Plus qu’une simple poursuite des questionnaires philologiques actuels par d’autres moyens, la philologie computationnelle peut, dans certains contextes, amener des déplacements ou des transformations dans les paradigmes de recherche en sciences des textes. Sans sortir du paradigme évolutionniste qui remonte aux origines de la discipline, la philologie peut ainsi reprendre à son compte des approches méthodologiques articulant démarche de test d’hypothèse, modèles et analyse de données. La massification des données permise par l’intelligence artificielle ouvre des perspectives d’analyse nouvelles, propice aux études macrostructurelles, comparatistes ou de longue durée.

A century has passed since Joseph Bédier observed the ‘bifidity’ of textual evolutionary trees (stemmata). A fundamental debate has developed around this observation, opposing the Bédiérist thesis of a methodological bias with that of law resulting of the process of text transmission. This article, focusing on the langue d’oïl domain, returns to this question of the ‘evolution’ of textual traditions, which is shown to be related to a more general problem specific to all the trees traced by evolutionary disciplines, that of ‘unbalanced trees’, well known in the Darwinian paradigm. I discuss how philology can be understood as a science of the evolution of texts, and how its fundamental problems can profitably be tackled with the support of computational methods. These methods can be used to test hypotheses, by combining modelling and data analysis, and by taking advantage of interdisciplinarity. The resources of artificial intelligence can also be mobilised to base these approaches on increasingly vast corpora, encouraging studies on a scale not previously envisaged.