The site of Qdeir 1 (Fig. 2) is situated in the El Kowm oasis in northern Syria (Fig. 1), and was first located in 1980 by Olivier Aurenche, with subsequent excavations by the ‘Mission d’El Kowm-Mureybet’ directed by Jacques Cauvin (1981), revealing a Final Pre-Pottery Neolithic B [PPNB] settlement dated 7100–5720 cal BC (Stordeur, 1993 and Stordeur, 2000; Table 1). Aurenche's survey demonstrated that the El Kowm oasis contains a number of prehistoric settlements (Fig. 1), including El Kowm, Qdeir 1 (Stordeur, 2000) and Umm-el-Tlel (Molist et al., 1992). A larger-scale excavation was initiated in the late 1980s, initially directed by Danielle Stordeur (in 1989, 1991 and 1993), and subsequently by Frédéric Abbès (in 1999, and 2001–2003). The site is estimated to be c. 2000 m², of which some 200 m² have been excavated (Abbès, 2015). While El Kowm and Qdeir — 8 km apart — were both occupied during the Final PPNB, they displayed sufficient occupational differences to warrant attribution to two distinct cultural facies (Cauvin, 1994). While El Kowm was a village with a long and complex settlement history, with at least five different levels of occupation spanning 7100–4230 cal BC (Stordeur, 2000), Qdeir 1 comprised an exceptionally well preserved nomadic camp, consisting of only a few architectural structures, a semi-permanent camp, and a workshop for the manufacture of flint implements (Abbès, 2015, Besançon et al., 1982 and Stordeur and Wattez, 1998). A total of four occupation phases have been defined (Ibid.; Gourichon, 2004), two with architectural structures (isolated houses 1 and 2; phases II and IV), and two without (phases I and III). These characteristics led Qdeir 1 to be classified as belonging to some specific facies of the PPNB, the ‘desert PPNB’. While distinct in size, location, and nature of occupation, these three sites have all produced evidence for the use of obsidian, an exotic raw material whose closest sources lay hundreds of kilometres to the north in central and eastern Anatolia (Fig. 1); this material was mainly recovered in the form of pressure blades and bladelets (Cauvin and Chataigner, 1998).

Part of the Qdeir 1 chipped stone assemblage was studied and published in the 1980s, albeit with modest attention to the obsidian itself, due to the limited number of artefacts found at this point (Calley, 1986; see also Cauvin, 2000).

Our study involved the analysis of each artefact that was available to us (the remaining pieces being in Syria), i.e. the 519 artefacts coming from the excavation campaigns of 1979–1980, 1989, 1991, 1993, and 2000. While exact figures were unfortunately not available to us (nor likely to be accessible in the near future given the political situation in Syria at the time of writing), the relative proportion of obsidian to flint within the chipped stone assemblage is known to be quite low (F. Abbès, pers. comm.). The volume of worked flint per cubic meter has been estimated at > 12,000 pieces, equivalent to 22 kg of raw material (Calley, 1986). Most of this flint appears to have been procured from an outcrop 4 km north-west of Qdeir, an easily accessible source that provided raw nodules of excellent quality (Calley, 1986). The flint was used primarily for the manufacture of bi-directional blades knapped from opposed platform/naviform cores (Calley, 1986; see also Abbès, 2003), a particularly productive technology, with estimates of 26/27 blades and 13/14 bladelets per core (Calley, 1986). Perhaps unsurprisingly, given the proximity of the source, the flint assemblage contains a significant quantity of early-stage reduction knapping debris, with a relatively high proportion of cortical waste (29%, Ibid.). As we document below, it is in stark contrast to the obsidian assemblages, which are dominated by end-products and late-stage manufacturing pieces. The relative proportion and range of modified flint implements is greater than that witnessed in the obsidian assemblage, with 31 arrowheads, 156 burins, 108 scrapers, and ten sickle blades reported from the 17,642 artefacts of the GY and GX surveys sectors from the 1980 excavation (Aurenche and Cauvin, 1982). Flint and obsidian seem to be the only raw materials employed by the community to manufacture flaked stone implements, with the exception of one chisel blade executed on greenstone (Aurenche and Cauvin, 1982).

In our study, we used two non-destructive geochemical characterisation techniques that have been widely employed for obsidian sourcing in the Mediterranean and Near East, namely SEM–EDS (cf. Acquafredda et al., 1999 and Le Bourdonnec et al., 2010) and ED-XRF (cf. Carter and Shackley, 2007, Carter et al., 2013 and Poupeau et al., 2010). The operating conditions for each method are detailed in an earlier paper (Orange et al., 2013); we present here instead a discussion on the repeatability and accuracy of our measurements (concepts discussed in Frahm, 2011).

The Qdeir 1 obsidian assemblage thus contains artefacts made of raw materials from both Cappadocia (Göllü Dağ, Nenezi Dağ) and eastern Anatolia (Bingöl B, Bingöl A [Fig. 9]), with the majority coming from the slightly closer Bingöl region (c. 450 km, compared to c. 520 km to the central Anatolian sources). This ‘dual access’ to both central and Anatolian raw materials is something we have come to associate with later PPNB communities of the northern Levant/Upper Mesopotamia (Cauvin and Chataigner, 1998 and Delerue, 2007).

While to an extent we have reaffirmed previous ideas concerning Final PPNB obsidian ‘diffusion zones’, we also provide new important information as to the form in which these raw materials circulated and how they were subsequently consumed. Our study shows that these four raw materials were worked in essentially the same manner, used to make fine prismatic blades by a skilled pressure technique, with the same modes of core preparation and reduction, leading to the end-products having much the same size and form. Only with the Bingöl B raw material do we have evidence for more than one mode of production, with the pressure technology augmented by a handful of soft hammer percussion blades, and a single larger blade from a skilled percussion opposed platform technique. The standardised nature of the unipolar pressure blades (all materials) is attested by the great regularity of the finished products, and the high ratio of 2-1-2’ blades, indicative of a highly skilled practice involving careful control of core preparation and the subsequent knapping sequence. As has been suggested at the nearby, but slightly earlier (middle PPNB) site of Tell Sabi Abyad II (Fig. 9), we might also be dealing with the products of specialist craftspeople (Astruc et al., 2007), although here at Qdeir 1 we do not have evidence for the use of crutch and small lever pressure debitage (Abbès, 2013). As noted previously, only a few obsidian core preparation elements were found in Qdeir (all materials considered), and only one cortical flake was reported. Thus, the first stages of the knapping sequence appear to have been undertaken elsewhere, which leads us to wonder which aspect of this industry would have been most valuable to these people — the exotic raw material (obsidian), or the restricted technical know-how (pressure technique) with which to work it? In sum, while the Qdeir 1 obsidian assemblage contains four different raw materials, there is no reason whatsoever to believe that they were procured via four distinct exchange networks. Instead, it appears far more likely that an intermediary population, perhaps in the Middle Euphrates or further to the north-west (discussed below), were responsible for procuring the different raw materials, and then consuming them within the same pressure-blade tradition. By extent, whomever was responsible for working the obsidian at Qdeir 1, be that a visiting skilled knapper, or a local community member, they too then continued to reduce these cores in the same manner.

Our first point of comparison is with other assemblages from the surrounding sites in the El Kowm basin, namely the El Kowm 2-Caracol village and Umm-el-Tlel 2. At the former site, we also have a mix of obsidian from both eastern and central Anatolian sources, mostly in the form of end-products, again unipolar pressure blades and bladelets (including corner-thinned examples), plus some knapping debris and a single core which appears similar to the example from Qdeir 1 (Cauvin, 2000). At Umm-el-Tlel 2 (Borrell et al., 2013 and Molist and Cauvin, 1990), some pressure blades were found but no cores; with only Bingöl B obsidian thus far attested from the site (Gratuze et al., 1993). The fact that El Kowm — the village site — appears to have a more restricted quantity and range of material (no cortical debris for example) than at the nomadic campsite of Qdeir 1, might be taken as evidence to support the theory espoused by V. Gordon Childe in the 1930s that nomadic groups may have ‘gravitated’ around sedentary villages, supplying them with various non-local products (Childe, 1934). That said, it should be noted that excavations at El Kowm have thus far been limited in nature, and that the known flint workshops have not yet been excavated. Moreover, in Qdeir the excavation technique allowed all microdebitage to be collected, whereas in El Kowm this was not the case (F. Abbès, pers. comm.).

If we then consider the Qdeir 1 data in its Final PPNB broader context, then one can point to what appears to be very similar sets of material (raw material and technology) from not only El Kowm and Umm-el-Tlel, but also Abu Hureyra (pers. obs.) and the slightly earlier (Middle PPNB) Tell Sabi Abyad II to the north (Astruc et al., 2007). With obsidian representing only a minority component of each of these sites’ chipped stone assemblages, and the fact that the material was often being consumed in a distinct manner to how flint was being worked (at Qdeir 1 the flint assemblage essentially relates to a naviform/opposed platform percussion blade technology [Calley, 1986]), one has to consider the idea that these obsidian assemblages could have been produced by one or two non-residential itinerant specialists (cf. Perlès, 1990). One might argue that these itinerants were actually members of the nomadic group who camped at Qdeir 1, perhaps occasionally leaving the group to travel north to access the obsidian quarries of Anatolia, and/or interact with those individuals who could access the distant materials for them.

The fact that at Qdeir 1 — a nomadic campsite — we have evidence for a greater range of knapping debris than its near neighbours suggests that this mobile community may have represented the means by which obsidian was circulated in this local area. While El Kowm has a certain amount of production debris, the Umm-el-Tlel assemblage is comprised exclusively of end-products. Here we might claim that the former community may have represented a more ‘attractive’ locale for the itinerant specialists to visit, whereas the latter populace only received their tools second-hand, with perhaps a reflection of local power relations and dominance in this arrangement. In theorising ‘attractiveness’, one might think of the political capital to be gained by an individual/sodality in their ability to attract a character such as the itinerant obsidian worker to their village, the latter potentially an exotic figure of know-how and experience (based on their skilled crafting abilities and their regional travels [cf. Carter, 2007: 100–102; Helms, 1993]). These skilled knappers might be viewed as the minstrels or rhapsodes of their time, the teller-of-tales, relaying cultural traditions as much as they were tool-makers. Their appearance at a village might be seen in terms of a performance (with the all too important host(s) in association), perhaps appearing on a cyclical basis, or drawn to the community to participate in a social gathering/festival surrounding such important occasions as marriage ceremonies, the meeting of far-flung trade partners, or the founding of a lineage's new house.

Further traces of these inter-community networks can be viewed through reference to the specificities of obsidian consumption, as for example evidenced by the common tool-type of the ‘corner thinned blade’. These distinctive implements, defined as blades that have deliberately had their corners thinned by retouch initiated from the end (often a medial break), and interpreted as sickle elements (Nishiaki, 1990), are well known from later PPN — Pottery Neolithic sites of the northern Levant/Upper Mesopotamia (Nishiaki, 2000: 205–209, Fig. 8.15). The Qdeir 1 examples (which are mainly made on Bingöl B blades, followed by a few Bingöl A and Göllü Dağ examples) provide a clear link between this community and such Final PPNB contemporaries as Abu Hureyra, Tell Sabi Abyad, Tell Kashkashok II, and Bouqras inter alia (Fig. 9). While these tools tend to be made of eastern Anatolian raw materials (exclusively at Bouqras), at both Abu Hureyra and Qdeir 1 there are a few examples made of Göllü Dağ obsidian (Bingöl B is the main material employed at our site, followed by Bingöl A). In sum, our integrated characterisation study indicates closely shared traditions in technical/raw material choices between these sites and suggests a close level of socio-economic interaction amongst these populations, underwritten perhaps via trade partnerships and inter-marriage, in essence what scholars have recently begun to describe as socio-techno ‘communities of practice’ (for a related discussion of this theme see Carter et al., 2013).

The northern Levant/Upper Mesopotamia is often viewed as a world whose cultural traditions and technical practices were linked to various surrounding regions, its communities’ use of both central and eastern Anatolian obsidians taken to be evidence for these peoples’ connections with their neighbours to the east and west (Delerue, 2007). Our Qdeir 1 data might indeed be seen as proving links to the east through the presence of a bullet-core which represents a clear connection to the technical traditions of populations in the eastern wing of the Fertile Crescent in the Jezireh and Zagros regions (Kozlowski and Aurenche, 2005: 23, fig. 1.3.5), while wedge-shaped nuclei were the norm in regions to the west (cf. Özdoğan, 1994: 271, fig. 1). Moreover, the Qdeir 1 pressure blades are also much the same size as those from Bouqras (10–12 mm in width), a site that has also produced the same scraper type (Stordeur, 2000).

Links with communities to the south and west are arguably less clear. The claim is usually based upon the common consumption of Cappadocian obsidian by later PPNB northern Levantine/Upper Mesopotamian communities (e.g., Abu Hureyra, Akarçay, Dja’de, Halula, Kerkh II, Ramad), and those living on the Levantine littoral (e.g., Byblos and Ras Shamra [Fig. 9]), southern Levant, Cyprus and central Anatolia (Delerue, 2007, fig. 320). This however is to focus alone on raw materials, an approach we above argued to be reductionist and problematic; the common use of a particular obsidian need not infer any connection between these communities, as there are many different ways in which these raw materials could have been procured and worked by contemporary populations (see Carter et al., 2013). A claim could be made for regional links with regard to shared technical traditions, with Göllü Dağ obsidian also being used to make pressure blades in the southern Levant at this time (Gopher et al., 1998: 645); conversely, with the pressure technique not being introduced into central Anatolia until the mid-seventh millennium BC (Carter, 2011: 12), we have an example of where shared raw material use likely does not indicate inter-regional connections.

In this paper, we focused on a synchronic vision of the obsidian economy at Qdeir 1; allowing us to improve our understanding of this raw material's consumption by the local community and by extent to further our knowledge as to the use history of various Anatolian obsidian sources during the Final PPNB.

In developing this study, the next obvious step would be to take a diachronic perspective (as at Çatalhöyük for example, see Carter et al., 2006), Qdeir 1 having four occupation phases within the Final PPNB (Stordeur and Wattez, 1998). In turn, if one had access to detailed absolute dates for the periods under consideration, volumetric data of the deposits dug in these phases, plus the counts and weights of obsidian, then one could attempt to reconstruct the amount of obsidian entering the site on an annual basis (cf. Cessford and Carter, 2005), which might further help us understand the value regimes surrounding the circulation and consumption of these exotic materials.

A further line of enquiry would be to turn our attention to the other northern Levantine/Upper Mesopotamian sites discussed in this paper (Fig. 9). At present the publications available on these sites’ obsidian assemblages do not provide the integrated form of characterisation study that we provide here, whereby most of our discussion has had to focus on patterns of raw material diffusion. However, the use of a common raw material is insufficient evidence to support claims of common cultural traditions and inter-community contact (Carter et al., 2013). Instead, what is required is a more ‘thick description’ approach that incorporates not only typo-technological considerations (and specificities of how items were modified, as for example with the corner-thinned blades), but also metrical data, such as mean width of prismatic blades. Such a study is in fact currently being undertaken by one of us (TC), with the analysis of 300 artefacts from Abu Hureyra (the data providing many parallels with Qdeir 1), whose results will contribute further to our understanding as to the relationships between these communities.