Semantics and epistemology: The meaning of ‘evidentials’ in a cross-linguistic perspective

Abstract

The author argues that meaning is encoded not only in words but also in grammatical categories. The meanings encoded in grammar (just like those encoded in the lexicon) are language-specific. If one attempts to identify the meanings encoded in different languages by means of the same, arbitrarily invented labels, one can only conceal and obfuscate the language-specific character of the categories to which they are attached. To be able to compare grammatical categories across language boundaries we need some constant points of reference, which slippery labels with shifting meanings cannot possibly provide. Universal (or near-universal) semantic primitives (or near-primitives) can provide such constant and language-independent points of reference. They offer, therefore, a secure basis for a semantic typology of both lexicons and grammars. At the same time, they offer us convenient and reliable tools for investigating the universal and the language-specific aspects of human cognition and human conceptualization of the world. In this paper, the author illustrates and documents these claims by analyzing one area of grammar in a number of different languages of the world: that area which is usually associated with the term of ‘evidentiality’. As the goal of the paper is theoretical, not empirical, the data are drawn exclusively from one source: the volume entitled Evidentiality, edited by Chafe and Nichols (1986). The author reexamines the data presented in this volume by experts on a number of languages, and tries to show how these data can be reanalyzed in terms of universal semantic primitives, and how in this way they can be made both more verifiable (that is, predictive) and more comparable across language boundaries.