Quantitative Methoden einer kognitiven Texttypologie: Automatische Genre-Klassifizierung als Rekonstruktion kognitiver Weltmodelle

This research combines automatic pattern recognition with parameters of a cognitive text typology to develop a method for genre classification based on annotated corpus data, considering text-linguistic parameters such as information density, elaboration measures or frequent event schemata. Using a corpus of Ob-Ugrian folk tales and other, primarily oral genres, classification and clustering algorithms are applied to evaluate the effectiveness of this method for the induction of quantitative text structure types. This aims to reconstruct text world models as genre-specific, schematic text structure models of human cognition, learned through typification of language use.