The project aims at a comprehensive explanation of the ways and means ensuring text coherence between sentences in Czech in cases where no discourse connectives are used (so called implicit discourse relations). Based on a corpus analysis of written texts, index of the coherence means will be determined and their interplay and relation to the syntactic structure will be examined. During the research, the hypothesis of compensation of complexity will be subject to verification. This hypothesis is based on the assumption that when a relation in a given context is let to be implicit, the reader needs to make a greater effort during decoding and therefore the given context does not contain other phenomena requiring increased decoding efforts (e.g. noncanonical word order, complex syntactic structure etc.). The research will by supported by psycholinguistic experiments aimed at text understanding.