Pronoun resolution in two-clause sentences: Effects of ambiguity, antecedent location, and depth of embedding

Abstract

The two experiments reported here were designed to study anaphoric pronoun resolution by measuring the effects of syntactic ambiguity, antecedent location, and depth of antecedent embedding on sentence-final reading comprehension time. Fifty-two college students read sentences in a word-by-word self-paced task and answered a question after each. Syntactic cues indicated the appropriate antecedent in the unambiguous sentences, while in the ambiguous sentences, the semantic context of the second clause cued the antecedent. Results of Experiment 1 showed that comprehension time was significantly faster for the unambiguous sentences. In both experiments, comprehension time in ambiguous sentences was faster when the antecedent was located early in the first clause than when it was located late in the clause. In Experiment 2, the interaction of embedding by location showed slower comprehension times for sentences with early deep antecedents than for sentences with late shallow antecedents. These results are consistent with predictions derived from Hobbs' (1978, Lingua, 44, 311–338) algorithm for a top-down left-to-right breadth-first search.