In multi-clause sentences, which clause carries the at-issue point is expected to be influenced by whether a clause is at the Right Frontier: Last-uttered clauses or clauses that subordinate these are expected to be at-issue. In a Dutch forced-choice experiment, we measure the rate at which comprehenders interpret an ambiguous pronoun to refer to one of two possible antecedents in a preceding sentence. We manipulated the type (matrix vs. subordinate) and position (sentence-early vs. sentence-final) of the clauses hosting the antecedents, as well as the topicality of the subject (mentioned in context vs.not mentioned in context). We find no effect of topicality, but we find that clause position and type influence the at-issue status of clauses within multi-clause sentences in Dutch: When multiple clauses are at the Right Frontier, sentence-final clauses are more likely hosts for at-issue content, and matrix clauses more so than subordinate clauses in this position.

Pronoun interpretation seems to be driven by structural factors, but also by factors related to meaning. In a forced-choice pronoun interpretation experiment, we compare the impact of the next-mention bias associated with transfer-of-possession-verbs on the interpretation of three Dutch pronominal forms that differ in the strength of their structural biases: reduced personal pronoun ze ‘she_reduced ', full personal pronoun zij ‘she_full', and demonstrative pronoun die ‘that'. In addition to replicating the common Goal-bias associated with transfer-of-possession verbs, results show significant differences in the proportion of pronoun resolved to the preceding subject between all three pronominal forms. However, the effect of the next-mention manipulation did not differ between pronominal forms. These findings are in line with a model of pronoun interpretation that combines structural and meaning-related factors, and present particularly strong evidence against models that posit that pronoun interpretation is the mirror image of pronoun production.