Automatic extraction of potential examples of semantic change using lexical sets

This paper describes ongoing work on automatically finding candidates for semantic change by comparing two corpora from different time periods. Semantic change is viewed in terms of distributional difference with a computational and linguistically motivated approach. The data is parsed, lemmatized and part of speech information is added. In distributional semantics, meaning is characterized with respect to the context. This idea is developed from Firth (1957) and is formulated according to ‘the distributional hypothesis’ of Harris (1968). A method is developed to describe distributional behaviour in order to track semantic change over time. We will explore statistically ranked lists of verbal predicate nominal object constructions and examine differences at the level of word types.