Several theoretical and methodological notes on the Grammar of Contemporary Czech

The article is concerned with the critical analysis of some aspects of the methodology and language theory on which Mluvnice soucasne cestiny [Grammar of Contemporary Czech] (Cvrcek et al. 2010) is based. First, the statement by the authors of Mluvnice soucasne cestiny concerning the character of the description of language and its relationship to reality is criticized. Specifically, it is argued that the use of corpora does not lead to the recognition of "real" language properties, since every description of reality is an interpretation. The idea concerning the representativeness of corpus is also called into question. Furthermore, a serious statistical deficiency is revealed. Namely, it is demonstrated that the assignment of difference expressed by a percentage as significant (or non-significant) is unacceptable - the results should be interpreted using common statistical methods. Next, fundamental problems of the semantic theory (naive mentalism) used in Mluvnice soucasne cestiny are discussed. A rather different view of a semantic analysis based on experimental methodology is mentioned as an alternative to the traditional approach. Finally, the langue-parole dichotomy, viewed by the author as a residue of Platonism, is argued to be inconsistent with the empirical methodological approach adopted by the authors of Mluvnice soucasne cestiny.