The project will investigate the meaning of scale in historical writings, and more precisely how scale is expressed through language in historical discourse. This question draws attention to the conceptual and linguistic mechanisms at play in building historical knowledge, when the historian moves between different layers of analysis, narration or consulted sources, involving different degrees of generality. A small historiographical corpus, in which variations of scale are clearly present, will serve to develop the digital approach/tools/methodology. Depending on the findings of the project, an extension of the research to other types of corpora is envisaged. The paper presents a workflow and first experiments using topic modelling to analyse scale representation in historiographical data. Further experiments with more documents, other models and visualisation tools, as well as eventually creating a pipeline for semi-automatic restructuration of data as zoomable texts are also envisaged.