Believing What You're Told: Politeness and Scalar Inferences

The experimental pragmatics literature has extensively investigated the ways in which distinct contextual factors affect the computation of scalar inferences, whose most studied example is the one that allows “Some X-ed” to mean Not all X-ed. Recent studies from Bonnefon et al. (2009, 2011) investigate the effect of politeness on the interpretation of scalar utterances. They argue that when the scalar utterance is face-threatening (“Some people hated your speech”) (i) the scalar inference is less likely to be derived, and (ii) the semantic interpretation of “some” (at least some) is arrived at slowly and effortfully. This paper re-evaluates the role of politeness in the computation of scalar inferences by drawing on the distinction between “comprehension” and “epistemic assessment” of communicated information. In two experiments, we test the hypothesis that, in these face-threatening contexts, scalar inferences are largely derived but are less likely to be accepted as true. In line with our predictions, we find that slowdowns in the face-threatening condition are attributable to longer reaction times at the (latter) epistemic assessment stage, but not at the comprehension stage.