The project aims to gain a better understanding of what makes an oath a ‘true’, performative utterance, whilst taking into account its ritual and gestural dimensions. The research is structured around several themes. The project examines what renders an oath legitimate or otherwise. The political, social and judicial uses of the oath make it a common practice, which may serve as an argument in favour of its legitimacy in normative sources. However, over-reliance on it sometimes leads to perjury—which is utterly condemned—and may also contribute to its devaluation, or even to more stringent restrictions on its use. The legitimacy of the oath is also based on the framework governing the ritual and the words spoken, through the establishment of norms and formulae; a comparative study of these across various linguistic and religious contexts enables us to better identify what was likely to constitute a performative utterance. Normative sources are not, however, sufficient to document the medieval practice of oath-taking: other sources, whether diplomatic or normative, enable us to grasp the gap between norm and practice, even if they also introduce a bias. Generally speaking, the issue of bias arising from the transcription of speech—which is, by its very nature, oral—arises regularly; it can, however, be resolved through the cross-referenced study of different records.