Side scrapers (alternatively known by the French term racloirs) have always been seen as one of the most distinctive retouched tool forms in Middle Palaeolithic industries and in the earlier literature were often regarded as a diagnostic feature of these industries (e.g. de Mortillet 1883). In fact, typical side-scraper forms, effectively identical to those encountered in the Mousterian and related industries, are known to occur throughout the greater part of the Lower Palaeolithic sequence extending back to the earlier Middle Pleistocene. Nevertheless, high frequencies of typical side-scraper or racloir forms are one of the most pervasive features of typically Middle Palaeolithic technology not only in Europe but in most parts of Africa and in at least the western parts of Asia (Klein 1989a).
In their most basic forms, side scrapers are very simple, almost elemental tools, characterized by two basic features (Bordes 1961a; Mellars 1964): first by a single, major retouched edge almost invariably located along one of the longest margins of the original flake blank; and second by the application of retouch obviously intended to produce a regular, sharp working edge along the retouched margin of the tool (Figs 4.1, 4.2). All the microscopic use-wear studies which have so far been carried out on typical side-scraper forms confirm that the zones of use were confined largely to these retouched edges (Semenov 1964; Keeley 1980; Beyries 1986, 1987, 1988a,b; Plisson 1988; Anderson-Ger-faud 1990). The specific functions of the tools seem to have been more variable, ranging from use as either cutting or slicing tools
(variously applied to wood, meat or skin) to more heavy-duty scraping of hide or bone. Clearly, the functions of these tools were much more flexible than their conventional English and French names would imply.
When defined in these terms there is scope for relatively wide variation in both the precise forms of side scrapers and in the character and treatment of the retouched edges. The clearest reflection of this can be seen in the complex system of typological divisions for various racloir forms proposed in the standard typology of François Bordes (1961a). Leaving aside the more complex forms of convergent and déjeté racloirs, Bordes advocated 17 separate divisions for the simpler racloir forms (see Table 6.1). The major distinctions in Bordes' system are based on the position of the retouched edges in relation to the main flaking axis of the tool and on the overall shape of the retouched edge itself. Thus a basic distinction is made between pieces where the retouched edge is aligned parallel to the main axis of the flake (defined as 'lateral racloirs') and those where the retouched edge is oriented transversely across this axis ('transverse racloirs'). Further subdivisions of these types are based on the forms of the retouched edges - whether straight, convex or concave. In addition to these basic forms Bordes recognized separate categories of tools retouched either exclusively or partially on the ventral surface of the flakes (racloirs sur face plane or a retouche alterne) and other forms with extensive retouch on both the dorsal and ventral surfaces (racloirs a retouche biface and racloirs a dos aminci). Tools with characteristic retouch on two separate edges are grouped together collectively as double racloirs, and further subdivided into six categories based on various potential permutations in the curvature of the two retouched edges (see Table 6.1; Figs 4.1,4.2).
From these distinctions it is clear that the manufacture of even relatively simple racloir forms left Middle Palaeolithic artisans with much flexibility in their choice of both overall design (i.e. the number and location of the retouched edges) and the precise patterns of retouch applied to the worked edges. Recent debates on the significance of these formal variations in racloir morphology have focused on two basic issues:
1. How far was the retouch applied to racloirs a deliberate feature applied in the initial stages of tool production and how far was it simply a result of ad hoc resharp-ening and reworking of the edges as they became progressively worn and blunted in the course of use (Rolland 1977, 1981; Dibble 1984a, 1987a; Rolland & Dibble 1990; Dibble & Rolland 1992)?
2. What significance can be attached to docu mented variations in the forms of the tools - either in terms of function or specific design norms or 'mental templates' that lay behind the conception and production of the tools (Dibble 1987a/b,c, 1989; Dibble & Rolland 1992; Kuhn 1992a)?
As discussed earlier, the question of tool reduction and resharpening has emerged as a central issue in Middle Palaeolithic technology with radical implications not only for the morphology of individual tool forms but also for the more general issue of inter-assem-blage variation in the Mousterian (see Chapter 10). In a series of publications over the past 15 years Nicholas Rolland and Harold Dibble have presented a stark alternative to conventional perspectives on side-scraper typology by suggesting that these tools may never have been planned as retouched tool forms but may simply have acquired their characteristic retouch in the course of progressive resharpening of their edges during use (Rolland 1977, 1981, 1988a, 1990; Dibble 1984a,b, 1987a, b,c, 1988a,b, 1989, 1991a,b; Rolland & Dibble 1990; Dibble & Rolland 1992). They envisage that the great majority of conventional racloir forms started their use-lives simply as unretouched flakes, which were only systematically retouched as their originally sharp edges became progressively worn and damaged through heavy use. The logical extension of this is that effectively all the documented variation in the relative frequencies of racloirs in different Middle Palaeolithic assemblages can be attributed largely to variations in the degree to which these tool-resharpening processes were carried out in the different sites. Thus taxonomically Denticulate Mousterian industries, in which the overall frequency of racloirs is low, are seen as lightly reduced industries in which few raw flakes were transformed into retouched racloir forms, while the various Ferrassie and Quina-type industries (in which overall racloir frequency is high) are seen as heavily reduced indus-
Figure 4.3 Resharpening flakes detached from the edges of racloirs or similar tools, from the penultimate-glacial levels of the Cotte de St Brelade (Jersey). After Cornford 1986.
tries in which a large proportion of the available flakes were transformed into systematically resharpened racloir forms (Roll-and 1977,1981,1988a; Rolland & Dibble 1990; Dibble & Rolland 1992).
These patterns in turn are thought to be closely related either to environmental factors (reflecting the relative abundance and accessibility of local raw material supplies) or to varying degrees of duration and intensity of occupation in the different sites (Dibble & Rolland 1992; Dibble 1991a,b). These issues will be discussed in more detail in Chapter 10 in the context of more general models of inter-assemblage variation in the Mousterian. The assumptions of these models, however, are directly relevant to the present issue of how far tool-reduction models can be invoked to account for the patterns of production of racloir forms.
There is no doubt that Rolland and Dibble have identified an important potential aspect of variation in Middle Palaeolithic technology which is in line with the findings of much recent research. Recent results of microscopic use-wear studies, for example, leave no doubt that many specimens of unretouched flakes in Middle Palaeolithic industries were subjected to heavy use, apparently for a range of different functions. Amongst others, the studies carried out by Sylvie Beyries on the assemblage from Corbehem in northern France (Beyries & Boeda 1983) and similar studies by Patricia Anderson-Gerfaud (1990) on assemblages from several southwestern French sites leave no doubt about this (see also Beyries 1986,1987,1988a,b, 1990; Keeley 1980). There is also now clear evidence from several sites that the retouched edges of typical racloir forms were systematically resharpened by the removal of deliberate resharpening spalls (Fig. 4.3). Typical specimens of these resharpening flakes have been recovered, for example, from Combe Grenal (Lenoir 1986), Marillac (Meignen 1988), La Cotte de Saint-Brelade (Cornford 1986) and La Micoque (Schlanger 1989), and in some cases have been refitted directly to the parent tools (e.g. Fig. 4.4). These discoveries leave no doubt that systematic resharpening of blunted or damaged edges of side scrapers was carried out commonly on a range of Middle Palaeolithic sites.
The central issue here is the relative scale on which resharpening was carried out in Middle Palaeolithic sites. Can we use these resharpening models to argue that typical side-scraper forms were rarely if ever manufactured as deliberate, a priori tool forms (as Rolland and Dibble imply) or was resharpening applied purely as a secondary technological device to extend the use-life of tools conceived and produced from the outset as retouched tool forms? The main points are as follows:
1. The evidence from many sites does not support easily the notion that systematic reduction of raw flakes into characteristic racloir forms was a product of either intensive patterns of site occupation or the scarcity of local raw materials. There are now several examples of sites where high frequencies of typical racloirs coincide with locations where local raw materials were both of high flaking quality and apparently available in almost limitless supplies. The most obvious example is the site of Combe Capelle, where a major outcrop of high quality flint occurs directly on the occupation site (Peyrony 1943; Bour-gon 1957). In this, as in other sites such as Champlost in Burgundy (Farizy 1985) and Biache-Saint-Vaast in northern France (Tuf-freau & Sommé 1988), there would have been an ample capacity for producing fresh, unretouched flakes for tool use, without any need for the intensive resharpening of flakes as these became worn and damaged through use. Why it should have been necessary in these contexts to apply highly 'economizing' strategies in the use of available flint supplies is by no means clear.
2. Similarly, there are now many sites where a high overall frequency of racloirs (in rela tion to other tool forms) can be seen to coincide with an equally high frequency of unretouched flakes. At Biache-Saint-Vaast, for instance, an industry comprising over 60 percent of retouched racloir forms occurs in the context of an assemblage which is dominated by unmodified Levallois flakes (Tuf-freau and Sommé 1988). A similar pattern is reflected at Champvoisy (Marne: Tuffreau 1989b) and again at Champlost in Burgundy (Farizy 1985: 406). In these and other cases it is difficult to see why intensive retouching and resharpening of tools should have been needed when there were large numbers of unretouched flakes immediately accessible and readily available for use on the occupation sites.
3. Even more problematic in terms of the racloir-reduction models is the available data on the relative sizes of side scrapers in different types of Mousterian assemblages. For example, several studies have shown that in assemblages comprising the lowest overall percentages of racloirs (i.e. those of the Denticulate variant) the average lengths of the side scrapers present in the assemblages are shorter than those recorded in some of the most heavily reduced industries, such as those of the Ferrassie and Quina Mousterian variants (Fig. 4.5; Table 10.1). This has been clearly documented, for example, by Rolland himself for the various levels of Denticulate, Quina and Ferrassie Mousterian at Combe Grenal (Rolland 1988b: 173, Table 9.4B) and is equally apparent in similar assemblages from the Abri Chadourne, La Quina, Hauteroche, Arcy-sur-Cure and elsewhere. Exactly how these observation can be reconciled with the hypothesis that the extent of systematic resharpening and reduction of tools is actually much greater in the Quina and Ferrassie industries than in those of the Denticulate variant (i.e. how systematic reduction can somehow make the tools larger) has yet to be explained.
Figure 4.5 Mean lengths of racloirs recorded in various levels of Ferrassie, Quina and Denticulate Mousterian at Combe Grenal, as documented by Rolland (1988b: Table 9.4b). The larger sizes of the racloirs recorded in the Ferrassie and Quina Mousterian levels clearly conflicts with the notion that these are heavily 'reduced' versons of the tools encountered in Denticulate industries.
Was this article helpful?