The comprehension of quasi-performative verbs in verbal commitments: New evidence for componential theories of lexical meaning

Quasi-performative verbs collectively satisfy certain constraints for performative speech act syntactic form but differ from one another in the extent of match between the speaker intentions and beliefs conveyed by their use and those specified for a “felicitous” or ideal speech act of a given type. Tested in two experiments was a componential model representing affirmative and negative forms of two speaker intentions— Desire and assuredness of Ability to perform some task—conveyed by four quasi-performative verbs (i.e., promise, agree, hope, and guess) when used in verbal commitments embedded in a dialog script. In Experiment 1, latencies for verifying questions probing the speaker intentions conveyed by the verbs supported the model: (a) Affirmative components were processed faster than negative components and (b) YES responses were faster than NO responses for affirmative components, while (c) YES responses were slower than NO responses for negative components. In Experiment 2, verbal commitment reading latency was shorter when speaker intentions—explicitly presented in the dialog script—were polarity consistent than when they were polarity inconsistent with the corresponding verb components. Further, verification latencies of Experiment 1 and posttrial judgments of “speaker committedness” from Experiment 2 indicated component-independent processing in comprehending these verbs. These findings argue strongly for a propositional representation of the pragmatic meaning of quasi-performative verbs. It is proposed that the particular salience of speaker intentions in speech act comprehension predisposes this aspect of quasi-performative verb meaning to be decomposed into corresponding components.