The challenge of historical data: from sources and corpora to answering research questions

Abstract

At the start of each research enterprise, historical sociolinguists have to deal with the key issue of the required historical data. In the present world of big data optimism, the idea may arise that the time-consuming compilation of specialised corpora is no longer needed. Can taking a shortcut still lead us to convincing results? In this article I discuss the crucial relationship between specific research questions and appropriate historical data. This methodological issue will be illustrated by concentrating on three historical-sociolinguistic research programmes, conducted at Leiden University: Letters as loot: Towards a non-standard view on the history of Dutch (2008–2013), Going Dutch: The construction of Dutch in policy, practice and discourse (2013–2018) and Pardon my French? Dutch-French language contact in the Netherlands, 1500–1900 (2018–2023). What do we learn from these large-scale projects which address different research questions and focus on different periods in the history of Dutch? The use of specific sources, handwritten material such as ego-documents, a multi-genre approach and details of corpus compilation will be discussed. The various approaches and results are considered against the background of methodological developments and current debates in historical sociolinguistics. I argue and conclude that the careful compilation of specialised corpora remains essential as a solid foundation for historical sociolinguistic research.