Révision critique du concept de débitage Levallois à travers l'étude du gisement moustérien de Saint-Vaast-la-Hougue/le Fort (chantiers I-III et II, niveaux inférieurs) (Manche, France)

The preliminary study of Saint-Vaast-La-Hougue's lower levels, conducted by Gerard Fosse in the early 1980s, ascribed the series to a notched and denticulated Mousterian with a slight degree of Levallois débitage. The analysis of the second site lithic assemblage, carried out within the context of a master's degree (Guette, 1999), allowed the revision of the previously proposed diagnosis. While the débitage seems to have been affected by the morphology of the available poor-quality pebble flints, and the barely restructured core convexities lead to lower productivity, the process appears to be verging on the Levallois predetermination concept even though some technical criteria seem to be lacking. These observations have led us to revise the Levallois concept and reaffirm the over-restrictive aspect of Eric Boëda's definition, which used a particular case to define a general rule, giving less importance to Levallois variability. The convexities notion, a technical criterion meant to guide the fracture wave, can actually be fulfilled in many ways. The knapper can not only intervene directly on them but can also either prepare them or exploit the natural morphology of the support. In the light of this discussion, we tackled the series of the first and third site of Saint-Vaast-la-Hougue. The unipolar method prevails, and requires a minimum of surface reshaping. Considered separately, some cores seem far from the Levallois débitage, even though they all depend on the same conception. The Saint-Vaast-la-Hougue site is particularly meaningful in this matter since it shows the different methods within the same lithic assemblages.