Abstract
The present thesis deals with, and elaborates on, the concept of sense inheritance (particularly with respect to suffixation in Czech). Drawing on more general remarks on lexical meaning, polysemy, and the semantic relatedness of a derived word and its base, the theoretical part of the thesis delineates the notion of sense inheritance, originating within the framework of construction grammar.
It is shown that the idea of sense inheritance has at least implicitly been present in Czech lexicology (cf. the derivational criterion for identifying senses of polysemous words) and lexicography. It is argued that the traditional Czech theory of word formation relies on the assumption of the semantic transparency or even compositionality of motivated words, which can hardly be squared with empirical evidence concerning sense inheritance, nevertheless.
In the empirical part of the thesis semantic relations are analyzed that hold between derivatives within 43 word families of polysemous bases (denoting body parts), and subsequently the sense-inheritance properties of the 80 most frequent (hence established and lexicalized) derivatives with the agentive suffix -tel and polysemous base verbs are contrasted with those found in 100 hapax legomena derived from polysemous bases with the same affix. This enables not only the identification of general sense-inheritance tendencies in Czech word-formation (e.g. the fact that a word derived from a polysemous base inherits only 40% of its senses on average), but also the differentiation of the structural aspect of sense inheritance from its processual aspect.