Historical linguistics can resort to neither recordings nor interviews of native speakers. The only type of material that can be dissected is of a textual nature. This leads to two main issues: first, it has been widely observed in the literature that the choice of a certain register over another shapes the results of the investigation. 1 For instance, the syntax of a poem will offer different perspectives on a language than the syntax of a recipe. Second, and this is particularly relevant to French (amongst other languages), the written language often retains archaisms that are not found in the oral tongue. However, building a corpus of non-literary material may lead to consistent findings in diachronic studies. More precisely, the use of legal material is put forward here. This choice stems from other studies investigating different languages and is discussed here within the French historical context. The present article comes from a doctoral project that recently sought to create a digitalized corpus of unexplored texts in order to investigate the evolution of pronouns in the diachrony of French, and subsequently produce a quantitative analysis within the framework of generative linguistics. 2 Medieval France has produced a sizeable array of texts that allow researchers to carry out studies on the language’s history. The productivity of medieval scribes can be