Third-person overt pronoun and zero reference in estonian: insights from two experiments

This study reports two experiments that investigate the variation between two Estonian third-person referential devices - zero reference vs. the overt pronoun ta. First, in a speech-restoration paradigm (Experiment 1), we test whether the structure of a reference chain affects referential choice. Second, drawing on Experiment 1, we designed an acceptability judgement task (Experiment 2) to explore a possible sys tematic variation between zero reference vs. overt pronoun in different sentential configurations (two separated sentences vs. one coordinated sentence). Our findings suggest that sentential configuration affects referential choice. This result accords with earlier observations that salience-only accounts cannot fully explain referential choice. Furthermore, this study supports the form-specific account of reference, by showing that zero reference is more sensitive to sentential configuration than the overt pronoun. We suggest that the use of referential devices reflects a dualistic nature of discourse processing, and local coherence as well as discourse segmenting effects should be considered for a convincing explanation of referential usage patterns.