Answers are grounded in this paper and cite their sources.
Communication may include ambiguous pronouns that refer to different antecedents. Behavioural studies show reduced processing costs for ambiguous pronouns, indicating an ambiguity advantage and shallow processing. This raises the question: “How does ERP evidence for ambiguous pronouns align with psycholinguistic research suggesting that incremental interpretation may involve shallow processing?” We present results from an EEG/ERP experiment in German examining plural pronoun contexts (e.g. sie/they) with antecedents in either set-subset relations (e.g. artists ⊇ sculptors) or non-subset relations (e.g. artists ⊉ journalists), and verbs that are either ambiguous (e.g. browsed) or unambiguous (e.g. carved). No ERP effects were observed at pronoun and ambiguous verb regions, suggesting shallow processing. The set-subset unambiguous context showed an N400 relative to the non-subset case, indicating costs when an ambiguous interpretation is not possible. This negativity was positively correlated with surprisal estimates from a Generative Pre-trained Transformer, supporting a shallow processing strategy under less predictive contexts.
No takes yet. Share an insight, caveat, or question.
Çokal et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: