Linguistic research on anaphora (Hankamer & Sag, 1976; Sag & Hankamer, 1984) suggests that anaphors can be divided into two classes: Surface anaphors that find their antecedents in some level of linguistic representation, and deep anaphors that find their antecedents in a discourse model or a corresponding mental representation. In three experiments, we tested the hypothesis that the syntactic form of the antecedent for a subsequent anaphor would affect the difficulty with which surface anaphors but not deep anaphors would be comprehended. In a “makes-sense” judgement task, surface anaphors were judged to make sense more often when the antecedent was introduced in a phrase that was syntactically parallel to the anaphor than when it was syntactically non-parallel. In contrast, the syntactic form of the antecedent did not affect judgements to the deep anaphors. Parallelism did, however, influence comprehension times for both types of anaphors. The results provide qualified support for the hypothesis that deep and surface anaphors access different types of representations during comprehension.
No takes yet. Share an insight, caveat, or question.
Tanenhaus et al. (1990) studied this question.
Synapse has enriched 5 related papers. Consider them for comparative context: