In this paper, we argue that there is another approach to the study of historical pragmatics beyond those explicitly mentioned in Jacobs and Jucker (1995). We label this approach "sociophilology". Moreover, we demonstrate how this approach can be effectively pursued by combining two corpus linguistics techniques: corpus annotation and "keyness" analysis. Specifically, we draw from the Sociopragmatic Corpus (1640-1760), an annotated subsection of comedy plays and drama proceedings taken from the Corpus of Dialogues 1560-1760, as a means of identifying the statistically-based style markers, or key items, associated with a number of social role dyads (including examiner to examinee and master/mistress to servant). We will show how such an approach might be used to uncover differential distributions of personal pronouns, interjections, imperative verbs, politeness formulae, etc., and how, by combining qualitative analysis with quantitative analysis, one can scrutinise such material for pragmatic import.