Eventive-selection bias in relative clause attachment in English: Evidence from native and non-native speakers

This study investigates how the selectional properties of matrix verbs influence the parsing of reduced relative clauses in English. Building on the Pseudo-Relative-first hypothesis (Grillo & Costa, 2014), we test whether a processing bias toward eventive complements after perception verbs affects attachment preferences in reduced relative clauses—a structure that has not previously been examined in this context in English. Using a self-paced reading paradigm, the experiment compares full and reduced relative clauses in sentences with perception and non-perception matrix verbs and examines both attachment choices and processing times. Participants included adult native speakers of English and adult native speakers of Norwegian, Portuguese, Armenian and Catalan/Spanish for whom English was a second or third language. Results from mixed-effects models show that perception verbs significantly increase high-attachment interpretations in reduced relative clauses, but not in full relative clauses, supporting the prediction that eventive complements guide parsing decisions. Reading time data further indicate increased processing costs for reduced relative clauses, consistent across groups. Together, the findings provide evidence that reduced relative clauses in English are sensitive to eventive-selection biases, and that this sensitivity is homogeneous across several multilingual populations.