Date of Award


Document Type

Campus Access Dissertation



First Advisor

Amit Almor


Discourse coherence often relies on anaphoric expressions such as repeated names, overt pronouns, and null pronouns. Gordon, Grosz, and Gilliom (1993) showed that repeated names are harder to process than pronouns when the antecedent is the subject of a sentence, an effect which they called the Repeated Name Penalty (RNP). The RNP has been demonstrated in English, and it has also been extended to Mandarin Chinese (Yang, Gordon, Hendrick, & Wu, 1999), which is a fully pro drop language. However, it is not clear whether the RNP extends to null subject languages like Spanish. In four self-paced sentence-by-sentence reading experiments we examined the difference in the processing of Spanish discourses containing repeated names, overt pronouns and null pronouns.

The results from this study confirm that the RNP extends to Spanish and they show that overt pronouns cause a processing delay when they refer to salient antecedents, an effect which we call the Overt Pronoun Penalty (OPP). Although both the RNP and the OPP can be conceptualized as penalties for overspecificity, there is a typological distinction between them. While the RNP appears to be a universal phenomenon, independent of the anaphoric inventory of any specific language, the OPP appears to be confined to a subset of pro drop languages. We propose that the OPP reveals the narrow pragmatic function of the overt pronoun in null subject languages like Spanish, in which the verbal morphology encodes the grammatical features that are sufficient for the identification of the antecedent of an anaphoric expression.