The case of literally true propositions with false implicatures

The truth-conditional content of the proposition expressed has been the subject of many debates. The different approaches differ as to the extent and the type of pragmatic inferences allowed into this notion. But a more recent approach offers a seemingly completely orthogonal discourse-based classification. We argue that this classification is crucial for assessing meaning and truth. To test this claim we conducted two experiments. Our stimuli were highly relevant Particularized Conversational Implicatures (PCIs), which are predicted to not influence truth-conditional contents. In both experiments, we were interested in the condition where the target sentence was embedded in a story that rendered it literally true while its PCI was false. Results from these experiments indicate that PCIs do influence truth judgments. Literally true proposition with false PCIs were quite often viewed by participants as not true, and even as false to varying degrees. We argue that such judgments testify that speakers evaluate truth based on the strength of the interpretation, regardless of the type of pragmatic inference involved. Thus, not only explicated inferences, but also strong implicatures affect truth judgments.