Corpus Methods for Age Categorization Research
In the hypercommunicative present-day context, ageism (like sexism and racism) is perceived not just as an issue involving social behavior towards age groups, but also as an issue involving the categorization of those groups, either through explicit labeling or through implicit framing. To get a good grip on such categorization phenomena, a spectrum of different methods may be used, but specifically also (given the huge amount of digitally available texts that should be included in the description), we will need appropriate methods for corpus semantic research.
Against that background, this paper explores the usefulness of a distributional semantic approach (in the sense of Lenci 2018, ‘Distributional models of word meaning’, Annual Review of Linguistics 4:151-171) for the study of age categories in large corpora. Focusing on the Dutch terms ouderen and senioren ‘older people’, we will distributionally investigate whether the semantic relationship between these two near-synonyms changes over a timespan of 20 years, and, to the extent that there is a shift in usage, which underlying dimensions structure the relationship.