2 Large blade from Çayönü Tepesi, CV building, Cell Building sub-phase 3 (Late PPNB) 

Contexts in source publication

Context 1

... A nearly complete blade, 28.5 cm long, was found broken in three main fragments (Fig. 5.2 , ÇT S3-1 CV building/cell 3). The mesial and distal fragments of the blade have been previously described (Binder 2005 : Fig. 5), but the proximal part was only recently identifi ed and refi tted. The original length of the blade may have been as much as 33 cm, if we estimate that 5 cm are missing from the present distal end, which is uncurved and measures 24 mm in width and 3.8 mm in thickness. The proximal section is 31.9 mm wide and 8.4 mm thick, and the mesial section is 29.5 by 5 mm. The regularity of the blade’s edges and arrises (Inizan et al. 1995 ) is impressive, and its thinness slightly decreases towards the distal end. The profi le is moderately curved without infl exion or undulation. These characteristics are consistent with pressure blade production using a lever. The blade is four-sided in the proximal section, the code of débitage being 4321 (Binder 1984 ) , but it becomes trapezoidal and asymmetric (321) in the medial section. A long scar measuring more than 68 mm long, together with rather hinged short scars, is evidence of core preparation to detach this large blade (Fig. 5.8 (1)). The butt is small, ovoid and dihedral (7.8 mm wide and 2.1 mm thick). The butt shows a tiny inclination to the left edge, and its edge angle is greater than 90° (approximately 95°). The pressure point is located on the dihedral, defi ned by two tiny fl ake scars. The lip is clearly developed, and the absence of any cracking or damage suggests the use of a pressure stick armed with an antler point to detach the blade. • A large proximal fragment, 17.2 cm long, comes from a blade that was probably 20–25 cm in length (Fig. 5.3 (1), ÇT70 R2-10/4 CV building/cell 3). It is as wide as the previously described specimen (32 mm) and somewhat thicker (7.8 mm under the bulb, decreasing regularly to 6.4 mm at its mesial break). Its ventral face is perfectly regular, without any undulation, and the profi le is almost straight (Fig. 5.8(2)). The butt is small (8.8 mm wide and 2.8 mm thick), with an oval and slightly concave surface that bears two tiny fl ake scars, probably produced by pressure, giving a platform angle of 80°. The detachment of the blade ...

Context 2

... from the distal half of the blade blank. A slight undulation of the dorsal side and arrises is mirrored on the ventral side. The overall regularity and slight curvature testify to a pressure technique, very probably with a lever, given that the mesial section of the blade is larger, about 28–30 mm. From an open area in sector E04 (6550–6500 cal B.C., Fig. 5.7 (2)), a mesial fragment 4 cm long with a distal inverse notch comes from the distal half of a large blade (the width decreases from 20 to 17 mm, and it is 4 mm thick). The regularity and symmetry of the section suggest that the fragment is that of a central blade detached from a very well-treated pressure core, possibly using a lever. • In sector E03 (6750–6600 cal B.C., Fig. 5.7 (3)), a mesial fragment of a very regular obsidian blade was recovered. Truncated by an inverse notch at both ends, it is 6.3 cm long, 31.4 mm wide, and 6 mm thick. The remarkable regularity of the edges, the arrises and the ventral side, and the wide width of the blank suggest that the blade blank was detached by pressure using a lever. The detachment of blades by the pressure technique is characterized by regularity, reduced curvature and thinness (Pelegrin 1988 : 48; 2003 : 63; Tixier 1984 : 66). Indeed, the mechanical conditions of a pressure technique, immobilization of the core, permanence of the compression along the fracture propagation and absence of shock, which would generate vibrations and therefore undulations, are the only means of detaching such a regular and fragile column of volcanic glass. The blades presented here bear the scars of two to four previous removals that are also highly regular, implying a very controlled and repeatable mechanism of detachment. We made careful experiments on obsidian both using indirect percussion and pressure (standing pressure and pressure with a lever), and this after years of experience of these techniques with fl int as a raw material (Pelegrin 2002a ) . Obsidian blades can be detached in series using indirect percussion, but they are far to be as regular as pressure blades (Pelegrin 2000, 2003, 2006 , this volume). In this respect, we fully share Crabtree’s opinion ( 1968 : 459) that ‘the impact from the percussor causes excessive undulations and waves on both the core and blade; the dimen- sions of the blade cannot be controlled with regularity; the bulbs of force are much too large, and the curve of the blades and termination of the ends cannot be con- trolled’ (see also Figs. 5.4 , 5.8 ). In addition, the fragility of obsidian leads to a high rate of proximal breaks when trying to produce relatively thin blades. These proximal breaks, which are rarely produced by pressure detachment, occur even more frequently with the use of indirect percussion than with direct percussion. They clearly occur during the detachment itself (and not after, as do simple medial breaks) because they produce distal ripples and hinged termination of the blade, thus spoiling the regularity of the distal end of the core. The extreme sensitivity of obsidian to breakage explains why, beyond 12–15 cm in length, irregularity of curvature and termination seems inevitable, even when using an elastic support for the core (which has a regulating effect on the detachment of fl int blades) (Pelegrin 2000, 2002b, 2003 ) . There are two practical ways to produce large blades by pressure: using the full weight of the body transmitted by a crutch in a standing position and using a lever. During a recent colloquium held at Pennsylvania State University (Hirth 2003 ) , some of the most experienced specialists agreed that more than length, it is the width of a blade that is dependent on the force of the pressure, as Crabtree ( 1968 : 468) stated: ‘the wider the blade, the greater the amount of pressure that is required’. In working fl int, for example, the maximum width of pressure blades detached using a relatively long crutch placed at belt level by a person in a standing position can reach about 20 mm when using an organic (antler) pressure point and even 21 or 22 mm when using a copper pressure point (harder than antler, copper helps to detach thicker butts). Blades with these maximum widths have been observed in different archaeological contexts (Pelegrin, this volume). With obsidian, our attempts at using the standing pressure crutch technique produced blades with widths of up to 26 mm, confi rming an earlier observation of ours that obsidian could yield blades which were 30% wider than fl int, using an identical technique and level of effort (Pelegrin 1988 , see also Kelterborn, this volume). Crabtree ( 1968 : 468) concluded that the maximum size of the obsidian blades that he could produce using his standing technique was ‘1 in. wide and 8 in. long’, while the ‘Mexica’ technique reconstructed by Clark and replicated by Titmus could be used to detach blades up to 24 mm wide (Titmus and Clark 2003 ) . The width of the almost complete blade from Çayonü (Fig. 5.2 ; 31.9 mm) is clearly larger than that which can be achieved with the standing pressure technique; a more powerful device had to be used to detach the blade, one which involved the use of a lever, such as the one we used in our experiments (Pelegrin, this volume). Our analysis of the proximal portions of four large obsidian blades found at Çayonü Tepesi (Figs. 5.4 (3), 5.8 ) indicates that the point used for their detachment was probably made of an organic material. Three of the detachment butts are ovoid and plane, the fourth one is ovoid and dihedral; the clear lips and the absence of cracks on the butts favour an organic point, probably antler (experiments from Pelegrin in Astruc et al. 2007 ) . This is even more apparent for the fourth butt: the point of pressure is located on the dihedral which did not suffer of any damage that would be caused by a copper point. At Sabi Abyad I, a proximal fragment of a large blade has been found at the surface of the Tell, in the operation 3 area. Its ovoid, plane butt is similar to those of Çayonü Tepesi’s large blades. The analysis of the large blades of Çayönü Tepesi and Sabi Abyad I brings a new perspective to lithic specialization within Neolithic communities in the Near East. The production of large blades using a lever occurred as early as the second half of the eighth millennium cal B.C. at Çayonü Tepesi, likely between 7340 and 7080 cal B.C. This is the earliest evidence of this remarkable technique. It was thus testifi ed in the Balikh Valley a thousand years later, between 6100 and 6500 cal. B.C. The production of large obsidian blades demonstrates a remarkable level of technical specialization for these early periods. Pressure detachment with a lever was a technique likely practised by a few highly qualifi ed specialists, who were possibly already fully trained in the standing pressure technique. To carry out this type of blade production, successive choices had to be made in order to reach the optimal exploitation of both raw material and technical investment and to avoid accidents that would lead to the waste of several blades or of the entire core. Risk levels associated with the various techniques would have been under constant evaluation, and substantial experience in pressure blade production would have been necessary to develop and control the whole production system, to manufacture the tools and to control the numerous practical details or adjustments. Experimental research (Pelegrin 1988 ) has shown that the technical knowledge needed to produce medium-sized blades by standing pressure is considerable. However, the necessary expertise is much greater when the goal is to produce a standardized series of long blades. At both sites, the lengths of the nearly complete blades – 27.2 cm at Çayönü Tepesi and 28.6 cm at Sabi Abyad – allow us to estimate the length of the original cores as 32 cm or more. A very high level of understanding of the mechanical properties of obsidian is necessary to shape such huge cores and to produce these large, wide blades. The initial core preparation has to be of very high quality, as any irregularity on the production surface will have a direct effect on the regularity of the ventral sur- faces and edges of the blades. Once the critical roughing out by stone percussion is fi nished (no deep or hinged scars are allowed), the next stage is a patient shaping using direct stone percussion or indirect percussion for the detachment of transversal fl akes, alternating from three to four axial crests; then the detachment of several large covering fl akes by direct percussion, using a hard wood hammer, and, fi nally, shaping the crests by a subtle direct or indirect percussion or even by pressure fl aking. The goal is to correct the volume that will be transformed into blades by defi ning the convenient convexities and avoiding any deviation – bumps or hollows – from an ideal of ±2 mm. Experimental reproduction by J. Pelegrin has shown that crested or under-crested blades (the fi rst series of blades which serve to remove the pre-shaped surface of the core) can tolerate such irregularities if they are broad and thick enough, without reproducing these irregularities on their scar or without becoming hinged. Diffi cult choices also have to be made when conducting the subsequent blade removal. The repartition of arrises on the core has to be strictly controlled, leading to different possible rhythms of débitage (convergent, divergent, inserted and adjacent unidirectional or alternating) (Astruc et al. 2007 ) . In this respect, it is crucial to realize that each blade detachment is anticipated not only to visualize the fi nal product but to control the effect of its removal on the geometry of the core. This requires meticulous attention to the preparation of each detachment not only to avoid accidents such as edge crushing, hinging and excessive plunging but to actually detach the expected blade with the most precision. ...