%A Dr Ernst-August Gutt %J Logical Relations in Discourse %T Logical connectives, relationships and relevance %X This paper was written in the context of a workshop (held in 1989) on logical connectives in discourse. It addresses difficulties that arise from attempts at analysing logical connectives by assigning logical relations to them, taken from a typological listing. As the workshop itself showed, such attempts typically face the problem of on the one hand needing to be broad enough to cover all the instances of relations found in texts and on the other hand needing to be specific enough to differentiate types of relationships from each other. Applying the relevance-theoretic framework proposed by Sperber and Wilson, this paper argued that the root of the problem is that (logical) connectivity in discourse is not created by a fixed set of interpropositional or rhetorical relations, but by the search for relevance, through the inferential interaction between a given utterance and its context. This explains the virtually unlimited variety of relations that can and do arise in texts, without reliance on any defined set of relations. After a brief introduction to the relevance-theoretical framework, the paper applies it to the analysis of the connective ?m in Silt?i, an Ethio-Semitic language. It attempts to show how the variety of relations associated with this connective can be explained in terms of the interaction of a simple semantic property with different pieces of contextual information, accessed in the search for relevance. The paper then proposes and illustrates three methods of testing the validity of such relevance-theoretic analyses of connectivity, two of which can be carried out experimentally. %K relevance theory, inference, logical connective, implicature, pragmatic connective %P 1-24 %E Eugene E Loos %D 1999 %I Summer Institute of Linguistics %L cogprints3112