Kolloquium

Talk by David Lukeš (Charles University Prague)

To this day, MDA is strongly associated with its founding father of more than 30 years ago, Doug Biber. He is often closely involved with MDA-based studies, whether of English or other languages, offering both technical and interpretive support. In the case of the from-scratch MDA of Czech I’ve been involved with (see https://korpus.cz/mda), he has also provided extremely valuable feedback at various points of the project. However, we’ve also striven to maintain an independent spirit, and not just blindly follow guidelines. Given the rather ambitious scale of the project, this has led to various methodological insights which complement, or sometimes even depart from, the Biberian mainstream, including ways to determine the number of dimensions to settle on, the influence of text length on position within MDA space (and the related issue of using text samples vs. full texts), and novel applications of MDA (e.g. corpus comparison). Many of these are relevant for any quantitative approach to register analysis.