Çatalhöyük in the Context of the Middle Eastern Neolithic

This review aims to show how the new results from Çatalhöyük in central Turkey contribute to wider theories about the Neolithic in Anatolia and the Middle East. I argue that many of the themes found in symbolism and daily practice at Çatalhöyük occur very early in the processes of village formation and the domestication of plants and animals throughout the region. These themes include a social focus on memory construction; a symbolic focus on wild animals, violence, and death; and a central dominant role for humans in relation to the animal world. These themes occur early enough throughout the region that we can claim they are integral to the development of settled life and the domestication of plants and animals. Particularly the focus on time depth in house sequences may have been part of the suite of conditions, along with environmental and ecological factors, that “selected for” sedentism and domestication. 105 A nn u. R ev . A nt hr op ol . 2 00 7. 36 :1 05 -1 20 . D ow nl oa de d fr om w w w .a nn ua lr ev ie w s. or g by U ni ve rs ity o f W as hi ng to n on 0 9/ 15 /1 1. F or p er so na l u se o nl y. ANRV323-AN36-07 ARI 13 August 2007 17:13 Anatolia: Asian region of Turkey, although the main focus here is on the region from central to southeastern Turkey Epipalaeolithic: time period between the Palaeolithic and the Neolithic. Associated with a changed lithic technology and more intensive subsistence strategies Kebaran: Epipalaeolthic groups in the Levant prior to the Natufian include those with material culture assemblages incorporating microlithic tools Levant: region in the eastern Mediterranean that now includes Israel, Palestine, the West Bank, Syria, Jordan, and Lebanon Natufian: cultural group that has distinctive material culture, lasts from approximately 12500 to 10000 bc, and is associated with predomesticated cultivation Pre-Pottery Neolithic A (PPNA): cultural group found in the Levant from ∼10000 to 8700 cal bc INTRODUCTION Çatalhöyük in central Turkey was first excavated by James Mellaart between 1961 and 1965. At that time the main impact of the site was to show that early settled villages existed outside the Fertile Crescent of the Middle East. The site also had a wide impact because of Mellaart’s (1967; Todd 1976) reconstructions of elaborate shrines with complex paintings, installations, and sculptures. Much of the symbolism of the Neolithic of the Middle East has been interpreted in terms of the bull and mother goddess themes that Mellaart thought were so prominent at Çatalhöyük (see for example Cauvin 1994). Since the 1960s, our understanding of the Neolithic of the Middle East has changed substantially. In particular, new finds from throughout the region have pushed back the dates of early settled life and have shown that the process is diverse—for example, the differences between the Levantine sequence and that in southeastern Turkey are marked. However, our understanding of Çatalhöyük has also changed as a result of new excavations started by Hodder in 1993 (Balter 2005; Dural 2007; Hodder 1996, 2000, 2005a,b,c, 2006, 2007). For example, it is clear that the symbolism at Çatalhöyük is part of domestic cults and that female imagery is only a small part of a diverse set in which mother and goddess characteristics are hard to find. The main focus of this review is on how the new results from Çatalhöyük fit into or challenge wider theories about the Neolithic in Anatolia and the Middle East. Çatalhöyük, dated to 7400–6000 bc (Cessford 2005; all dates here are calibrated), occurs a long time after the first sedentary settlements in the Middle East (which emerge in the period between the twelfth and ninth millennia bc) and well after the first domesticated plants (in the ninth millennium bc but see below for the debate about the dates). The Levantine sequence, described below, involves Epipalaeolithic groups such as the Kebaran and Natufian (the latter from approximately 12,500 bc to 10,000 bc) with increasingly intensive hunting, gathering, and cultivation of wild plants; followed by the Pre-Pottery Neolithic A (PPNA) from 10,000 to 8700 bc and Pre-Pottery Neolithic B (PPNB) from 8700 to 6800; followed by the Pre-Pottery Neolithic C (PPNC) and Pottery Neolithic (PN). Because of the polycentric character of the processes of sedentism and domestication (Gebel 2004) throughout the Middle Eastern and the Anatolian region, it is incorrect to use these terms and sequences outside the Levant, and other terms have been proposed for Anatolia (e.g., Özbaşaran & Buitenhuis 2002). However, the Levantine sequence is best understood and documented and provides a benchmark for the sequences elsewhere. As comparatively well known as the Levantine sequence may be, there remains little consensus about the causes of the emergence of sedentism in agglomerated villages and the domestication of plants and animals. Despite the late date of Çatalhöyük, the detailed evidence and the long-term projects at the site allow insight into the character of prepottery and early pottery agglomerated settlement in the region. The site has remarkably dense settlement (3500 to 8000 people in 13.5 ha) and was occupied for a long period. The Neolithic East mound is 21 m high, has 18 levels of occupation, and lasts 1400 years before settlement relocated to the West Mound on the other side of the river (the Çarşamba Çay in the flat Konya Plain) during the early Chalcolithic in the early sixth millennium bc. The Neolithic economy was based on a wide range of domesticated and wild plants (Fairbairn et al. 2005, Hastorf 2005) and based only partially on domesticated animals (sheep and goat—cattle and pig were not domesticated through the main Neolithic sequence according to Russell & Martin 2005). Çatalhöyük can thus provide some insight into the ways in which people lived in these early villages. Early theories of agricultural origins in the Middle East were based on single environmental, climatic, and population density causes. The last glacial maximum occurred at 106 Hodder A nn u. R ev . A nt hr op ol . 2 00 7. 36 :1 05 -1 20 . D ow nl oa de d fr om w w w .a nn ua lr ev ie w s. or g by U ni ve rs ity o f W as hi ng to n on 0 9/ 15 /1 1. F or p er so na l u se o nl y. ANRV323-AN36-07 ARI 13 August 2007 17:13 24,000 to 18,000 years ago when the region was cold and dry. The gradual change to warmer and wetter conditions after this time suffered a setback in the Younger Dryas (11,500 to 10,000 bc) during the second half of the Natufian. Bar-Yosef (2001) is among many that see the Younger Dryas conditions leading to intensification and then to PPNA and the first agriculture. One limitation of the climatic argument is that scholars now indicate that sustained domestication of plants did not occur at the end of the Younger Dryas in the PPNA but considerably later in the PPNB (Colledge et al. 2004, Nesbitt 2002, Willcox 2002). External causes of change have tended to be balanced during recent decades by theories that focus on social factors such as prestige exchange (Bender 1978), feasting (Hayden 1990), and symbolism (Cauvin 1994). Social factors may have provided the driving forces behind sedentism and intensification. Although evidence from the Natufian onward of large-scale communal building works, and of open areas used for roasting pits, is widespread, evidence throughout the region and period of marked social ranking, except at Çayönü in southeastern Turkey, is lacking (Özdoğan & Özdoğan 1990). Certainly recent finds have shown with great clarity that initial sedentism was closely tied to ritual. Landscapes may have been drawn together at ritual centers to which people came for initiation, feasting, burial, exchange, marriage, etc. (Schmidt 2000). In fact several of the early sites seem to have been ritual centers, whatever other functions they may have had. In north Syria and southeast Turkey, at sites such as Tell ‘Abr 3, Jerf el Ahmar, and Göbekli, one finds large PPNA buildings, circular and semisubterranean, which have generally been accepted as communal ritual buildings. Those at Tell ‘Abr 3 are 7–12 m in diameter (Yartah 2005). The internal furnishings of these communal buildings are certainly elaborate, but we need to avoid getting caught in a possibly inappropriate opposition of ritual versus domestic. At Pre-Pottery Neolithic B (PPNB): cultural group found in the Levant from 8700 to 6800 cal bc PN: Pottery Neolithic Tell ‘Abr 3, building B2 was dug 1.55 m into virgin soil and had a bench within its circular walls. This in turn was lined with stone slabs polished and decorated with wild animals. Bucrania (cattle skulls) were deposited in a bench. But in another building, M1, a hearth was found, and on the floor were found limestone basins and bowls as well as grinding stones (Yartah 2005). Indeed, Yartah (2005) argues that the large early PPNA communal buildings at Mureybet and Jerf el Ahmar are not elaborate ritually and symbolically and were probably used for stockage and multiple functions. But at the end of PPNA Yartah suggests that there is less evidence of economic functions and much decoration and ritual—e.g., at Jerf el Ahmar, Tell ‘Abr 3, and probably Göbekli. However, the interpretations of these buildings, their communal and domestic versus ritual nature, remain problematic until detailed accounts of floor residues and discard practices are available. The forensic work on the floors at Çatalhöyük shows that floors can be carefully cleaned and abandoned and that microresidues of activities can be discerned only with careful analysis. This work showed that the supposed “shrines” at Çatalhöyük were actually used as domestic houses (Bull et al. 2005, Matthews 2005, Middleton et al. 2005). Even if, as seems likely, social and ritual gathering was an important component of the processes that created permanent sedentary gatherings of people, we are left with the issue of why people adopted more elaborate and larger-scale social and ritual practices, including the fashioning and erection of large monoliths and semisubterranean circular structures, and all the investments of labor necessary for large-scale feasting and ritual. Disadvantages of economic intensification and of collective living in one spot can be cited: hard work (seen in stress markers on skeletons) and depletion of resources, sanitation, disease, etc. (Larsen 1995). So by which process did people submit themselves to greater work and intensification to achieve the benefits of social and ritual elaboration and sedentary village life? www.annualreviews.org • Çatalhöyük 107 A nn u. R ev . A nt hr op ol . 2 00 7. 36 :1 05 -1 20 . D ow nl oa de d fr om w w w .a nn ua lr ev ie w s. or g by U ni ve rs ity o f W as hi ng to n on 0 9/ 15 /1 1. F or p er so na l u se o nl y. ANRV323-AN36-07 ARI 13 August 2007 17:13 Many authors have summarized the social relations of hunter-gatherers (e.g., Ingold 1999, Meillassoux 1972, Sahlins 1972). In general scholars argue that in hunter-gatherer societies, the means of production are collectively owned, groups achieved reciprocal rights to the resources of other bands by asking permission, and studies show a lack of accumulation of personal wealth, with storage being only a technique for preparing for seasonal shortfalls. Ingold (1999) discusses the notion of “collective access” (p. 401), and social relations are immediate (Woodburn 1980) in that there is a lack of temporal depth in the relations between self and other (Ingold 1999). Formal institutions that structure social rules and regulations (p. 406) are relatively lacking. People trust good hunters, but they trust the hunters not to reduce their autonomy. A leader cannot place a person under obligation or compulsion because this action is a betrayal of trust. Such descriptions of hunter-gatherer society are difficult to apply to societies in the millennia that approach the domestication of plants and animals. An investment of labor already accompanied the more intensive economies of the Kebaran and Natufian, and social relations could be decreasingly described as immediate. We find little evidence for storage beyond that needed to tide over from season to season, and accumulation of personal wealth is limited right up into the PPNB. But there is undoubtedly an increased focus on temporal depth. As people depended more on things, and on intensive resource extraction and cultivation, they would have needed to depend on others to provide objects (in exchange), to tend objects (fields and animals, houses and boats), to construct objects (houses), to discard objects (organizing refuse and discard in dense villages), etc. One of the conditions that made agriculture possible in the Middle East was a changed relation to time and history. Rather than immediate and short-term relationships, societies in the region developed a strong sense of temporal depth tied to specific places well before domesticated plants and animals emerged. Intensive collecting and early farming involved delayed return systems (Woodburn 1980). But for delayed return systems to be viable (“selected for”), given the harder work and restrictions involved, there had also to be wider structural changes. One of these was a greater sense of temporal depth, history, and memory. Temporal depth is the main focus of this review, but I briefly consider two other regional conditions of possibility for sedentism and the emergence of farming. These possibilities include a symbolic focus on wild animals, violence, and death and a central dominant role for humans in relation to the animal world. REPETITIVE PRACTICES IN THE HOUSE AND MEMORY CONSTRUCTION One of the main results from the new excavations at Çatalhöyük is that the buildings Mellaart (1967) saw as static entities are now understood as the by-products of continuous processes. The new project has documented the extraordinary sequences of plasters on floors, walls, and relief sculptures. These monthly and yearly replasterings with their associated residues often occurred up to 450 times in houses that lasted 70 to 100 years. A house was then often rebuilt in the same place. The old house was dismantled, often carefully and with much careful cleaning and placing of objects, and filled in with clean soil, and the new house was built on the stumps of the walls of the previous house. In some places we have up to 6 rebuildings in the same place. The repetition of the ordering of social space within these building sequences is remarkable and has led to the hypothesis that social life was organized at least partly through the routines and practices of domestic socialization (Hodder 2006, Hodder & Cessford 2004). Embedded within a complex symbolic world, the daily activities within houses formed and reformed the social world. 108 Hodder A nn u. R ev . A nt hr op ol . 2 00 7. 36 :1 05 -1 20 . D ow nl oa de d fr om w w w .a nn ua lr ev ie w s. or g by U ni ve rs ity o f W as hi ng to n on 0 9/ 15 /1 1. F or p er so na l u se o nl y. ANRV323-AN36-07 ARI 13 August 2007 17:13 As well as these continuities in practices and functions in houses at Çatalhöyük, one finds very specific house-based continuities in the art and symbolism (Düring 2006, Hodder 2006). For example, the building that Mellaart called VI.8 had VII.8 below it, and in both cases investigators found stylized hands in horizontal rows. VIII.8 and VII.8 both had vulture scenes. But perhaps the best example was the repetition of the paired leopards in VII.44 and VI.44. An individual leopard and rather stylized fighting leopards were found in two other buildings (VIII.27 and VI.80), but these differed from the distinct pairings in building 44 in Levels VII and VI. Although some evidence shows feasting and prestige exchange at Çatalhöyük, the bulk of the evidence suggests that status and power were very much based on the control of people and their socialization within domestic units. But how widely applicable is this view that socialization through daily routines in houses (Watkins 2004, 2006) was an important mechanism for creating and maintaining social relationships and access to resources? This article looks at how similar interpretations might be relevant elsewhere (Nadel 2006), even though the preservation of detailed activity sequences is usually not as good as at Çatalhöyük. Some evidence at Çatalhöyük also indicates a practice of burying the dead beneath the floors of houses and then digging up and recirculating selected human heads before the final burial of these heads in foundation and abandonment deposits. Some evidence demonstrates the digging up of early relief sculptures and animal heads and their use in later houses and installations in houses. A good case can be made (Hodder 2006, Hodder & Cessford 2004) that the houses that invested more in the construction of long-term memories in these ways were also more socially and ritually successful. These houses tended to have more burials and to be more elaborate in terms of internal fixtures (Düring 2006, Hodder 2006). The “ancestral houses” are not larger than other houses, and they do not have more storage or productive facilities (Cutting 2005). Although plastered skulls have long been recognized in the Neolithic of the Levant, does evidence indicate the type of recirculation seen at Çatalhöyük? Can we argue that social power everywhere was based on the control of history and links to the past? We could argue that the repetition of houses in the same place results from the crowding and permanence of settlements. However, the specific continuities in function and art just alluded to at Çatalhöyük cannot be explained in this way; neither can the digging down and retrieval of earlier skulls and sculptures. In any case we see that repetition of house sites occurs very early in small, relatively short-term settlements. Certainly, by the time of the PPNA and PPNB the decreased residential mobility and intensity of habitation would have produced greater internal site organization (N.B. Goodale & I. Kuijt, circulated manuscript, 2006; Nadel 1998). But even in densely occupied settlements a number of strategies can be taken in locating new houses above, by, or near older houses (Tringham 2000). Rather, it seems that the repetition of houses and the construction of house-based memories were formative processes that played a part in producing sedentism, long-term duration in one place, and agglomerated settlement. Of course, repetitive practices took place early in the Palaeolithic. These involved repeated seasonal uses of the landscape in such a way that certain sites that provided shelter, such as cave sites, were returned to over long periods of time. For example, Ksar Akil in Lebanon has 23 m of deposit covering the period from the Middle Palaeolithic through the Upper Palaeolithic to the Kebaran EpiPalaeolithic. In the upper levels there was a “fine and complex stratigraphy” (Bergman 1987, p. 3). Kebara cave also has deposits spanning the Middle Palaeolithic and Natufian periods, or from ∼60,000 to 10,000 bc. The Middle Palaeolithic deposits show repeated use of part of the cave for hearths, while an inner part of the cave was used as a dump area (Goldberg 2001). The hearth area has deep www.annualreviews.org • Çatalhöyük 109 A nn u. R ev . A nt hr op ol . 2 00 7. 36 :1 05 -1 20 . D ow nl oa de d fr om w w w .a nn ua lr ev ie w s. or g by U ni ve rs ity o f W as hi ng to n on 0 9/ 15 /1 1. F or p er so na l u se o nl y. ANRV323-AN36-07 ARI 13 August 2007 17:13 deposits of overlapping hearths, each of which results from several episodes of combustion (Meignen et al. 2000, p. 14). These multiphase hearths indicate long periods of repetitive use in the same depression (p. 15), and similar processes are found in other sites in the Middle East. Many fire installations were vertically superimposed (p. 16) at Kebara, but the placing of these hearths was not exact. Rather investigators found a zone in the cave where, over a long period of time, people made hearths. Each hearth involved refirings, but the hearths themselves created a vertical palimpsest of overlaps. A part of the cave was generally used for hearths, but investigators did not find specific backward reference. The Kebaran in the Levant has lowland aggregation sites of 25–50 people and upland camps of 14–17 people, and there may have been seasonal cycles of aggregation and dispersal. Little architecture has been excavated, but evidence shows a possible twicea-year or even year-long occupation in the early Kebaran at Ohalo II about 21,500 years ago (Nadel 1990). The largest hut at Ohalo II had three successive floors and erect stones as well as a probable stone arrangement under them (Nadel 2006). The mud floors were covered with rich artifact debris, probably in situ. Nadel (2006) suggests a clear focus on continuity of place. Burial beneath floors probably occurred in the Kebaran at Kharaneh IV and Ein Gev (Valla 1991). At Ein Gev 1 in the Jordan Valley in Israel investigators found a fourteenth-millennium-bc Kebaran site on the eastern side of the Sea of Galilee (Arensburg & Bar-Yosef 1973). A hut was found dug into the slope of a hill. “The hut was periodically occupied as indicated by six successive layers which accumulated within it” (Arensburg & Bar-Yosef 1973, p. 201). Each layer had a floor 5–7 m in diameter littered with artifacts and bones, covered by a sandy layer that included artifacts. In section, the floors clearly repeat each other, and from one of the middle floors a grave was cut. Evidence does not indicate specific repetitions of feature or artifact placements, but this example clearly indicates some specific backward reference in the location of a house structure, even in the absence of permanent occupation. In the Natufian we see some degree of sedentism. ‘Ain Mallaha has animals and birds from all seasons (Valla 1991), and commensals (such as the house mouse) indicate sedentism. Settlements occur in the hill zones of Israel, Jordan, Lebanon, and Syria, and related sites are found to the north in Mureybet and Abu Hureyra. The later Natufian starts at the same time as the Younger Dryas climatic deterioration. In the Levant in the later Natufian, many but not all hamlets dispersed and became more mobile (Bar-Yosef 2001). But in the Taurus in southeastern Turkey and adjacent areas, the response to the Younger Dryas may have been greater sedentism at sites such as Hallan Çemi (Bar-Yosef 2004). Investigators noted both base camps and short-term intermittent sites in the Natufian. In the short-term sites, there is little evidence of repetitive practices, for example, at Hatula and Beidha (Byrd 1989, Ronen & Lechevallier 1991). Even in substantial Natufian sites we find little evidence of structured repetition. Valla (1991) notes that it is often difficult to follow coherent levels of habitation in Natufian sites, and it is difficult to show the absolute contemporaneity of buildings (see also Kenyon 1981, Moore et al. 2000). However, in the early Natufian site of Wadi Hammeh 27 in the central Jordan valley there is “a continuity in spatial arrangement of constructed features through successive phases” (Edwards 1991, p. 125). The earliest evidence of Natufian occupation at Hayonim Cave is Grave XIII “which was covered by the floor of Locus 3”—that is, by one of the structures with undressed stone walls (Bar-Yosef 1991, p. 86). At ‘Ain Mallaha we definitely find superpositioning of houses. In the “ancient level” houses, 131, 51, and 62–73 succeeded each other on the same spot (Perrot 1966). And in the “recent level” houses we find another sequence of houses dug into each other (houses 26, 45, and 22). In the Final Natufian at Mallaha, each major building had 110 Hodder A nn u. R ev . A nt hr op ol . 2 00 7. 36 :1 05 -1 20 . D ow nl oa de d fr om w w w .a nn ua lr ev ie w s. or g by U ni ve rs ity o f W as hi ng to n on 0 9/ 15 /1 1. F or p er so na l u se o nl y. ANRV323-AN36-07 ARI 13 August 2007 17:13 a succession of floors, one on top of another, with no sterile layers between (i.e., no abandonment fill) (N. Samuelian, H. Khalaily, F.R. Valla, circulated manuscript, 2003). At Çatalhöyük important evidence for memory construction is the removal, circulation, and reuse of human skulls. By the end of the Natufian evidence indicates the removal of the human skull after death, although in the absence of evidence for circulation and reuse, this does not by itself indicate the construction of historical links to ancestors. Skull removal may have had other roles such as healing, divination, etc. Skeletons were found within the houses at Mallaha, but the stratigraphical positioning is often unclear in Valla (1991). According to the reanalysis by Boyd (1995) the 131–51–62–73 sequence of buildings started with 12 skeletons beneath the floor of 131. He draws attention to the continuity of activity in the same place starting with a set of burials. At Çatalhöyük the focus on repetitive practices in the house and on memory construction is associated with careful and elaborate abandonment practices, including the placement of objects and the filling of houses with clean earth before rebuilding. For societies in which temporal depth and memory construction are important, ending and starting buildings are likely to be significant events surrounded in ritual. Did such practices already occur in the Natufian? In the ruins of one house at Mallaha investigators found several boar heads (Valla 1991), which could indicate ritualized abandonment processes. In what he called Abri 26 at Mallaha, Perrot (1966) found a child skeleton and necklace on the abandoned floor. Complete basalt artifacts were found discarded or cached on interior floors at Wadi Hammeh 27 (Edwards 1991), but it is not clear whether they were just abandoned in a context of use or whether this act was ritualized in some way. In the PPNA in the Levant, settlements were 0.2 to 2.5 hectares in size and are thus 3 to 8 times larger than the largest Natufian sites (Bar-Yosef 2001). The houses were often oval and semisubterranean, with internal hearths and plaster floors. As in northern Syria, mounds were often long-lived. Jerf el Ahmar had at least 10 building levels comprising ∼800 years of settlement (Akkermans 2004, p. 287). PPNA and related sites were also often much more structured than most Natufian sites. Nadel (1998, p. 9) has noted that “in Natufian and other Epipalaeolithic sites, it is common to find the entire range of typological variability in each site, and even in each locus . . . However, in PPNA cases, it is common to find typological differences between assemblages from contemporaneous loci at a site.” N.B. Goodale & I. Kuijt (circulated manuscript, 2006) have noted a similar shift in the way that sites are formed, as a result of their work at ‘Iraq ed-Dubb in Jordan. Here a late Natufian occupation “had fairly nondelineated use of space compared to a more delineated use of space during the PPNA.” We see much more evidence of repeated use of the same space or house in the PPNA throughout the region. Qermez Dere in northern Iraq has good evidence of rebuilding in the same place (Watkins 2004, 2006). In Phase II at Mureybet on the Middle Euphrates investigators found round houses that were superimposed on an Epi-Natufian house xxxvii. “Trois niveaux d’habitation en maisons rondes se superposent directement à la maison xxxvii de la phase IB. Il s’agit manifestement de la reutilization du meme espace d’habitat en continuité directe avec la période épinatoufienne” (Cauvin 1979, p. 26). In part of the site they found five levels of occupation in this phase. At Jericho in Trench DII Kenyon (1981) found a huge amount of very repetitive surfaces adjacent to the tower in PPNA— between the tower and adjacent circular enclosures. It is inside the settlement that one sees most residential continuity in PPNA and PPNB deposits, although, on the whole, walls were cut down further than at Çatalhöyük. In PPNA in Squares EI, EII, and EV there were 24 main building phases. In most cases there Kenyon saw only 2–4 floors for each www.annualreviews.org • Çatalhöyük 111 A nn u. R ev . A nt hr op ol . 2 00 7. 36 :1 05 -1 20 . D ow nl oa de d fr om w w w .a nn ua lr ev ie w s. or g by U ni ve rs ity o f W as hi ng to n on 0 9/ 15 /1 1. F or p er so na l u se o nl y. ANRV323-AN36-07 ARI 13 August 2007 17:13 building phase. “Some of the houses lasted through several phases, but usually with rebuildings almost from the base of the walls. Associated with most of the phases was usually a long succession of surfaces, particularly in the courtyard areas linking the various buildings” (Kenyon 1981, p. 269). The greater delineation of space in PPNA sites has already been noted and is relevant to abandonment and foundation processes. One can find more evidence of refuse management practices, with separate middens and more cleaning out of houses on abandonment (Hardy-Smith & Edwards 2004, Rosenberg & Redding 2000; N.B. Goodale & I. Kuijt, circulated manuscript, 2006). In PPNA at Jericho in trenches EI, EII, and EV Kenyon found one building with a central stone lined post socket under which was an infant burial (Kenyon 1981), which may represent a foundation deposit. In Square M1 in PPNA in phase xlii in house MM the clay floor had a foundation of cobble stones. “Set in the cobbles, but sealed by the clay floor, and therefore contemporary with the construction of the building, were two burials” (Kenyon 1981, p. 232). Skull removal also occurred in the PPNA (Bar-Yosef 2001). At Jerf el Ahmar in northern Syria, in Village 1/east there Stordeur found a sunken building with wooden posts to hold up the roof. At the bottom of one of these posts “two human skulls were found” (Stordeur 2000, p. 1). These findings begin to suggest the specific use of skulls to build histories in houses, although the use of skulls in this way may have been simply protective or magical. Yet the use suggests that links to the past and past individuals were of increasing salience. Turning to the PPNB in the Levant, ‘Ain Ghazal has frequent floor replasterings (Banning 2003), but perhaps the best evidence is from the extensive excavations and soundings at Jericho. As in PPNA, walls are built on walls and floors are repeated inside houses. So in EI, EII, and EV, in phase xlvii “the levels in the northern room of the eastern range [of rooms] were gradually raised by a series of floors. . . . The numerous floor levels suggest a prolonged period of use” (Kenyon 1981, p. 295). But the best evidence for repeated surfaces was in the outside, courtyard areas between buildings. The courtyards had alternating layers of clay or mud floors and spreads of charcoal (Kenyon 1981, p. 294). Kenyon found hearths in these areas, but she did not plan these; therefore, we cannot determine whether location of hearths was repetitive in outside areas. In Jordan at PPNB Beidha, “the inhabitants were extremely conservative in their siting of the different elements of the village” (Kirkbride 1966, p. 14). In one building at Beidha the total thickness of the multiple plaster layers was more than 5.5 cm, and parallels were drawn with Çatalhöyük (p. 18). At Abu Hureyra 2 “each house was usually constructed on the remains of an earlier one, and the form of that building largely determined the plan of its successor” (Moore et al. 2000, p. 262). The rooms of the ruined house were filled in and the stubs of the walls cut down. “The houses in Trench E were rebuilt four, and the houses in Trench B no fewer than nine times” (p. 266). Floors were renewed at least 2–3 times, and sometimes up to 10 times. Walls also had mud plaster or whitewash refreshed several times during a room’s life. “The hearths were often set in the same place in successive houses” (p. 265), e.g., the series of hearths in houses of phases 2–7 in Trench B. “We conclude from this that the builders of a new house often remembered not only the plan but also the internal arrangements of its predecessor, and considered it appropriate to replicate both” (p. 265). “We know, too, that in some instances they themselves were the descendants of the inhabitants of the earlier structures” (p. 266) because some distinctive skeletal and dental traits that are probably genetically transmitted were identified in house burials. In southeastern Turkey at Çayönü there seems at first sight to be much more evidence of conformity within phases than between phases because houses changed in form 112 Hodder A nn u. R ev . A nt hr op ol . 2 00 7. 36 :1 05 -1 20 . D ow nl oa de d fr om w w w .a nn ua lr ev ie w s. or g by U ni ve rs ity o f W as hi ng to n on 0 9/ 15 /1 1. F or p er so na l u se o nl y. ANRV323-AN36-07 ARI 13 August 2007 17:13 from round to grill to channeled to pebble paved to cell to large room. We see a striking homogeneity of building types in each building layer (Özdoğan & Özdoğan 1990, p. 72). Thus investigators indicate more of a focus on horizontal similarity than on vertical continuity. However, even here Özdoğan & Özdoğan (1990, p. 73) argue that “in every building layer, the foundations of the new building are always directly on top of the preceding one, without disturbing or reusing its stones.” Several buildings are mentioned as having several rebuilds, and the Skull Building went through at least five major rebuilds. At Aşıklı Höyük in central Turkey, dated to the late ninth and early eighth millennia bc, “in one of the excavated rooms, ‘room A’ (trench 3K . . . ) 13 floor levels have been recognized” (Düring 2006, p. 73). At this site variation between houses in memory construction is a possibility. Only 35% of rooms have hearths at this site, but in the deep sounding, a building was knocked down and rebuilt in the same place at least 7 times, a practice that continued throughout the entire 8-m-deep sequence in the mound (Esin & Harmankaya 1999). In each rebuilding a hearth is seen in exactly the same position. Given the relatively small percentage of buildings with hearths, this evidence suggests that some buildings passed down the practices of hearth use, whereas others did not. We also find much continuity at the site in terms of the location of the major street by the “ritual complex” and the location of midden areas (in the deep sounding). The emphasis on continuity of houses seen at Aşıklı Höyük and Çatalhöyük is also found elsewhere in the Ceramic Neolithic in central Anatolia (Düring 2006, p. 236). Much evidence indicates repetitive practices in houses and memory construction in the PPNB and related groups in the Middle East and Turkey. Evidence also suggest abandonment and foundation practices, although walls were generally cut down much more than at Çatalhöyük. Burning of houses, associated with death, is found as part of abandonment practices (Verhoeven 1999, 2000, 2002). Heads tend to be found in groups in the Levant, sometimes with features plastered on, but how much they were circulated is not clear. Investigators found male and female skulls, as well as subadults, raising the question of whether the skulls represent ancestor veneration at all rather than apotropaic or other protective functions (Bonogofsky 2004, Talalay 2004). However, the depositional contexts of some skull deposition suggest practices that may have involved backward or forward reference. The skull of a child was found between the stones of the foundations of Wall E180 at PPNB Jericho (Kenyon 1981). In phase lxi in a room in a house in EI, EII, EV the cranium of an elderly man was set upright in the corner about 15 cm below floor level. In EIII-IV a plastered skull was found in a building fill. GoringMorris (2000, p. 119) argues that many PPNB burials definitely stratigraphically predated the construction of the overlying architectural features and floors. For example, “in at least three instances at Kfar HaHoresh burial pits clearly stratigraphically underlie and are sealed by plaster surfaces” (p. 119). In some cases we see a time lapse between burial and/or skull removal and the making of the floor. Thus buildings “remembered” the location of the burials or skulls. Sometimes there is evidence of markers above the burials or skulls. Goring-Morris suggests that constructing buildings in relation to earlier buildings may have started at Mallaha in the Levant (see above). Special abandonment practices are found at Çayönü—for example, in the Cell phase investigators found blocking of doorways, and intact artifacts are abandoned in cell rooms (Özdoğan & Özdoğan 1990). Charnel houses or buildings for the dead occur at Çayönü (the Skull Building) and at Abu Hureyra and Dja’de el Mughara (the Maison des Morts) in Syria (Akkermans 2004, p. 289). Through much of the region in the PPNB evidence indicates circulation and handing down of artifacts through time. Practices of www.annualreviews.org • Çatalhöyük 113 A nn u. R ev . A nt hr op ol . 2 00 7. 36 :1 05 -1 20 . D ow nl oa de d fr om w w w .a nn ua lr ev ie w s. or g by U ni ve rs ity o f W as hi ng to n on 0 9/ 15 /1 1. F or p er so na l u se o nl y. ANRV323-AN36-07 ARI 13 August 2007 17:13 stone recirculation and reuse were found at Çayönü. Standing stones up to 2 m high were found in the plaza and in the Skull and Flagstone ceremonial buildings. “Some of the standing stones were intentionally broken and then buried under the subsequent reflooring of the plaza” (Özdoğan & Özdoğan 1990, p. 74). At Jericho in the PPNB levels Kenyon found a large bituminous block (Kenyon 1981, pp. 306–7). It had been carefully flaked and was obtained from the Nebi Musa district some 17 miles away. Investigators found it in the foundation of wall E223 of phase lxv. But it fit exactly into a niche of the earlier phase lxiv, where it probably stood on a stone set on a pillar of earth on which there were traces of plaster. So this stone had a role in phase lxiv and was then reused in the foundation of lv. In phase lxiii this same room had a distinctive green clay floor, all suggesting that this part of the building had a special character over three phases. Overall, then, the pre-Neolithic and Neolithic societies of the Middle East and Turkey were increasingly concerned with temporal depth. Evidence increasingly suggested repetitive practices in houses, and sometimes in outside areas (e.g., courtyard or midden areas at Jericho and Aşıklı Höyük), as well as in public spaces such as paved streets (at Aşıklı Höyük). Evidence of specific memory construction as houses are built over burials, or skulls and other objects are circulated and passed down through time is also increasing. The concern with time depth, history, and memory reaches its apogee in the PPNB at the same time that domesticated plants appear in quantity, but it starts to emerge at least by Kebaran and Natufian times, even in contexts in which sedentism is limited. It is difficult to explain the focus on temporal depth as the result of living in dense villages. Rather, the emergence of greater temporal depth was a necessary condition for dense settled life, the delayed returns of intensive subsistence systems, and the shift to domesticated plants and animals, as well as for the staging of large-scale feasts, exchanges, and marriages. THE RITUAL VALUE OF VIOLENCE, DEATH, AND SEX The new work at Çatalhöyük has shown that much of the symbolism, far from having a focus on a nurturing mother, centers on violence, death, and perhaps sex. The horns, teeth, claws, and beaks of animals and birds are preferentially brought into the house and installed on or in house walls (Hodder 2006, Russell & Meece 2005). Investigators found scenes of humans teasing, baiting, and killing wild animals, and they found phallic images (Meskell 2007). At Çatalhöyük the prevalence of such imagery in domestic and ritual contexts demonstrates that much social life was embedded within these themes. They also found scenes in the paintings that suggest that death and violence were linked to rites of passage in which the social order was created and recreated through moments of transcendence and transformation (Bataille 1962, Bloch 1992, Hodder 2006). It is of interest, therefore, that themes of death, violence, and sexuality dominate the symbolism of earlier sites such as Göbekli Tepe (Schmidt 2003). This particular suite of ideas, linked to ceremony and public feasting, was central to the creation of large-scale social order on which settled village or early town life depended. At Nahal Oren and other Natufian sites Noy found carved stones with incised decoration, and animal heads carved on bone handles (e.g., of sickles) (Noy 1991). The sickle shafts from El Wad and Kebara are in the form of deer and goat heads (Henry 1989). Fox (Vulpes sp.) teeth are widely used as raw materials for pendants (Goring-Morris & BelferCohen 2002, p. 70). In the Natufian we see a marked rise in the numbers of raptor talons (Goring-Morris & Belfer-Cohen 2002, p. 71) and pendants of bone and canine teeth (Henry 1989). Investigators also found phallic objects. “Natufian art also had an erotic element” seen in a calcite statuette from Ain Sakhri (Henry 1989, p. 206). In the PPNA wild cattle imagery is found throughout the region (Goring-Morris & 114 Hodder A nn u. R ev . A nt hr op ol . 2 00 7. 36 :1 05 -1 20 . D ow nl oa de d fr om w w w .a nn ua lr ev ie w s. or g by U ni ve rs ity o f W as hi ng to n on 0 9/ 15 /1 1. F or p er so na l u se o nl y. ANRV323-AN36-07 ARI 13 August 2007 17:13 Belfer-Cohen 2002). At Tell ‘Abr 3, a series of stone slabs line the bench around the walls (Yartah 2005) in building B2. These are polished and decorated with wild animals— gazelle, panther, aurochs—as well as with geometric designs. The panthers are spotted and highly stylized and look rather like lizards. Bucrania are deposited within a bench, but there are also bucrania on view in smaller buildings, interpreted as houses, at the site. At Jerf el Ahmar investigators also found a building with four cattle bucrania probably suspended on the interior walls (Stordeur 2000, Yartah 2005). At Jerf el Ahmar there is also serpent decoration on the stone slabs of the benches of the large circular buildings (Stordeur 2000), along with a separate depiction of a vulture (for parallel symbolism at Hallan Çemi and Nemrik 9 see Rosenberg & Redding 2000, p. 45; Kozlowski 1992). At Göbekli Tepe in the PPNA and early PPNB, megalithic pillars have reliefs of snakes, foxes, wild boar, cattle, gazelle, wild ass, lion, scorpions, spiders, water birds, and centipedes. The fox and wild boar have erect penises (Schmidt 2003). The reliefs also show a headless human body with an erect penis. In the PPNB there continues to be a widespread symbolic focus on the fox, wild cattle, wild boar, and birds of prey (Goring-Morris & Belfer-Cohen 2002, pp. 70–71). At Nevali Çori, Hauptmann (1999) reconstructs a large stone statue of a man holding his erect penis. This association in the early village societies of the Middle East with violence, sex, and death in the symbolic imagery could be interpreted in many ways (Hodder 2006, Verhoeven 2002). But perhaps at the simplest level, we can say that these associations give power (Guenther 1999). The powers to give feasts, to provide, and to protect would be enhanced by the images of violence, sex, and death. At Çatalhöyük there is an association between feasting deposits and wild male cattle. The art shows large numbers of people engaged in the killing of dangerous animals such as bulls that then appear in the feasting residues and in the installations in houses. In Building 1 at Çatalhöyük, a set of wild goat horns covers and perhaps protects a bin of lentils (Hodder 2006). The key aspect of giving a feast may not have been simply the provision of calories, but also the demonstration of intercession with and control of wild animals and the use of their powers to protect and nurture. The demonstration of power in relation to wild animals and animal spirits created the basis for building the long-term social structures of sedentary and then agricultural societies. The ability to harness the power of animals may have attracted followers and allowed the creation of trust and dependencies. The existence of an elaborate symbolic world of violence, danger, and sexual power, and the ability to intercede with the ancestors, may have created the conditions in which sedentary life and intensive delayed-return economies became possible (selected for). HUMAN AND MATERIAL AGENCY In many hunter-gatherer societies animals must be hunted with respect (Fowler & Turner 1999, p. 422). Appropriate prayers must be offered to the spirits of the animals if humans are to expect the animals to yield up their lives to the hunters. If humans cease predation the animals will do less well and decline in numbers (p. 422). There is a relationship of friendship and respect, of reciprocity and complementarity between hunter and game animal (Guenther 1999). “Hunters maintain relations of trust with their animal prey . . . assuming that animals present themselves with hunters in mind, allowing themselves to be taken so long as hunters treat them with respect and do nothing to curb their autonomy of action” (Ingold 1999, p. 409). Powerful hunters attract animals as they attract followers. They inhabit a “giving environment” so that “far from seeking control over nature, their aim is to maintain proper relationships with these beings” in the natural world (Ingold 1999, p. 409). www.annualreviews.org • Çatalhöyük 115 A nn u. R ev . A nt hr op ol . 2 00 7. 36 :1 05 -1 20 . D ow nl oa de d fr om w w w .a nn ua lr ev ie w s. or g by U ni ve rs ity o f W as hi ng to n on 0 9/ 15 /1 1. F or p er so na l u se o nl y. ANRV323-AN36-07 ARI 13 August 2007 17:13 But already at Göbekli we see something very different from this perspective. In the PPNA and early PPNB there are 2–3 m high stone stele on which have been carved the images of wild animals, birds, and insects described above. “The pillars themselves clearly have an anthropomorphic meaning” (Schmidt 2003, p. 3) and the human figure is entirely dominant. This is no longer a balanced relationship. Even more clearly at Çatalhöyük are scenes in the art showing people teasing, baiting, and playing with wild animals such as bulls, bear, wild boar, and stags. In these cases the humans are dominating animals and interfering with them. This increased importance of the dominant or active human in the symbolic or mythic world may also be linked to the ancestors. But the result was that the conditions were created for the domestication of animals, which involved interfering with and dominating animals (Cauvin 1994, Ingold 1999).