A contrastive description of English and Czech using the methodology of n-gram extraction

This dissertation examines phraseological patterns in three registers (parliamentary debates, newspaper reporting, children's fiction) in English and Czech. It identifies and analyses recurrent word sequences through n-gram extraction, aiming to characterise the phraseology of each register and compare them cross- linguistically, while observing how the n-gram method can be adapted to accommodate for the typological properties of each language. Czech is particularly challenging in this respect due to its morphological and positional variability. The dissertation comprises three case studies, each focussed on a different register. The first case study explores different n-gram lengths using a small corpus of a specialised register - parliamentary debates, suggesting that for a comprehensive register characterisation, different lengths should be combined. It notes the importance of discourse-structuring patterns and the problem of overlaps between n-grams. In the newspaper study, I extract n-grams containing prepositions - a convenient starting point given their frequency and involvement in text- structuring. N-grams are complemented with collocation analysis, revealing some evaluative prosodies and semantic preferences of patterns and suggesting that the newspaper register is not purely..