New evidence on the Pseudorelative-First Hypothesis: Spanish attachment preferences revisited

Abstract This paper is aimed at testing the Pseudo Relative-First Hypothesis in Spanish, a proposal that may settle the long-standing question of cross-linguistic variation in attachment preferences. This hypothesis predicts that whenever a Pseudo Relative (PR) is obtainable, it will be preferred for parsing over a genuine relative clause (RC). Assuming that PRs only allow for high attachment (HA), it follows that HA will be obtained when a PR is possible. To test this hypothesis, two experiments previously conducted in Italian will be replicated in Spanish with sentences containing PR-ambiguous and unambiguous RCs. In experiment 1 PR-availability is manipulated by modifying structural conditions, while in experiment 2 the PRs are only manipulated through semantic conditions. The results obtained show that PR-possible contexts do not yield the predicted HA. It will be argued that this finding, together with the data provided by the Italian experiments, only partially support the PR-First Hypothesis.