D-linking or set-restriction? Processing Which-questions in Dutch

Abstract

Research on Wh-questions suggests that Which questions are harder to process than Who questions (e.g., Who/Which athlete won the competition?). According to the Discourse (D)-linking Hypothesis, Which-questions differ from Who-questions in that Which questions need a link to a preceding discourse, while Who questions do not. However, this difference in processing may also be caused by differences in “set-restriction.” Who is much less restrictive in the set of potential referents it presupposes than Which N (e.g., Which athlete). A self-paced reading study investigated how Who and Which N questions were processed compared to questions involving the generic Which person, which refer to the same relatively unrestrictive referential set as Who. Our results showed that Which N questions were significantly more difficult than Which person or Who questions in object initial structures, supporting the hypothesis that increased processing cost for Which should be explained by a mechanism of set-restriction inherent to Which N questions. Additionally we found that the syntactic role of the possible referents in the discourse context affects question processing before the readers encountered disambiguating information.

Keywords:

• Sentence comprehension
• Discourse linking
• Set restriction
• Wh-questions

Notes

¹De Vincenzi (1996) included subject- and object-first wh-questions in an experiment with healthy adults. However, this research is difficult to incorporate in this discussion as only part of the results are discussed, with focus on interpretation in the light of De Vincenzi's Minimal Chain Principle.

²Cowles (2003, Chapter 3) suggests that with restricted sets readers keep all possible alternatives active in memory. This kind of approach might provide an explanation for processing difficulty; however, it suggests that the problem lies in working memory resources, see discussion below.

³Although note that this is not De Vincenzi's (1996) interpretation of her results; see page 7 of this paper.