Annotated samples for: When does semantic change lead to semantic loss? Metaphor vs inference-driven metonymy

This dataset contains semantically annotated random samples for five adjectives - <i>bright</i>, <i>dumb</i>, <i>narrow</i>, <i>strange</i> and Dutch <i>eng</i> 'frightening' - analyzed for the paper 'When does semantic change lead to semantic loss? Metaphor vs inference-driven metonymy' (Ceuppens & De Smet, accepted). The dataset was created to investigate how different mechanisms of semantic change - specifically metaphor and inference-driven metonymy - affect the likelihood of retention or loss of source meaning over time. Random samples from different time periods were drawn from EEBO, CLMET, BNC, COHA and the Dutch corpora Couranten Corpus, Dutch C-CLAMP, Sonar and CGN. Each sample was annotated for sense of the adjective in context according to the data coding protocols described in Ceuppens & De Smet (accepted). This dataset may be useful for researchers looking to replicate the study's results or build upon its findings.