THE USE OF SUPERFICIAL AND MEANING-BASED REPRESENTATIONS IN INTERPRETING PRONOUNS - EVIDENCE FROM SPANISH

In this paper, we address the question of whether a representation of the surface form of a text is directly implicated in the interpretation of definite pronouns in that text. According to an influential theory proposed by Sag and Hankamer (1984), it should not be, because definite pronouns are model-interpretive anaphors that take their meaning from elements in a representation of content. We report three experiments in Spanish, a language with non-semantic gender, in which pronouns can match their antecedents on the basis of morphosyntactic properties alone. The first experiment suggested that a surface representation might not be implicated in the interpretation of pronouns, since a gender match speeded only the interpretation of pronouns referring to people and not those referring to things. However, a questionnaire study (Experiment 2) confirmed that our strategy of modelling the Spanish sentences in Experiment 1 on sentences used in English studies had affected the results, and a further on-line experiment provided evidence that the interpretation of pronouns referring to things can be speeded by a gender match (Experiment 3). We discuss the implication of these findings for theories of text comprehension.