On an Apparent Freedom of Czech Word Order. A Case Study

The aim of the present contribution is to document, on the material of the Prague Dependency Treebank (PDT), that the assumed freedom of Czech word order is not really a freedom but that it is guided by certain principles, different from the grammatically given principles determining the word order in some other European languages such as English, German or French. After a short introductory section summarizing the traditional views on Czech word order (Section 1) we briefly characterize our approach to the information structure of the sentence (TFA) and its representation in the annotated corpus of Czech (PDT, Section 2). In Section 3 we present the results of the automatic procedure for dividing the sentence into its topic and focus part and compare these results with human annotators decisions. In Section 4 we put forward and test the hypothesis on the order of elements in the focus part of the sentence, adding some observations that follow from the tests. A summary of our investigations is given in Section 5. 1 Traditional views on Czech word order Since the pioneering studies of Vilém Mathesius [17] topic–focus articulation is considered to be the primary factor determining the word order in Czech; other factors influencing the Czech word order are then considered to be secondary, be it the grammatical or the rhythmical factors (see [5, p. 609]). According to Uhlířová [23], among the word order principles governed by grammar there is the basic position of a congruent attribute before the governing noun and the modificiation of manner before the verb; grammatical as well as rhythmical aspects govern the position of the so-called clitics, typically the second sentence position, i.e. the position after the first member of the sentence carrying the stress (the so-called Wackernagel position). A special attention is paid to the position of the so-called rhematizers; their position is closely related to their scope.