Digital Humanities Day #3

Language users adapt their language to the medium they use: for example, a given author writes differently in a professional report than a personal text message, or speaks differently in a formal interview than in private conversation. At the same time, individuals also show characteristic features in their language use that allow us to draw inferences about the author of a certain text. We are creating a corpus of texts from the same authors in different social media (blogs and Twitter) that enables us to study the interplay of stable individual characteristics of language use and variable adaptation to the medium. We have collected blog posts and tweets from over 40 authors, comprising hundreds of documents. In an initial pilot study, we show a linguistic analysis of intensifiers in this dataset and their variation across media and individuals. We also present our experiences with legal and ethical challenges in using this kind of social media data for (linguistic) research.