This paper seeks to approach the topic of historical language choice from a quantitative perspective, arguing that solid baseline evidence drawn from a substantial dataset is a much-needed complement to the largely qualitative findings of previous research. We propose a methodological framework which enables us to examine the sociolinguistic factors that condition language choice in the private domain. Illustrating the possibilities of our methodology, we present a case study on Dutch-French language choice in the Northern Low Countries (i.e., the present-day Netherlands), focusing on nineteenth-century family correspondence. Our paper shows that a careful selection procedure is crucial in order to achieve a balanced representation of language choice in a large-scale dataset. With respect to our analyses, the role of French in private letters turns out to be relatively small against the prevalence of Dutch. However, interesting patterns become visible when looking at regional differences, gender constellations and familial relationships. These quantitative findings can therefore constitute an interpretational frame for qualitative studies on historical language choice in the Dutch-French context and beyond.