Neil Bermel: Methodological problems in research in variation in historical Czech

Neil Bermel (The University of Sheffield)

Goals: to examine the potential uses and limitations of two interfaces and two historical Czech corpora in the exploration of a research question (case form variation). The two corpora in question are DIAKORP and DIAKON: the first a 3.5m-token corpus of manually corrected texts, the second a 145m-token corpus including uncorrected texts, with a pronounced emphasis on the 19th and 20th centuries. The two interfaces are KonText and SyD: most Czech linguists will be familiar with the former, with its extensive search capabilities, but the simple graphic visualisations in SyD can be a useful tool for initial soundings into a hypothesis. Time allowing, we can also look at similar features in a related language using Google Ngram Viewer.