The Lithic Technologies of the Epipalaeolithic Hunter-Gatherers in the Negev, Israel
Implications from Refitting Studies
Les technologies lithiques des chasseurs-cueilleurs épipaléolithiques du Negev, Israël : Implications à partir d’études de remontages
p. 239-267
Résumés
In the course of intensive systematic surveys and excavations in the Western Negev Desert, Israel, dozens of Upper Paleolithic (c. 46-23 ka cal BP) and Epipaleolithic (23-11.5 ka cal BP) sites were investigated. Traditional lithic techno-typological studies enabled assignment of the assemblages to a series of distinct temporally and spatially well-defined sociocultural entities. Field observations indicated the potential for systematic refitting of the lithic assemblages. An extensive refitting program was conducted incorporating 30 assemblages with > 10,000 conjoins. This study focuses on reconstructing diachronic and synchronic changes in the chaîne opératoire of microlith production during the Epipaleolithic following refitting studies. The most significant contributions of this study are the realization that the Epipaleolithic cultural entities in the Negev all belong to the same conservative technological tradition, the Wide-fronted (W-fronted) approach to core exploitation, and that changes were gradual. Still, several diachronic and synchronic differences within the microlith production sequences between the industries were also observed. In contrast, typological changes were more rapid. Each cultural entity is distinguished by discrete morphological types of microliths that appeared over a relatively short time span of 1,000-2,000 14C years. Finally, this research provides further evidence supporting the validity of refitting studies as a significant methodological tool used to refine technological and typological classifications.
Au cours de prospections systématiques intensives et de fouilles dans le désert du Néguev occidental, Israël, des dizaines de sites du Paléolithique supérieur (c. 46-23 ka cal BP) et de l’Epipaléolithique (23-11.5 ka cal BP) ont été retrouvés. Les études traditionnelles typo-technologiques de leur matériel lithique ont permis leur attribution à une série d’entités socio-culturelles distinctes, bien définies temporellement et spatialement. Les observations de terrain ont signalé le potentiel pour des remontages systématiques des assemblages lithiques. Un programme intensif de remontage a été conduit en incorporant 30 assemblages comprenant plus de 10 000 raccords. Cette étude se concentre sur la reconstitution des changements diachroniques et synchroniques dans la chaîne opératoire de la production des microlithes pendant l’Epipaléolithique à partir de l’étude des remontages. Les contributions les plus significatives de cette étude sont la conclusion que toutes les entités culturelles épipaléolithiques du Néguev appartiennent à la même tradition technologique, l’approche de l’exploitation du nucléus, dite Wide-fronted (exploitation de la face large), et que les changements sont graduels. Cependant, plusieurs différences diachroniques et synchroniques au sein des séquences de production des microlithes entre les industries peuvent être observées. En revanche, les changements typologiques sont plus rapides. Chaque entité culturelle se distingue par des types morphologiques séparés de microlithes, qui apparaissent en un espace relativement court de 1000-2000 ans. Enfin, cette recherche atteste à nouveau la validité des études des remontages, qui se révèlent être un outil méthodologique puissant pour préciser les classifications typologiques et technologiques.
Entrées d’index
Mots-clés : Epipaléolithique, technologie lithique, études de remontages, désert du Néguev
Keywords : Epipaleolithic, lithic technology, refitting studies, Negev Desert
Texte intégral
Introduction
1Archaeological cultural entities rarely present the possibility of reconstructing diachronic and synchronic technological behavior in a confined geographic region. The Western Negev sites during the Epipalaeolithic (23–11.5 ka cal BP; table 11.1) present the perfect setting for such a study, as they are comprised of single, well-defined archaeological horizons. It was assumed that regionalism was a distinctive feature of the Epipalaeolithic technological production system (e.g. Goring-Morris 1987, 1995; Henry 1995; Richter 2011) and that these differences could be highlighted through refitting studies and the reconstruction of the main chaînes opératoires (production sequences). The following discussion focuses on the reconstruction of the microlith chaînes opératoire, which is the main production sequence present within the assemblages. Less distinctive chaînes opératoires for massive and medium-sized tools (usually endscrapers) were recognized but not presented, as they are less culturally and technologically distinct (Goring-Morris et al. 1998; Marder 2002). It was noted that some of the observed technological trends and changes, especially those of the Early Epipalaeolithic, are also integral to the Late/Terminal Upper Palaeolithic production system, and their roots may even be detected earlier in Early Ahmarian Upper Palaeolithic technological concepts and modes of production (Davidzon & Goring-Morris 2003; Goring-Morris & Davidzon 2006; Belfer-Cohen & Goring-Morris 2007; Leplongeon & Goring-Morris 2018).
table 11.1
14C Chronology* | Habitual use of microburin technique | No use of microburin technique |
LATE EPIPALAEOLITHIC | ||
12,500‑11,500/100 | Harifian Shunera XXIV; Shunera IX | |
13,500‑11,500 | Late Natufian Shunera XVIII; Givat Hayil I Miscellaneous Shunera II? | |
15,000‑13,500 | Terminal Ramonian/Early Natufian Nahal Sekher 23; Shunera VII; Azariq XV | |
MIDDLE EPIPALAEOLITHIC | ||
17,750‑15,000 | Ramonian Nahal Lavan 1003; Nahal Lavan 1009; Shunera XXI | |
18,500/17,750‑15,000 | Mushabian Azariq XII; Azariq VIIB; Har Qeren VIII; ?Shunera II? Proto-Mushabian Shluhat Qeren II; Azariq X | Geometric Kebaran Azariq XVI; Azariq XVIII; Shunera XIIB; Azariq VIIA, Azariq II |
EARLY EPIPALAEOLITHIC | ||
Intermediate Kebaran/Geometric Kebaran | ||
20,000‑18,700/500 | Nizzanan Hamifgash IV | Late Kebaran Shunera XVII |
22,500‑20,000 | Early Kebaran Azariq IV | |
25,000‑23/22.000 | Nebekian? | Masraqan/Late Ahmarian Azariq XIII; Shunera XVI |
UPPER PALAEOLITHIC | ||
30,000‑23,000? | Arqov/Divshon Shunera XV | |
46,000‑25,000 | Early Ahmarian Nahal Nizzana XIII |
2The chaîne opératoire approach was developed by French ethnologists whose focus was the study of various steps in production processes in order to reconstruct the technical activities and technological traditions in a given culture (Mauss 1947; Cresswell 1983, 1990; Lemonnier 1976, 1986, 1993). This approach was adopted by prehistorians to study the technical activities in given archaeological sites. It is based on the reconstruction of the different stages of flint production, from the acquisition of raw material to the final product and the abandonment of the desired and/or used objects (See summary in Pelegrin 1985, 1995; Pelegrin et al. 1988; Soressi & Geneste 2011; Tostevin 2011; Delage 2017).
3Using the chaîne opératoire concept reveals and facilitates the identification of different technical traditions that are passed down from one generation to another. Prehistoric humans possessed a range of different techniques in order to obtain the final product. Repetitive and continued use of preferred techniques should enable prehistorians to identify the technical systems of different social groups (Pelegrin 1985, 1995; Pelegrin et al. 1988). Thus, the identification of social groups enables researchers to reconstruct temporal and as well as chronological techno-economic changes that occurred in specific regions among hunter-gatherer technological systems (Valentin 1995; Klaric 2007).
4The complexity of this approach is not overlooked, for it demands the meticulous descriptive analysis of each technical stage, step and action. At the same time, it is also necessary to deal with the most abstract levels of analysis such as the concepts of knapping, knowledge, psycho-motor behaviors and technical skills. It seems that to penetrate the “prehistoric mindset” is an illusion and to describe it objectively is an almost impossible task (Dibble 1995; Shott 2003; Bar-Yosef & Van Peer 2009; Shott et al. 2011). Yet, the combination of refitting studies and the chaîne opératoire approach present a reliable methodological tool for reconstructing synchronic and diachronic changes within limited areas of the Negev during the Upper Palaeolithic and Epipalaeolithic (Volkman 1983; Ferring 1988; Goring-Morris et al. 1998; Marder 2002).
5In recent decades refitting of chipped stone assemblages has become a significant tool in investigating prehistoric cultures (e.g. Leroi-Gourhan & Brezillon 1966; Cahen et al. 1979; Cziesla 1990; Villa 1991; De Bie & Caspar 2000; Pigeot 2004; Churmans & De Bie 2007). It is commonly used in reconstructing technical traditions, technological systems, intra-site and inter-site variability, as well as site formation processes. In the Levant the refitting method has also been used as an integral part of the analysis of Palaeolithic and Neolithic chipped stone assemblages (e.g. Volkman 1983; Gilead & Fabian 1990; Goring-Morris et al. 1998; Becker 1999; Rosen 2000; Marder 2002: Davidzon & Goring-Morris 2003; Goring-Morris & Davidzon 2006; Mitki et al. 2013). The validity of the refitting method was evaluated by Marks, who argued that the application of refitting to all in situ sites is problematic, since it is very time consuming and cannot realistically be applied to large assemblages (Marks 1983). Moreover, Bordes also doubted the efficiency of reconstructing contemporaneous activity areas on the basis of refitting, since refits from adjacent concentrations are not necessarily proven to have existed simultaneously (Bordes 1980). In contrast, Cahen categorically and metaphorically stated that refitting provides… “a feeling of closeness to prehistoric man, that is given by the rejoining of flakes that were struck apart thousands of years ago” (Cahen 1987). Likewise, Villa emphasizes that refitting provides distinct information concerning behavioral aspects of prehistoric humans, and also increases the precision, objectivity and legitimization of inferential processes, as well as contributing to the credibility of the results obtained (Villa 1991).
The Epipalaeolithic chrono-stratigraphic framework
6In this section a short overview of the Southern Levant’s chrono-stratigraphic framework is presented with particular emphasis on the Negev and Sinai. The Terminal Pleistocene has been broadly sub-divided into: 1) Early Epipalaeolithic (25‑18.5 ka cal BP); 2) Middle Epipalaeolithic (18.5‑15.5 ka cal BP), and 3) Late Epipalaeolithic (15.5‑11.5 ka cal BP) (Goring-Morris & Belfer-Cohen 2017: tab. 71.1).
Early Epipalaeolithic (25‑18.5 ka cal BP)
7The Early Epipalaeolithic displays considerable continuity from the preceding Upper Palaeolithic. This continuity is apparent in the microlithic component and in the basic foraging subsistence mode (e.g. Gilead 1983, 1988). Nonetheless, innovations and high lithic typo-technological variability and diversity can be observed, along with various Southern Levant sociocultural aspects within distinct and limited geographical and ecological niches (e.g. Goring-Morris 1995; Goring-Morris et al. 2009; Maher et al. 2012b). The diversity of this period may reflect the harsh conditions prevailing during the Late Pleistocene, when overall carrying capacities decreased considerably, restricting settlement to isolated refugia in low-lying regions and those with plentiful water sources (e.g. Maher et al. 2011; Rosking et al. 2014; Jones et al. 2016; Goring-Morris 2017; Maher 2017; Vardi et al. 2018). However, it still in debate if the high degree of lithic variability observed in the Early Epipalaeolithic developed in a more linear progression with minimal contemporaneity between cultural entities (e.g. Bar-Yosef 1970; Goring-Morris 1995; Goring-Morris & Belfer-Cohen 2017), or developed more synchronically as suggested by the existence of multiple microlithic industries during this time span (e.g. Hovers & Marder 1991; Richter & Maher 2013). Within this refugia during the Early and Middle Epipalaeolithic periods, prolonged use of aggregation sites (c. 20,000 m2) was documented at Kharaneh IV and Jilat 6 at the Azraq Basin in the northern-central Jordanian Plateau. The sites exhibit multi-phase stratigraphy with thick archaeological accumulation (at least 2 m) including evidence of structures, burial on the hut floor (Kharaneh IV), hearths, middens, postholes and ash dumps, as well as high densities and concentrations of marine shells.
Kebaran Complex (22‑18.5 ka cal BP)
8The Kebaran techno-complex is comprised of numerous industries, which exhibit both regional and chronological diversity. The geographical distribution of the Kebaran is relatively widespread, focusing on the coastal plain (including Lebanon), and on both sides of the Jordan Rift Valley and including the Western Negev.
9A detailed study of the Kebaran complex was conducted by Bar-Yosef (Bar-Yosef 1970; Bar-Yosef & Vogel 1987), who identified four groups based on the lithic assemblages (A-D). The Kebaran can be sub-divided into two periods: the Early Kebaran is characterized by broad or narrow micropoints and large curved or pointed bladelets, while the Late Kebaran is distinguished by obliquely truncated and backed bladelets (the “Kebara point”) and curved backed bladelets (Bar-Yosef 1981).
10Kebaran sites are quite small, ranging between 50 and 150 m2 and rarely exceeding 250 m2 (Hovers et al. 1989; Goring-Morris & Belfer-Cohen 2017). Only one burial was discovered, at the site of Ein Gev I, below a semi-subterranean structure. A few art objects were unearthed, like the engraved plaques and nodules from Urkan e-Rubb II (Hovers 1990) and ‘En Qashish (Yaroshevich et al. 2014, 2016). Mortars, pestles, querns and mullers occur in some sites (e.g. Bar-Yosef 1970; Yeroshevivich et al. 2016).
Nizzanan (20‑18.7/5 ka cal BP)
11An assemblage dominated by minute triangles made by the microburin technique, defined as microgravette points at Ein Gev IV, was here assigned to the Geometric Kebaran (Bar-Yosef 1970; Bar-Yosef & Vogel 1987). It was re-designated by the newly defined term “Nizzanan” entity (Goring-Morris 1987, 1995). The Nizzanan is primarily distributed in the east of the Rift Valley (especially the Azraq Basin) and in the western Levant from Lebanon to the Negev (Byrd & Garrard 2017). Stratigraphic considerations, particularly 14C evidence, indicate that the Nizzanan postdated the Nebekian and Qalkhan, was coeval with the Kebaran and predated the Middle Epipalaeolithic Geometric Kebaran and Mushabian (Goring-Morris 1995; Byrd & Garrard 2017: fig. 74.3; Goring-Morris & Belfer-Cohen 2017).
12Technologically, the lithic industry has laminar aspects. The cores are commonly single-platform pyramidal types, though opposed platform varieties also occur (Goring-Morris 1987, 1995). Minute scalene and isosceles triangles made by the microburin technique and microgravettes characterize the lithic assemblages, which also include well-made endscrapers, burins (particularly dihedral variants) and pièces esquillées. Nizzanan sites vary in size, from small and ephemeral (a few square meters) to huge “aggregation sites” at Wadi Jilat 6, and Kharaneh IV, on the order of 20,000 m2 (see above).
Middle Epipalaeolithic (18.5‑15 ka cal BP)
Geometric Kebaran (18‑15 ka cal BP)
13The Geometric Kebaran probably developed from the local Mediterranean Kebaran, although, alternatively, earlier 14C dates from the Negev hint that it might have developed earlier in the marginal areas from the indigenous Kebaran. It seems that the duration of the Geometric Kebaran was not synchronic: in the south it was replaced by the Mushabian earlier than in the north, where it continued and later developed into the Natufian (Bar-Yosef & Belfer-Cohen 1989; Belfer-Cohen & Goring-Morris 2013).
14Geographically, the Geometric Kebaran is the most widespread Epipalaeolithic entity, from Wadi Feiran in southern Sinai to the Euphrates in northern Syria (Umm-el Tlel), and from the Mediterranean eastward to the Saudi Arabian border (Bar-Yosef & Belfer-Cohen 1989; Belfer-Cohen & Goring-Morris 2013). Moreover, Geometric Kebaran lithic assemblages exhibit a high degree of typological standardization in comparison with other Epipalaeolithic industries. The trapeze/rectangle and its variants absolutely dominate most lithic assemblages. Elongated endscrapers on blades are also distinctive (Goring-Morris 1987, 1995). It seems that the Geometric Kebaran displays typo-technological similarities to the Middle Hamran industries on the eastern side of the Rift Valley and exhibit greater variability than other industries of this entity (Henry 1995, 2017).
15Only fragmentary remains of semi-subterranean architectural structures were found in the Mediterranean zone (Ein Gev III, Haon II), while in the Negev and Sinai windbreaks may have been used. Sites are commonly small, ephemeral, often with only a single hearth (Goring-Morris & Belfer-Cohen 2017). In the Mediterranean zone, Neve David displays a large area (c. 1500 m2) and deep stratigraphy (at least 1 m), but with few features. The animal economy seems to follow the patterns of the earlier Upper and Epipalaeolithic, namely based on ungulate hunting (Yeshurun et al. 2015).
16Recent studies of Geometric Kebaran sites have revealed complex, elaborate mortuary behavior at the sites of Neve David (Kaufman 1987, 1989; Yeshurun et al. 2015), Wadi Mataha (MacDonald et al. 2016) and especially ‘Uyun al-Hammam (Maher et al. 2011). Built features, mostly single articulated interments accompanied by ground stone vessels and occasionally red ochre as well as worked bones were found within the burials (Stock et al. 2005; Bocquentinet et al. 2011; Maher 2007; Maher et al. 2011) Most interesting are the remains of at least 11 individuals discovered at ‘Uyun al-Hammam, two of which were associated with fox remains (Grave I and VIII) and also possible human skull removal within Grave I (Maher 2007; Maher et al. 2011). These findings provide strong evidence that key aspects of these complex mortuary traditions such as human-animal burials associated with graves goods and possibly also skull removal, which are common in the Natufian as well as in the Neolithic, occurred earlier in the Epipalaeolithic.
Mushabian (c. 18.5/17.75‑15 ka cal BP)
17The Mushabian complex was first defined by Phillips and Mintz (Phillips & Mintz 1977), and was considered to be of North African origin. The stratigraphic evidence and radiometric data show that the Mushabian was partially coeval and slightly later than the Geometric Kebaran. The Mushabian geographical distribution was restricted to the arid zones of the Negev and Sinai southward from the Beer Sheva Valley (Valla et al. 1979; Hermon 1996). However, similar and probably coeval industries to the Mushabian were documented in the Azraq Basin (Henry 1995, 2017; Byrd & Garrard 2017) and in southern Jordan. These industries reveal habitual use of the microburin technique, a higher frequency of non-geometrics versus geometrics and the presence of arch-backed bladelets as well as La Mouillah points. The Mushabians and their successors the Ramonians mainly settled the lowland dunes (Bar-Yosef & Phillips 1977; Goring-Morris 1987, 1995), although a few sites also occur in the central Negev Highlands (Marks & Simmons 1977; Belfer-Cohen & Goring-Morris 2019).
18The “classic” Mushabian lithic technology is represented by rather short, stubby blade/let blanks produced mainly from single platform cores that tend to rather globular shapes. Typologically, the lithic assemblages are distinguished by La Mouillah points and by arched- and blunt-backed to scalene bladelets habitually made using the microburin technique (Phillips & Mintz 1977; Goring-Morris 1987, 1995). More recent research demonstrates that Mushabian origins should be sought in Transjordan (Henry 1995; Goring-Morris 1995; Goring-Morris et al. 2009), where there is solid evidence for the early habitual use of the microburin technique (i.e. Nebekian, Qalkhan and Nizzanan).
Ramonian (17.75‑15 ka cal BP)
19This entity was first described as “Negev Kebaran” by Marks and Simmons (Marks & Simmons 1977; Kaufman 1983; Vardi et al. 2015) who divided it into two phases, the Harif phase and the Helwan phase. Many scholars (Bar-Yosef & Belfer-Cohen 1989) considered these assemblages “Late Mushabian” while Goring-Morris (Goring-Morris 1987, 1995) also views them as originating in the Mushabian, but believes that this industry has enough distinct characteristics in terms of lithic assemblages and settlement patterns to designate it as an independent industry (the Ramonian) within the Mushabian technocomplex. The “fossile directeur” of this industry is the obliquely truncated and backed bladelet with a straight to concave back, called the Ramon point. These points are absent from “classic” or early Mushabian assemblages. Subsequently, during the transition from Middle to Late Epipalaeolithic the Terminal Ramonian/Early Natufian assemblages are dominated by Ramon points and Helwan lunates, both often made on chalcedony.
Late Epipalaeolithic
Terminal Ramonian/Early Natufian (15‑13 ka cal BP) - Late Natufian (13.5‑11.5 ka cal BP)
20The Natufian emerged abruptly at the end of the Late Glacial Maximum, possibly first in the Mediterranean zone and in limited areas of the Negev and Transjordan (Belfer-Cohen & Goring-Morris 2013; Goring-Morris & Belfer-Cohen 2013). The Natufian saw the first appearance of complex hunter-gatherer groups in the Near East, with large-scale permanent settlements located in ecotones on both sides of the central mountain ridge in the Mediterranean zone (the core area), while the coastal plain was largely abandoned. Major innovations and intensifications in certain realms occurred in the material culture and settlement patterns (Byrd 1989; Henry 1989; Belfer-Cohen 1991; Valla 1995; Bar-Yosef 1998; Grosman & Munro 2017; Ibáñez et al. 2017). These included sophisticated architecture, the appearance of cemeteries, a major increase as well as a change in the composition of ground stone, chipped stone and bone tools, along with the proliferation of artistic and decorative endeavors. Subsistence included broad-spectrum exploitation of animals (but with a specialization in gazelle hunting), consumption of vegetal resources and the domestication of dogs probably as a result of social rather than subsistence needs (e.g. Techernov 1984; Valla 1984; Techernov & Valla 1997; Stiner et al. 1999 Grosman et al. 2008; Valla et al. 2007; Edwards 2012; Valla 2012: Nadel et al. 2013; Yeshurun et al. 2013, 2014; Grosman et al. 2016; Weinstein-Evron et al. 2018).
21Within the Negev and northern Sinai increase in site density is documented during the onset of the Younger Dryas (Grosman 2013) with the appearance of the Late Natufian and Harifian. Their settlement pattern was composed of residential bases with stone-built architecture in the Negev Highlands (Henry 1976; Scott 1977; Marks & Larsen 1977; Goring-Morris 1991, 1998; Goring-Morris et al. 1999). Such residential bases were accompanied by smaller task-specific sites in the lower altitudes of the northwestern Negev dunefields (Scott 1977; Goring-Morris 1987, 1991, 1997). The presence of Red Sea and Mediterranean mollusk species in the Late Natufian and Harifian as well as exotics raw materials (e.g. malachite, green stones and basalt) suggests that these entities were highly mobile and possibly maintained a regular exchange system with other regions of the Southern Levant (Bar-Yosef Mayer & Zohar 2010; Kurzawska et al. 2013).
The refitted assemblages
22During the course of the Emergency Archaeological Survey of the Negev from 1979 to 1984, dozens of Upper Palaeolithic and Epipalaeolithic sites were systematically surveyed and excavated in the Western Negev dunes (Goring-Morris 1987; Goring-Morris & Rosen 1987). For many of the assemblages the potential for refitting was evident. Since 1996 an extensive refitting project has been underway, encompassing more than 25 assemblages from the Upper Palaeolithic and Epipalaeolithic periods (Goring-Morris et al. 1998). More than 10,000 assemblages objects were refitted, including broken items and conjoinable pieces, both to cores and/or nodules, and artifacts conjoining to each other without cores. The number of conjoins for individual assemblages was as high as 1,125 items, a refitting rate of 0.1‑32.2% (table 11.2). For most assemblages, conjoin rates were well in excess of 200 items (10%), overall; the average refitting rate calculated previously, for 24 assemblages, was 12.5% (Goring-Morris et al. 1998).
table 11.2
RAW COUNTS | REFITS | |||||||
PERIOD/Entity/Site | Cores | Waste | Tools & Mbt | Total | Tools & Mbt | % of Tools & Mbt | Total Items Refitted | Total Refitting Rate |
UPPER PALAEOLITHIC | ||||||||
N. Nizzana XIII | 18 | 998 | 40 | 1056 | 17 | 42.5 | 521 | 49.3 |
EARLY EPIPALAEOLITHIC | ||||||||
Masraqan | ||||||||
Azariq XIII | 40 | 3639 | 1188 | 4867 | 41 | 3.5 | > 1125 | 23.1 |
Shunera XVI | 35 | 4181 | 243 | 4459 | 49 | 20.2 | > 975 | 21.9 |
Kebaran/Nizzanan | ||||||||
Azariq IV | 13 | 1622 | 199 | 1834 | 45 | 22.6 | 440 | 24.0 |
Hamifgash IV* | 29 | 1543 | 565 | 2138 | 40 | 7.1 | 237 | 11.1 |
Shunera XVII | 1 (3) | 529 | 407 | 925 | 41 | 10.1 | 192 | 20.8 |
MIDDLE EPIPALAEOLITHIC | ||||||||
Geometric Kebaran | ||||||||
Azariq XVIII | 12 | 2543 | 519 | 3075 | 19 | 3.7 | 270 | 8.8 |
Azariq XVI | 25 | 5764 | 1256 | 7043 | 66 | 5.2 | 496 | 7.3 |
Azariq II | 18 | 1514 | 389 | 1921 | 16 | 4.1 | 137 | 7.1 |
Shunera XIIB | 5 | 1224 | 331 | 1560 | 22 | 6.6 | 85 | 5.4 |
Shunera I | 18 | 871 | 209 | 1095 | 4 | 1.9 | 167 | 15.2 |
Proto-Mushabian | ||||||||
Shluhat Qeren II* | 15 | 4910 | 1379 | 6304 | 27 | 2.0 | 344 | 5.5 |
Azariq X* | 37 | 2415 | 535 | 2987 | 22 | 4.1 | 193 | 6.5 |
Mushabian | ||||||||
Azariq XII* | 62 | 4421 | 1688 | 6162 | 47 | 2.8 | 802 | 13.0 |
Azariq XX* | 30 | 4006 | 1569 | 5605 | 15 | 1.0 | 150 | 2.7 |
Har Qeren VIII* | 3 | 653 | 22 | 678 | 9 | 40.9 | 248 | 36.6 |
Ramonian | ||||||||
Nahal Lavan 1003* | 6 | 229 | 104 | 340 | 8 | 7.7 | 72 | 21.2 |
Nahal Lavan 1009* | 8 | 742 | 645 | 1399 | 30 | 4.7 | 230 | 16.4 |
Shunera XXI*-sample | 25 | 4026 | 1719 | 5770 | 78 | 4.5 | 543 | 9.4 |
LATE EPIPALAEOLITHIC | ||||||||
Terminal Ramonian/ Early Natufian | ||||||||
N. Sekher 23* | 28 | 1095 | 514 | 1635 | 32 | 6.2 | 360 | 22.0 |
Shunera VII* | 43 | 2035 | 755 | 2831 | 46 | 6.1 | 446 | 15.8 |
Azariq XV* | 12 | 2408 | 320 | 2740 | 81 | 25.3 | 539 | 19.7 |
Late Natufian | ||||||||
Shunera XVIII* | 11 | 681 | 212 | 904 | 1 | 0.5 | 69 | 7.6 |
Givat Hayil I* | 44 | 4863 | 1904 | 6811 | - | 0.0 | 4 | 0.1 |
Harifian | ||||||||
Shunera XXIV* | 8 | 563 | 154 | 725 | 17 | 11.0 | 100 | 13.8 |
Shunera IX* | 46 | 1705 | 450 | 2201 | 15 | 3.3 | 206 | 9.4 |
MISCELLANEOUS | ||||||||
Shunera II* | 24 | 3773 | 580 | 4377 | 53 | 9.1 | 860 | 19.6 |
TOTAL REFITS | 600 | 62211 | 17686 | 80454 | 838 (493) | 4.7 | > 9498 (4485) | 11.8 |
23From the extensive project conducted, 13 of the Early to the Late Epipalaeolithic assemblages are included in the present study (Marder 2002) (figure 11.1; table 11.1). These focus on the Middle (Geometric Kebaran, Mushabian, and Ramonian) and earlier part of the Late Epipalaeolithic (Terminal Ramonian/Early Natufian). Other parts of the Epipalaeolithic are presented here by one Late Kebaran and one Nizzanan sites, and by comparison with Early Ahmarian/Masraqan and Late Natufian and Harifian sites.
24All sites selected for the following analysis are open-air occurrences, deriving from thin archaeological horizons that commonly do not exceed 10 cm in thickness. The site sizes are small, rarely exceeding 100 m2, usually with one or more hearths. Most of the sites were almost completely excavated, therefore from 75‑95% of the sites were collected (Marder 2002: tab. iii-1). Standard retrieval techniques were based on ¼ m2 grids, with sediments sieved through 2‑3 mm mesh. In only a few instances, where sites were extensively deflated on fossilized dune surfaces, were the assemblages collected without a grid and/or without sieving (i.e. Nahal Lavan 1003, Nahal Lavan 1009).
25Analyzed assemblages varied from 340 to 7,043 chipped artifacts (excluding debris), though most rarely exceeded 3,000 artifacts (table 11.2). The refitted assemblages include debitage, chunks, cores, microburins and formal tools, comprising 40‑80% of the total assemblages. Chips were generally excluded from the analysis since they were too diminutive and fragmentary to conjoin. Only when cores or nodules were almost entirely refitted and relatively few artifacts left to complete the “puzzle” were attempts made to conjoin the chips.
26From each traditionally defined Epipalaeolithic socio-cultural entity, two to three assemblages were refitted, while for the Early Epipalaeolithic (i.e. Late Kebaran and Nizzanan) only one site was suitable for detailed analyses (table 11.1). Refitted tools and microburins were analyzed by the same method as the other items. However, additional and supplementary data were provided from Goring-Morris’ work (Goring-Morris 1987) since the typological and technological information from the conjoins is biased as a result of the very low refitting rates for microliths (c. 1‑2%) and the inconsistent success rates for microburins and other tool types, which ranges from 0.0‑14.4% (Goring-Morris et al. 1998).
Results
27The microlith chaîne opératoire is the main production sequence, comprising 70- 95% of the Epipalaeolithic assemblages of the region (Marder 2002: tab. iv: 1a-b). Within this reduction sequence a few ad hoc tools such as burins occasionally were derived from the initial shaping-out stage (e.g. Hamifgash IV, Nahal Lavan 1009, Azariq XVI). In the Late Epipalaeolithic assemblages, microliths were produced (sensu lato) together with standardized tools such as endscrapers (Geometric Kebaran sites; figure 11.2A) or “spokeshave” notches/denticulates (Terminal Ramonian/Early Natufian sites; figure 11.3). These tools were produced from well-defined locations within the core reduction sequence.
The microlithic chaîne opératoire
Raw Material Exploitation
28High quality flint is widely available throughout most of the Negev. Sites of any period are rarely located more than 20 km (one day’s walk) from high quality flint sources. Chert and chalcedony sources are available in numerous bedrock outcrops and from Pleistocene fluvial terraces (Goring-Morris 1987, 1997; Ashkenazy 2014: 325‑326; Leplongeon & Goring-Morris 2018).
29Based on the refitting studies both diachronic and synchronic changes were noted in raw material procurement and exploitation patterns. This phenomenon was previously described primarily concerning the use of translucent chalcedony at the beginning of the Late Epipalaeolithic Terminal Ramonian/Early Natufian and Late Natufian (Goring-Morris 1996; Goring-Morris et al. 1998). In the Terminal Upper Palaeolithic/Early Epipalaeolithic, discoidal or elliptical nodules of fine-grained flint were preferentially selected; these may not always have been available in the immediate vicinity of sites (Davidzon & Goring-Morris 2003; Goring-Morris & Davidzon 2006).
30Observations indicate that irrespective of the site or entity involved, the brittleness of the nodules (whether flint or chalcedony) frequently caused major problems during knapping. The preference for chalcedony is probably related to it being less brittle, so that debitage and tools were less likely to break during knapping (Marder 2002: tab. iv-2). In addition, its physical qualities result in more standardized end products. Finally, chalcedony has a certain aesthetic quality and that could possibly indicate even some symbolic value (Goring-Morris & Belfer-Cohen 2002; Edwards 2007).
31During the Upper Palaeolithic and Early Epipalaeolithic there is evidence for careful selection of nodule shapes illustrated by the preference for discoidal nodules at Shunera XVII and the more elliptical nodules at Hamifgash IV. Starting with the Middle Epipalaeolithic there seems to be less emphasis on nodule shape. During the Late Epipalaeolithic, with the introduction and later habitual use of chalcedony, the nodules selected tend to diminish in size from cobble size (0.5‑1 kg) to fist size (0.2‑0.1 kg). In contrast, the dimensions of the regular flint nodules selected did not change diachronically.
Core Shaping-out
32This section synthesizes the core-reduction sequence of the studied sites by means of diachronic and synchronic comparisons. During the Epipalaeolithic core setting-up was not carefully practiced, and the preform stage was characterized by minimal pre-planning and investment in basic configuration (crest and platform preparation, platform setting-up, etc.)
33The most common method of shaping-out, during this time period, was an initial splitting of the nodule into several blocks (Marder 2002: fig. vii: 3‑4) (figure 11.4), using a powerful blow on an anvil. This was achieved by vertical and/or horizontal removals, usually from a globular, amorphous or discoidal nodule. After the nodule was split, each piece (some displaying the flake’s ventral face) was separately decorticated and flaked. The knapper usually took advantage of the flat split surface and used it as the striking platform (figure 11.4). However, occasionally the split surface was used as the lateral face or back of the core necessitating preparation of a striking platform (Marder 2002: fig. vii: 4). In most cases the platform was formed by the removal of a cortical item from the nodule’s top (Marder 2002: fig. vii: 5).
34The splitting method was systematically practiced during the Early and Late Epipalaeolithic, and less commonly used during the Middle Epipalaeolithic. This method may have been favored by the knapper as it is a highly efficient way to reduce small globular nodules. The increased use of the splitting method during the Late Epipalaeolithic may also be linked to the use of chalcedony. As chalcedony was not locally found, there was a need to maximize utilization and the splitting method made the need for nodule cortification redundant (figure 11.4).
35Nodule decortication mainly involved the removal of cortical blade/lets from single platform cores, with slight differences between the assemblages. The systematic "onion-like" peeling of mainly incurvate and occasionally overpassed blade/lets from single platform unfaceted cores is evident throughout most of the Epipalaeolithic (Marder 2002: app. 1, figs 1‑2) (figure 11.5). The unidirectional flaking resulted in mostly globular cores. During the Late Epipalaeolithic (Terminal Ramonian/Early Natufian assemblages) the decertification was done by standardized unidirectional-convergent removals of mainly incurvate or overpassed blade/lets, forming a single-platform unfaceted pyramidal core (figure 11.4). This “onion-like” peeling of the nodules, evident in the refitted series, differed from the local Upper Palaeolithic reduction sequences. The later entailed pre-planning and systematic shaping-out of specifically selected narrow-sided or keeled nodules (Monigal 2001, 2003; Davidzon & Goring-Morris 2003; Goring-Morris & Davidzon 2006). In the Early Epipalaeolithic (figure 11.6) there are rare examples of core reduction sequences that resemble Upper Palaeolithic sequences.
Blank Production
36Throughout the Epipalaeolithic blanks were produced from single-platform cores. Wide striking platforms with minimal platform rejuvenation were the norm. In all industries the majority of blanks for microliths (60‑80%) were removed from the core’s hub (the internal part of the nodule once all cortex was removed). A sequential removal of a cluster of 5‑15 potential blanks was possible when core convexity and the removal angle was appropriate (usually > 65°< 80°). Between each series the core’s removal surface was abraded. One nodule usually furnished one or two series and, occasionally, even three (figures 11.7; 11.8: 1‑3; 11.9: 1‑3, 5; 11.10).
37In industries where the microburin technique was used (Mushabian, Ramonian and Terminal Ramonian/Early Natufian) more blanks were derived from the decortication and immediate pre-abandonment stages (up to 40%), since a wider variety of blanks could be utilized (figure 11.10: B). Blank dimensions indicate a high level of standardization among the different industries, with a few exceptions, mostly related to the use of chalcedony. At Shunera XVII (Kebaran) elongated (> 60 mm) incurvate bladelets were the preferred blank for Kebaran point production. At least a third of the blank’s length and width was reduced during backing (Marder 2002: fig. 4). At Hamifgash IV (Nizzanan), there is an evident bimodal pattern in blank size selection for the production of scalene/arch backed bladelets (Goring-Morris 1987). Within the Late Epipalaeolithic (Terminal Ramonian/Early Natufian assemblages), different blank sizes were selected for the production of lunates and Ramon points (Goring-Morris 1987). This variety can be related to a difference in raw material selection (i.e. quality, shape and size, chalcedony versus regular flint), as Ramon points were manufactured on both chalcedony and other flint types and Ramon points on chalcedony tended to be shorter (usually 30‑35 mm) and narrower (7‑12 mm) than such points on regular flint (> 40 mm; figure 11.10). In contrast, the lunates were smaller and almost exclusively produced on chalcedony flint; they usually do not exceed 35 mm, except at Azariq XV (Marder 2002: tab. vi-2).
38A close look at the refitted series from all Epipalaeolithic sites suggests that the majority of bladelet blanks were rejects and not merely unused blanks. At Azariq XII a detailed study of the blade/lets conclusively showed they were too thin (2‑3 mm), too narrow (< 12 mm) or too short (< 32 mm) to be modified into microliths. Since most of the recovered debitage was not suitable for further modification, their metric attributes are not representative of actual blank dimensions, but rather merely represent “real” discarded waste material (Marder 2002: 190).
Core Maintenance and Discard
39Maintenance was not regularly practiced or organized in a specific set of operations. Debitage plane rejuvenation, striking platform renewal and correction of knapping mistakes were rare (Marder 2002: tabs 1‑3; fig. vii-11) (figure 11.11: A). Such maintenance took place only following the unsuccessful removal of potential blanks, usually resulting in irregular shapes.
40Lateral and distal blunting were regularly practiced at different stages of core reduction (figure 11.11: B) (Marder 2002: fig. vii-9). Blunting involved the systematic flaking of small chips, followed by rubbing with a precursor (similar to platform abrasion) contributing to flaking surface convexity and extending the core’s distal edge (figure 11.11: B). Blunting is evident on more than 50% of the Middle Epipalaeolithic cores. During the Late Epipalaeolithic this practice declined sharply (c. 30%) (Marder 2002: tab. vii-1). The main objective in forming a blunt edge was to facilitate handling the core during knapping. Blunting is not a technical function, but can also be perceived as idiosyncratic behavior (Gunn 1977; Schlanger 1994; Foulds 2010) relating to the way the individual knapper gripped the core during knapping.
41Knapping generally involved the utilization of a single removal surface throughout the reduction sequence, with minimum rejuvenation, in order to extract the maximum number of potential blanks. Core platforms were usually not rejuvenated (c. 45%) at all following initial platform preparation. Consequently, overall platform rejuvenation rates range between 0.3 and 1.2% per platform, and only one or at most two core tablets were removed for each rejuvenation procedure. The site of Shunera XVII is an exception, reflecting systematic platform renewal. Two to three rejuvenation cycles per core are noted, each cycle included the removal of three standardized core tablets. In this respect, the Early Epipalaeolithic Kebaran assemblage at Shunera XVII may reflect the tail end of the local Upper Palaeolithic Ahmarian tradition of intensive platform renewal (Davidzon & Goring-Morris 2003: fig. 50a-b; Monigal 2003: figs 11.2: 3; 11.5: d; Goring-Morris & Davidzon 2006: fig. 10). More than half (56.3%) of the technological core tablets, defined following their refitting position within the sequence, do not actually classify as such typologically. These items are rarely rounded in shape, nor do they display scars of the former core debitage plane. Rather, most are flakes with hinged or pointed distal ends and a faceted striking platform (figure 11.12).
42Knappers appear to have focused on continued utilization of the main platform, with minimal platform alteration and shifts in the direction of removals, although occasionally, maintenance procedures were necessary during flaking. These activities emphasized minimal modification of the core’s lateral faces and distal end. In rare cases only crested blades were prepared within this rejuvenation process.
43Two major problems were commonly encountered during reduction that required the knapper’s attention. The first concerns the dorsal concavity of the core’s flaking surface (figure 11.13), while the second reflects increasing flint brittleness in some cases toward the center of the nodule (figure 11.14). Dorsal concavity was maintained throughout the reduction sequence, and just before core abandonment. The problem of raw material brittleness caused major difficulties throughout the reduction sequence, often restricting the number of potential blanks produced per core or even forcing the knapper to prematurely abandon the core. Preliminary observations from refitting indicate that more than half of the items were broken during knapping. Many represent “knapping accidents” in the form of proximal languettes (Marder 2002: app. 2: B13) and Siret breaks, which result from rebounding created by the supported core and the pressure applied by the flintknapper’s blow through the core’s flaking surface (figure 11.14). In order to evaluate the assumption that broken languettes indeed represent knapping accidents, a knapping experiment was conducted (knapping conducted by Boe Madson). Two single platform blade/let cores were knapped using soft organic direct percussion on fine-grained flint from the Western Negev. During reduction almost half the blank broke. The majority (> 80%) of breaks were languettes.
44Most cores were abandoned close to the point of exhaustion (c. 90%). This is reflected by their small dimensions, often shorter than the average blank lengths (Marder 2002: tab. vi-6) (figure 11.14). Many of the discarded cores display scars of hinge-fractured and overpassed blades. Some of the larger cores (> 50 mm), may reflect premature abandonment as they display standardized blade/lets scar patterns and systematically abraded removal surfaces (Marder 2002: app. 1, fig. 16).
Microlith Production
45Some Epipalaeolithic industries are characterized by the production of a single microlith morphotype, such as the obliquely truncated bladelet in the Kebaran, the trapeze/rectangle of the Geometric Kebaran and the Ramon point in the Early Ramonian. Other industries display two distinct morphotypes such as the microgravette and arch-backed/scalene bladelet in the Nizzanan, the arch-backed/scalene and blunt/splay backed bladelets of the Mushabian, and the Ramon point and lunate of the Terminal Ramonian/Early Natufian (Bar-Yosef 1970; Goring-Morris 1987; Henry 1995).
46The systematic manufacture of distinct morphotypes may reflect a combination of hafting differences and tool function, as well as stylistic proclivities (Bar-Yosef 1987). It is unclear whether the presence of two morphotypes within an assemblage reflects two separate lines of composite equipment manufacture or rather a single composite item incorporating both types. Refitting clearly indicates that in the case of the Mushabian, the different microlith types derive from the same reduction sequence, reflecting a divergence in blank selection for each morphotype. A similar pattern is indicated for the Terminal Ramonian/Early Natufian.
47During the Levantine Epipalaeolithic two distinct approaches to blank modification into microliths have been described. The first involves straightforward retouch/backing and/or simple snapping, as was regularly used during the Kebaran and Geometric Kebaran and sometimes during the Natufian (especially in the Mediterranean zone) (Goring-Morris 1995). In the Kebaran, well-made, elongate and slightly incurvate blade/lets were modified into backed points, while Geometric Kebaran assemblages display intentional snapping of long-blade blanks into two or three segments, modified by backing into trapeze/rectangles. Blank segmenting into several microliths is supported by high tool-to-core ratios, often in excess of 20, in most Geometric Kebaran assemblages in the Negev and Sinai (Goring-Morris 1987). Therefore it seems that these blades were backed on a fixed anvil and then snapped perpendicularly (e.g. Azariq XVI, Shunera XIIB) (Marder 2002: app. 1, figs 14: 2‑3, 6; 18: 1‑4). Henry suggested that perpendicularly snapped-backed microliths at Geometric Kebaran sites may often actually represent complete microliths that were used “as is” rather than broken fragments (Henry 1989). Among the refitted assemblages, and in a number of other Geometric Kebaran assemblages in the Negev and Sinai, Couze retouch, i.e. retouch on a languette fracture (Bordes & Fitte 1964) to attain the truncations on some trapezes/rectangles has been described on snapped blanks (Bar-Yosef & Goring-Morris 1977).
48The second approach systematically utilized the microburin technique, which likely originated east of the Rift Valley during the Early Epipalaeolithic Nebekian (Henry 1995; Byrd & Garrard 2017; Goring-Morris & Belfer-Cohen 2017). It was first habitually used for certain morphotypes in the Negev and Sinai from the later Early Epipalaeolithic Nizzanan and subsequently, during the Mushabian, Ramonian, Natufian and Harifian. It should also be stressed that within both the Nizzanan and Mushabian industries, the microburin technique was systematically used for a certain morphotype (i.e. arch-backed/scalene bladelet), but not for another (i.e. microgravettes and blunt/splay backed-bladelets). The microburin technique involved the preparation of a notch at the desired (usually distal) end. Subsequently the blank was broken by a small blow or by a twisting motion into a piquant trièdre (or, less commonly, a La Mouillah) and a (distal) microburin. In most industries for which the technique was used, only a single microburin was removed from one extremity. In Terminal Ramonian and Natufian assemblages (table 11.1) lunates were prepared by the removal of a microburin from both ends of the blanks. In the Nizzanan (at Hamifgash IV) occasional medial microburins may attest to the production of two microliths per blank.
49On the basis of this study, the following functional advantages and constraints of the microburin technique can be suggested:
- The main advantage in using the microburin technique is that it provided excellent control in achieving an oblique pointed shape, which was more difficult to accomplish using regular retouch or backing. Refitting indicates that almost any elongated item with more or less parallel edges could suffice as a blank. This is especially valid when the technique was applied to the blank’s distal extremity, when almost any shape, thickness, profile or inclusions (including hinged and overpassed items; figures 11.7‑11.8, 11.10) could be modified and removed using the technique. For the Ramonian in particular the technique was applied almost exclusively to the distal tips.
- The microburin technique is less effective for producing microliths with perpendicular (rectangles) or almost perpendicular truncations (trapezes; e.g. Geometric Kebaran), due to the angle of breakage, partially explaining the absence of microburins at Geometric Kebaran sites both in the Southern and Central Levant (figure 11.2: B). This absence also can be described as a rigid cultural preference of the Geometric Kebaran toward snapping and straightforward backing, as in the Negev, this cultural entity chronologically overlaps with the Mushabian, which did use the microburin technique.
- The microburin technique is an economizing technique, enabling the serial preparation of a large quantity (10 +) of blanks by simple actions and with minimal investment of time and energy, while at the same time widening the pool of potential blanks for selection. We postulate that this technique would be systematically employed in sites or in certain activity areas where intensive retooling took place. This may explain the higher indices of the microburin technique (as well as geometric microliths) in the lowland sites in the Negev, where more intensive retooling took place in comparison with the Natufian basecamps in the highlands, where retooling activity was perhaps less intensive (Goring-Morris 1999).
50During core reconstruction two main types of microlith rejects (accidents de travail) or incomplete microliths were discerned. These are often traditionally classified among the tools (i.e. notches, and La Mouillah points). The first type comprises broken (commonly more or less perpendicularly) retouched lateral notches, often located toward one end of the blank. These spalls appear to have snapped prematurely in an irregular manner often due to too deep or wide a preparation of the notch or as a result of chalk inclusions (figures 11.8: 2‑3; 11.9: 3).
51The second type includes La Mouillah points (follow Goring-Morris 1987: 186 in using a very restricted definition for this type). All refitted La Mouillah points were originally spalls of the microburin technique, for they display irregular shapes (i.e. shouldered, irregularly backed edges) and/or deep concave backing, which shortened the blank’s length (figures 11.9: 2, 11.10: B). Thus, it is not always clear whether La Mouillah points represent unsuccessful and hence intentionally discarded items, unfinished tools (as tentatively indicated by the refitting study), or whether some may have functioned as tools.
52Refitted sequences from the present study and other studies concerning the Upper Palaeolithic and Epipalaeolithic periods show that microlithic chaîne opératoire presents the use of two distinct knapping concepts. These concepts are not unique to sites in the arid region and characterize the entire Southern Levant (e.g. Bar-Yosef 1991; Shimelmitz 2002; Nadel 2003; Kuhn et al. 2004; Ashkenazy 2014; Kadowaki et al. 2015; Cheng 2019). The first method is the “Upper Palaeolithic volumetric concept,” not exclusively used for microlithic production. It is characterized by a highly standardized reduction sequence that focused on the core’s narrow front or nose in order to produce well-made, standardized, narrow incurvate blade/let blanks, usually from unfaceted single platform, pyramidal or rectangular shaped cores, called here the “Narrow-fronted concept” (N-fronted). The second method, the “Epipalaeolithic volumetric concept,” almost exclusively used for microlithic production, is characterized by a flexible and versatile utilization of the core’s wide front in order to produce irregular wide blade/lets blanks from predominately unfaceted, single platform pyramidal, globular or even amorphous cores (designated here as the “W-fronted concept”).
53The N-fronted concept initially appears at the “transitional” Middle‑Upper Palaeolithic and was continuously used throughout the Ahmarian and Masraqan Upper Palaeolithic industries and into the Early Epipalaeolithic (e.g. Jones 1983; Volkman 1983; Belfer-Cohen & Goring-Morris 2002; Davidzon & Goring-Morris 2003; Goring-Morris & Davidzon 2006; Goring-Morris & Belfer-Cohen 2018).
54The N-fronted concept results in overall blank shapes and sizes that resemble the “final products,” with straight or slightly incurvate profiles and convergent pointed forms, which were modified by minimal retouch into el-Wad points or fashioned into retouched bladelets by minimal “Ouchtata” and semi-steep retouch. The production of such well-defined blanks reflects meticulous selection of raw material types, shapes, sizes and qualities; highly standardized setting up of the core (regular preparation of an atypical primary crest, obligatory use of an elongated striking platform and a narrow debitage plane); and regular maintenance activities, including intensive uniform platform rejuvenation by the removal of well-defined core tablets.
55This concept displays a high level of knowhow and a high degree of efficiency. During knapping a large proportion of the initial nodule volume was reduced (up to two-thirds) resulting in a relative low number of blanks produced. Superficially this indicates non-economizing (i.e. low cost‑benefit ratio) knapping behavior. Nevertheless, closer examination of the cores after refitting indicates that blank production was actually very efficient, since the blanks comprise most of the core’s hub volume. Even when the core was near the point of exhaustion, well-made blanks could still be produced.
56The W-fronted concept dominates the Middle and Late Epipalaeolithic assemblages, although it also appears sporadically during the Early Epipalaeolithic (Goring-Morris et al. 1998; Marder 2002). This concept is characterized by an economical, highly efficient high cost‑benefit ratio and a flexible knapping strategy, with a superficial and poor level of knowhow. Minimal energy was invested in setting up the core form (i.e. platform preparation, decortication) and in platform rejuvenation. Blank selection was flexible as a result of the common, systematic use of the microburin technique (except the Kebaran and Geometric Kebaran), which further increased the range of potential blanks for selection. Short, squat, irregularly shaped blanks were commonly preferred. In consequence only a small amount of the nodule’s initial volume was reduced (1/4‑1/3) relative to the high number of potential blanks produced.
57Both concepts are conservative, and existed over long periods of time (N-fronted for c. 25,000 years and W-fronted for c. 10,000 years), irrespective of shifts in mobility, settlement patterns and economy. For instance, the Natufian knapping concept does not differ significantly from Mushabian or even Geometric Kebaran knapping strategies, despite the shift from a mobile hunter-gatherer-based economy to a sedentary-based one (Belfer-Cohen 1991).
Discussion
58The most significant conclusion of this study is that the Epipalaeolithic cultural entities all belong to the same conservative technological tradition, the W-fronted approach to core exploitation, and that changes were gradual. This is particularly true concerning the Epipalaeolithic of the Negev, but is probably also applicable to the Epipalaeolithic of the entire Levant (see above). A widespread “microlithization” process took place with the onset of the Epipalaeolithic, with composite tools replacing a number of Upper Palaeolithic tool types. Within this process a clear typological shift in the types of retouch occurred, from simple Ouchtata, fine or semi-abrupt into abrupt retouch. In addition, bladelet manufacturing increased and a wide range of different morphotypes of backed microliths appeared. Blade/let modification was achieved by invasive retouch and/or the introduction of microburin technique (Belfer-Cohen & Goring-Morris 2002; Richter 2011; Byrd & Garrard 2017; Goring-Morris & Belfer-Cohen 2017; Henry 2017). The initial shapes of the blanks became insignificant as a result of the habitual use of backing and/or the microburin technique. The selective use of the microburin technique reflects a discrete trait, and can be seen as a marker of specific cultural entities. The general transition to the use of non-standardized blanks is thought to reflect a shift toward an overall economized and high cost‑benefit, flexible tool production (Goring-Morris et al. 1998; Belfer-Cohen & Goring-Morris 2002).
59This process of “microlithization” was probably connected to the use of composite tools, variations in specific hafting techniques, a shift from maintainable to reliable hunting technologies and, perhaps, the adoption of the bow and arrow and a shift in food resources, mobility and settlement patterns as well as demographic changes (e.g. Bleed 1986; Valla 1987; Bar-Yosef & Kuhn 1999; Belfer-Cohen & Goring-Morris 2002; Kuhn & Elston 2002; Yaroshevich et al. 2010, 2013). In sum, the technological production system was more responsive to changes in tool function, hafting methods and raw material constraints, while it was less flexible in terms of changes in knapping traditions, styles, or idiosyncratic traits of the knapper. One of the most significant results of this study is that in contrast to the rapid, typological changes noted in previous research, the rate of technological change within the Epipalaeolithic period was much slower. Each cultural entity (e.g. Kebaran, Geometric Kebaran, Mushabian, Natufian, etc.) is distinguished by discrete morphological types of microlith (e.g. rectangle/trapeze; lunate/triangle; arch-backed/scalene bladelet, etc.), which appear over a relatively short time span of 1,000‑2,000 14C years. These typological changes are caused mainly by stylistic and ethnicity forces and to a lesser extent by mobility patterns, raw material and/or functional considerations as some have claimed (e.g. Barton & Neeley 1996; Neeley 2002).
60This study shows that refitting techniques are exceptionally efficient in small-size sites (rarely exceeding 100 m2) with thin archaeological horizons (rarely exceeding 10 cm in thickness and 3,000 artifacts), which were rapidly covered by sand after abonnement. The sites presented were almost completely excavated, thus 75‑95% of the assemblages were collected and analyzed, increasing the adequacy of refitting methods. The average conjoining rate for the Upper/Epipalaeolithic Western Negev sites, calculated for 24 assemblages, is 12.5% and ranges between 15% and 30% (see above; table 11.2). In comparison with other refitted assemblages, the Western Negev sites fit well within the variation of the presented success rates. The success rate of these studies are not usually very high and a rate of 20% is considered good (Cziesla 1990: tab. 1; Laughlin & Kelly 2010: tab. 1). Our experience also indicates that refit success rates decline with an increase in assemblage size, and refits become unmanageable with large assemblages (> 6,000 artifacts), even when taking into account variability in raw material color and texture. In addition, our experience shows that sometime after refitting attempts decline in the rate of success, such attempts must be abandoned for several days or weeks, and then started again with a fresh mind. We found that this practice resulted in the number of refits multiplying. A similar result was obtained in experimental refitting conducted by Laughlin and Kelly (Laughlin & Kelly 2010). Their study shows that the number of refitting efforts increased considerably for flakes smaller than 5‑6 g. They also demonstrate that there is a significant negative relationship between assemblage size and success rates—the larger the assemblage, the lower the success rate. Similar to our results, they suggested it might be wise to restrict refitting to assemblages where the number of artifacts is < 2,000 if a high rate of success is deemed necessary (Laughlin & Kelly 2010: figs 1, 4).
61The utilization of refitting techniques as employed in this study makes a significant contribution to technological research. Techno-cultural industries have been traditionally defined on the basis of core type and aim, and to a lesser extent by their core shape (Bar-Yosef 1970; Bar-Yosef & Phillips 1977). The current work indicates that this approach must be taken with caution, and discarded cores should not be used as sole indicators of the technology used in specific industries. Following refitting we arrived at the conclusion that formal typology cannot be applied to discarded cores. Most of the cores initially defined as opposed platform types were redefined as single platform in light of the technological reconstructions. This change in perspective can be explained by the fact that one of the “platforms” was not intended for blank production, but was used to facilitate the knapper’s grasp during knapping and/or as part of the maintenance of the core’s distal end.
62Another contribution relates to typological definitions, as after refitting certain artifacts defined as belonging to a certain type must be redefined. For example, more than half of the core tablets produced in the process of core rejuvenation could not be classified typologically as such before refitting (figure 11.12). A second example are the La Mouillah points and notches that resulted from unsuccessful microburin technique preparation (see above).
63Previous studies on Epipalaeolithic assemblages indicated that many of the blade/lets were actually unused blanks (e.g. Bar-Yosef & Phillips 1977; Goring-Morris 1987). The current study, however, suggests that this was probably not the case, since a large proportion of the debitage at all of the sites studied were broken and/or displayed hinged or overpassed distal ends. Furthermore, the dimensions of these items were generally unsuitable as blanks. This conclusion should be viewed with caution, because it raises questions as to the significance of more traditional metrical and morphological technological studies based on samples of unmodified blade/lets among the debitage.
64In previous studies, refitting techniques were employed in order to better understand technological aspects, to conduct spatial analyses, and to reconstruct site formation processes (see above). The value of the present study is that it demonstrates that refitting as a methodological tool may contribute to a better understanding of the complexity of formal typological definitions.
Summary and concluding remarks
65One of the most important questions raised by technological studies is whether it is possible to define technological trends from which different social groups and structures can be identified. These technological trends include distinctions of knapping strategies and concepts of hunter-gatherer groups. This study demonstrates that the different Epipalaeolithic cultural entities shared a similar knapping concept and thus utilized the same knapping techniques. Nevertheless, several diachronic and synchronic differences between the industries were observed. The Mushabian differs from the Ramonian in core shaping and blank shape; The Ramonian differs from the Terminal Ramonian/Early Natufian, in raw material procurement strategies and exploitation pattern. Splitting of the core and the possible intentional heat treatment of chalcedony nodules is evident only in the Terminal Ramonian/Early Natufian (Goring-Morris et al. 1998; Marder 2002).
66The Epipalaeolithic cultural entities operated within a single technological system which included one principal variation, i.e. the presence or absence of microburin technique. Within this system, minor technological differences occurred through time and space. This technological system is characterized by efficiency and flexibility that permitted the manufacture of various microliths used in composite tools for diverse activities (such as utilization of vegetal resources and hunting) (e.g. Valla 1987; Valla et al. 1991; Edwards 2007; Richter 2007; Yaroshevich 2010, 2013). Such efficiency and flexibility are also illustrated through the knapping techniques, the presence of an “open” system, and in the use of pyrotechnology in the flint, ochre and plaster technologies (Koren 2000; Delage & Sunseri 2004). All these constitute the background for the transformation from a nomadic lifestyle of hunter-gatherers into a sedentary lifestyle that mainly occurred within the Mediterranean zone, although reflections of this process can be seen in the marginal areas, as demonstrated in this technological study (Goring-Morris & Belfer-Cohen 2013).
Acknowledgments
67We thank Angela Davidzon and Fawzi Ibrahim for the help in the laboratory work and for helpful supportive advice. The refits were photographed by Gabi Laron and Klara Amit and the artifacts illustration were drawn by Leonid Zaiger and Eli Cohen-Sasson. We would like to thank Mae Goder-Goldberger and Lotan Edeltin for their constructive comments on an early draft of this chapter.
Auteurs
Ben Gurion University of the Negev, P.O. Box 653 Beer Sheva 84105, Israel
Institute of Archaeology, The Hebrew University of Jerusalem, Jerusalem 91905, Israel
Le texte seul est utilisable sous licence Licence OpenEdition Books. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.
Not Just a Corridor
Human occupation of the Nile Valley and neighbouring regions between 75,000 and 15,000 years ago
Alice Leplongeon, Mae Goder-Goldberger et David Pleurdeau (dir.)
2020
La langue des bois
l’appropriation de la nature entre remords et mauvaise foi
Sergio Dalla Bernardina
2020
Living and working with giants
a multispecies ethnography of the Khamtis and elephants in Northeast India
Nicolas Lainé
2020
Nomad lives
From Prehistoric Times to the Present Day
Aline Averbouh, Nejma Goutas et Sophie Méry (dir.)
2021
L’art des ghostnets
Approche anthropologique et esthétique des filets-fantômes
Géraldine Le Roux
2022
Elusive Partners
Contemporary Anthropological Perspectives on Marine Species
Alix Levain, Hélène Artaud, Émilie Mariat-Roy et al.
2023
Des voies de l’ombre
Quand les chauves-souris sèment le trouble
Frédéric Laugrand et Antoine Laugrand
2023