Abstract
The article aspires to provide a detailed semantic classification and description of scalar modifiers in contemporary Czech (e.g. velmi, skoro, celkem, poněkud, alespoň, přibližně). Words of this class are seen as operating on semantic scales of two basic types – continuous scales, primarily represented by gradable adjectives, and discrete scales, primarily represented by numerals.
The function of scalar modifiers is to state some degree(s) of these scales as valid. As a precondition of the final description and classification, the semantics of gradable adjectives is broadly addressed with respect to the problem of semantic vagueness and pragmatic imprecision.
The well-known division into the categories of relative, maximal and minimal adjectives is applied to Czech and some tests based on inferential schemes are shown. The diachronic origin of scalar modifiers is discussed on the basis of grammaticalization theory.
After that, the actual semantic classification and description of Czech scalar modifiers is presented, based not on mere semantic intuition but supported by corpus co-occurence data.