This paper is based on data from the LANCHART (Language Change in Real Time) corpus. LANCHART is ‘data-based’ in that transcripts are orthographically normalized and both transcripts and annotations are time stamped and stored in a database. The corpus is structured according to generation, gender, geography, class, and time of recording. The socio-linguistic issue treated is the relationship between genre, as defined in the context of a so-called Discourse Context Analysis, and a particularly frequent and significant phonetic variable, viz. the (æ) variable in Modern Danish. Through repeated searches in the corpus we show that genres do have an effect on phonetic variation and that they frequently develop in real time, at least with regard to the patterning of the selected phonetic variable. The results are discussed in the context of intra-individual variation within the sociolinguistic interview, i.e., the timehonored style problem.