May 2nd, 2014, 14:00 – 14:45
In this paper, I will argue that the meaning of negative utterances is underspecified at a semantic level. Its target domain is either contextually accessible, unless its positive counterpart (POS) can be accessed in the current conversational context. A more positive situation is when a negative utterance (NEG) is followed by a corrective one (COR).
The first part of the talk has as a main goal to make explicit the relation carried by the three types of content implied by a negative utterance: POS, NEG and COR. Logical, semantic and pragmatic criteria will support the claim that negation has mainly three types of uses: descriptive, metalinguistic#1 and metalinguistic#2. These two types of metalinguistic usages are respectively cases of negation of a scalar implicature and negation of a presupposition.
The second part of the paper will discuss the case of metalinguistic#1 negation, where a scalar implicature of POS is defeated. The issue that will be discussed and argued for is the reasons why negative utterances with a scalar operator do not trigger by themselves an implicature, whereas their corresponding POSs do. Similar cases with entailments, presuppositions as well as with explicatures will be examined too, which show the limit rational inferences with negative utterances: some inferences are safe under negation, other are not. Implicatures, and more specifically scalar implicatures, are such cases. A logical analysis of semantic and pragmatic contents will be given for supporting the view that inferences under negation are not safe from a truth-functional point of view.
Jacques Moeschler, Universität Genf