Starting from Gaston Bachelard’s assumption that “all knowledge is an answer to a question”, the study proposes the use of comparative textual analysis to formulate research questions. A set of questions were derived via the TXM, Textométrie software, a tool for lexicometry and statistical analysis. Two historical documents on the Economic and Monetary Union (EMU) were examined, the Werner report and the Delors report (French versions). The “quest for questions” was based on the comparison of the documents, using the TXM specificities feature that highlights what properties are specific, as overuse or deficit, to a part versus the rest of a corpus. The documents were analysed both as entire units and as fragments (numbered parts and sections). The specificities were computed for the noun-adjective combination and parts of speech, the properties with specificity scores higher and lower than the TXM default positive and negative banality thresholds being selected for further enquiry and subsequently used to formulate research questions. Although further experiments, testing with other corpora and theoretical formalisation are required, the first results show that digital tools may serve not only as hypotheses or conclusions validators but also as means of discovering exploration paths to support interpretation and the construction of new knowledge in digital history.