Retouched notches at Combe Grenal (France) and the reduction hypothesis

A detailed attribute analysis of notched flakes from upper layers of Combe Grenal, a key site in the debates about the nature of Middle Paleolithic assemblage variation in France, is used to evaluate the applicability to Combe Grenal of the reduction model offered by Holdaway, McPherron, and Roth as an explanation for morphological and size variation within notched specimens. We conclude that traditional implement types can be viewed as arbitrary divisions in a continuum of size and notch abundance, which can be explained by reference to a model of differential reduction. Specimens with greater numbers of notches are inferred to have been more reduced than specimens with fewer notches. Notch dimensions did not change as more notches were added, but complex notches were more commonly constructed in the middle of the reduction process rather than at the beginning or end. The location and abundance of notches were probably constrained by the size and shape of the flake blank: more notches were added to long flakes and to their distal end, indicating regular patterns of blank selection and treatment. These inferences can be used to suggest that flake blanks may have been an important factor in constraining the position, frequency, and perhaps even kind of notches that were placed on retouched flakes. We discuss implications of this proposition.