Indirect anaphora in English and French: A cross-linguistic study of pronoun resolution.

There is disagreement within both linguistics and psycholinguistics concerning the use of unaccented third person pronouns to refer to implicit referents. Some researchers (e.g. Erk & Gundel, 1987) argue that it is impossible or highly marked, while others (e.g. Yule, 1982) maintain that it is not only acceptable but commonly used in normal discourse. However, both sides in the debate may be correct: While peripheral implicit referents (which evoke the means or the instrument by which a given state of affairs is established) are not easily referred to using pronouns, central or nuclear implicit referents are. We tested this hypothesis in two experiments, involving different languages (English and French). The results of both experiments show that pronominal reference to implicit referents caused slower reading times compared to explicit referents for peripheral referents only. We discuss these results with respect to Gundel, Hedberg and Zacharskis (1993, 2000) Givenness Hierarchy.