The present paper is a reply to the article Perspektivy korpusové lingvistiky: deskripce, nebo explanace by František Štícha (2015) which is a critique of recent studies by Radek Čech (2014) and Jan Chromý (2014). It is shown that Štícha’s argumentation is based on an inaccurate reading of the two criticized studies. Also, Štícha’s conception of corpus linguistics as a discipline which aims to capture the morphological and syntactical norm of well-educated people is rather limited. This narrow-minded view seems to be another reason of Štícha’s misunderstanding of the criticized papers.
Large synchronic textual corpora of the Czech National Corpus are built as representative: they contain a balanced quantity of texts of various styles, divided into three genre subcorpora: fiction, technical/scientific literature and journalism. Comparisons of these genres have been performed on phonological and morphological level; in this paper, I deal with differences between genres on the surface-syntactic level. I use an automatic syntactic annotation of the SYN2005 corpus in the formalism of the analytical layer of the Prague Dependency Treebank. I compare the frequencies of syntactic functions of nouns in the three genres represented by the corresponding subcorpora of SYN2005. I also present a more detailed analysis of four syntactic phenomena: subtypes of the function of attribute in non-prepositional genitive; frequencies of groups of the type pan Novák (Mr. Novák); frequencies of the function of agent in passive constructions expressed by nouns in non-prepositional instrumental and the ratio of the expression of the nominal part of a verbal-nominal predicate by nominative and instrumental. Significant differences found between genres in all the syntactic phenomena analyzed show that in comparing corpora one should carefully monitor their genre composition.