the literature on grammatical agreement in order to synlhesise and distinguish theoretical arguments and experimental data relative to the maximal and minimal input hypotheses. The maximal input hypothesis assumes that during the construction of a syntactic frame for the sentence, the grammatical encoder takes advantage of any relevant information, i.e. information within the lexicon, but also from the conceptual and morphophonological levels. In contrast, the minimal input hypothesis assumes that only a limited flow of lexico- syntactic information reaches the grammatical encoder. Experimental data on adult performance in grammatical agreement tasks are considered in different languages. Major differences are reported between English and romance languages like French, Italian and Spanish as to the question of the minimal vs. maximal input frameworks. It is suggested that contradictions between results have to be examined in the light of the variability within each language as well as between languages. We suggest that the use of a particular information (conceptual, lexical, morphophonological) depends simultaneously on its validity and on its cost, with these two indexes varying both within and between languages.
