# Resolution of Null and Overt Pronouns in Catalan: An Eye Tracking Study

## TL;DR

This study explores how people resolve subject pronouns in Catalan, a language that allows null pronouns, using eye-tracking to examine factors like gender and sentence structure.

## Contribution

The study reveals that positional prominence, rather than grammatical subjecthood, strongly influences pronoun resolution in non-canonical sentence structures in Catalan.

## Key findings

- Null pronouns strongly prefer subject antecedents, while overt pronouns show more variability.
- In non-canonical OVS sentences, first-mentioned antecedents guide resolution more than syntactic subjects.
- Overt pronouns rely heavily on gender cues in unambiguous contexts but show delayed resolution in ambiguous ones.

## Abstract

This study investigates subject pronoun resolution in Catalan, a null-subject language, by examining linguistic and cognitive factors. We focus on gender, pronoun type (null and overt), and syntactic function (subject and object), as well as the order of mention (first and second) of antecedents. Using eye-tracking in the Visual World Paradigm (VWP), we analyze the interplay between these factors in both canonical (SVO) and non-canonical (OVS) sentence structures. Experiments 1, 2 and 3 (Study 1) explore how gender cues and pronoun type influence pronominal resolution in SVO sentences. Our findings reveal that gender facilitates early resolution but primarily affects first-mentioned antecedents. Null pronouns exhibit a strong preference for subject antecedents. However, overt pronouns, contrary to offline studies, show no strong preference for object antecedents and instead tend to favor first-mentioned subjects in ambiguous contexts. In unambiguous gender-marked cases, overt pronouns heavily rely on gender cues for reference resolution, but in ambiguous contexts, they display delayed resolution patterns. These results suggest that overt and null pronouns do not exhibit the expected division of labor proposed by the Position of Antecedent Hypothesis (PAH). Instead, null pronouns robustly refer to subject antecedents, while overt pronouns show more variability. Experiments 4 and 5 (Study 2) extend this investigation to OVS sentences, revealing that linear order exerts a stronger influence on pronoun resolution than grammatical subjecthood. Specifically, first-mentioned antecedents, rather than syntactic subjects, guide resolution in OVS structures, particularly for overt pronouns. This challenges previous findings that overt pronouns prefer object antecedents and highlights positional prominence as a crucial factor in pronoun resolution in Catalan. Overall, our results contribute to a broader understanding of pronoun processing across null-subject languages and underscore the importance of positional prominence in pronominal interpretation and suggest that, in non-canonical structures, order-of-mention can override subjecthood in guiding resolution.

The online version contains supplementary material available at 10.1007/s10936-025-10189-7.