Modality and the semantics-pragmatics interface
This thesis explores certain aspects of the structure of lexical semantics and its interaction with pragmatic processes of utterance comprehension, using as a case-study a sample of the English modal verbs. Contrary to previous polysemy-based accounts, I propose and defend a unitary semantic account of the English modals, and I give a relevance-theoretic explanation of the construction of their admissible (mainly, root and epistemic) contextual interpretations. Departing from previous accounts of modality, I propose a link between epistemic modality and metarepresentation, and treat the emergence of epistemic modal markers as a result of the development of the human theory of mind. In support of my central contention that the English modals are semantically univocal, I reanalyse a range of arguments employed by previous polysemy-based approaches. These arguments involve the distributional properties of the modals, their relationship to truth-conditional content, the status of so-called speech-act modality, and the historical development of epistemic meanings: it turns out that none of these domains can offer reasons to abandon the univocal semantic analysis of the English modals. Furthermore, I argue that the priority of root over epistemic meanings in language acquisition is predicted by the link between epistemic modality and metarepresentation. Finally, data from a cognitive disorder (autism) are considered in the light of the metarepresentation hypothesis about epistemic modality. The discussion of modality has a number of implications for the concept of polysemy. I suggest that, despite its widespread use in current lexical semantics, polysemy is not a natural class, and use the example of the Cognitive Linguistics to illustrate that polysemy presupposes some questionable assumptions about the structure of lexical concepts. I propose a division of labour between ambiguity, semantic underdeterminacy, and a narrowed version of polysemy, and present its ramifications for the psychology of word meaning. In the final chapter, I extend the proposed framework for modality to the analysis of generic sentences, thereby capturing certain similarities between genericity and modality.