In two experiments, the study explores how listeners adapt their predictions to the perceived reliability of a speaker by investigating whether native German listeners use gender cues on articles to predict upcoming nouns (e.g. dermasc → Mannmasc/Käsemasc “the → man/cheese”) in error-free native-accented and foreign-accented speech. In Experiment 1, listeners robustly used natural gender (when der refers to a human referent, e.g. man), in both native-accented and foreign-accented speech. Yet, for grammatical gender with inanimate nouns (e.g. cheese), predictive use of gender was delayed in foreign-accented speech. In Experiment 2, the foreign-accented speaker made unrelated grammatical errors, but remained accurate with gender. Listeners continued to predict for natural gender, yet did not show reliable cue use for grammatical gender. Together, we interpret these findings to suggest that listeners integrate top-down indexical cues about the speaker, assumptions about their likely errors, and distributional cues about speakers’ performance to fine-tune their predictions.
No takes yet. Share an insight, caveat, or question.
Hopp et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: