Conventionalization of politeness and questions under discussion in pragmatic inferencing: Interpreting scalars in interpersonal contexts

This study investigates how scalar expressions are interpreted in interpersonal contexts. Prior research suggests that weaker scalars (e.g., some or youxie ) may serve as polite hedges that suppress scalar implicatures (SIs), especially in face-threatening acts. However, findings remain inconclusive, often neglecting the role of discourse structure and hearer-based inferences. Drawing on a frame-based model of politeness, we argue that politeness is not inherently encoded in scalar statements but emerges from hearers' inferences shaped by contextual expectations, particularly questions under discussion (QUDs) and conventionalized response strategies. Across three experiments in Mandarin Chinese, we first elicit the QUDs associated with scalar utterances, then examine conventionalized speaker responses to these QUDs, and finally assess hearers’ interpretations of scalar utterances. Results reveal that politeness-based interpretations, which can suppress implicature derivation, emerge primarily when the utterance is judged to serve a face-mitigating function, especially in contexts where the QUD highlights a stronger, more face-threatening alternative. By contrast, SIs are more likely to be derived when the speaker is perceived as aiming to be truthful, particularly when their utterance deviates from conventionalized expectations for indirectness. These findings challenge traditional speaker-intention-based accounts and support a hearer-driven, discourse-sensitive approach to politeness and pragmatic inferencing. • Politeness emerges from discourse context rather than scalar terms alone. • Questions under discussion and conventional politeness strategies shape scalar implicature interpretation. • Scalar expressions trigger social inference differently in face-sensitive contexts. • Scalar implicatures are suppressed only when politeness intentions are inferred. • Interpretive demands vary from routine to social reasoning based on discourse cues.