 |
|||||||||||||
|
|
|
|
|
|
|
 |
|||||
|
Contents:
7.1 The Basics of Discussion re Dispersal As Jones says, “the significance of agriculture cannot be elucidated in terms of its origins alone, but involves a more detailed understanding of the emergent structure of its continuing spread” (1996, p.97). There are two broad schools of thought:
7.1.2 Types of Evidence Available The adoption of agriculture took place independently in a number of different geographical regions and environmental zones around the world, all at much the same time, but it is clear that the processes that drove and characterized these changes and their spread from the core areas where they were first innovated were different in each of these areas. In this section the Near Eastern version of these events and the consequent dispersal of these economic practices and associated social changes is the only one discussed. There are a number of ways of trying to track the dispersal of agriculture in geographical, social and interactive terms:
7.1.2.1 The archaeological record A number of writers (summarized by Bellwood 2005, p.10) have pointed to the dangers of trying to discuss agriculture at different scales. At a local scale differences are clearly represented, while on a much wider scale these individualistic features are no longer visible. Establishing a balance between these perspectives is necessary to get a clear picture of what happened. The presence of desert on the east and the Mediterranean on the west helps to concentrate populations within the Levantine area, preventing straightforward expansion and forcing dispersal channels. 7.1.2.2 Genetics (both human and plant populations) Genetic studies can be used in two ways in this context – extrapolating backwards from modern genetic data or to use ancient genetic data to attempt to make contemporary comparisons with a view understanding population spread. 7.1.2.3 Comparative linguistics (extrapolating backwards from modern languages) All the discussions about linguistics that follow rely very heavily on Bellwood’s excellent ‘First Farmers’, chapter 9 (2005), and to a much lesser degree on Renfrew (1996) and Muzzolini (Adams and Friedman 1992). Explanations for language ancestry may give indications about shared language origins. Although languages may appear to be very different from each other, today, similarities indicate that certain components have been inherited from a common source language. The nature of language is a divergence, not convergence – in terms of regional developments, dialects, and linguistic forms: “the fundamental observation that the ancestral languages in a family have somehow spread outward from a source region and then diversified” (Bellwood p.182). Some linguists believe that recognizable traits of an evolved language can only be traced back as far as between 10,000 and 7,000 years, and that beyond this time language would not be identifiable as a member of an ancestral group. If a language type appears 100s or 1000s of miles from its source, it is necessary to find a mechanism to account for it. This may include only few processes:
Bellwood (2005, p.182) rejects language mixing as a viable option: “language families, like animal species cannot form because of convergence of taxa originally unrelated, such as Germaic and Semitic languages . . . to form a single genetic taxa within which all units share a common ancestry”. He goes on: “Historical data indicate that language shift alone, without population movement or some degree of dispersal by the population carrying the target language, has never created anything remotely equaling those intercontinental genetic groupings of languages with which we are concerned . . . . Imperial conquest by itself, without large-scale and permanent settlement by members of the conquering population, generally imposes little apart from loan words in the long term. Trade also is generally of little significance as a factor behind large-scale language spread”. He cites a number of examples including the spread of Arabic into areas like Egypt and the spread of English into America and Australasia, versus Latin which died out of all non-core areas of the Roman Empire and Dutch in Indonesia. He points out that it is difficult for humans to learn new languages beyond a certain age, that there needs to be a good reason to learn a new language and that there may be social resistance to new language adoption (p.197). He therefore emphasizes that genetic language spread can only have happened by actual population movement and integration and not by language shift alone” (p.193). It should be pointed out that Bellwood’s view on this matter is not the only one. Some linguists do support the idea of language shift. For example, Ehret (1988) believes that hunter-gatherer groups who are in contact with farming groups can adopt new languages even when their own subsistence practices remain unchanged. This does not agree with other studies, based on both ethnographic and historical data, which have suggested that existing populations do not adopt new languages. Looking at attempts to detect the age of language, its rate of spread, and possible geographical origins, Bellwood emphasizes the difficulties of the task, saying that language “does not yield time depth easily” (2005, p.193). Linguist Morris Swadesh developed what he called the glottochronolgical formula which claimed that “a random core vocabulary replacement in any language of 19.5% per 1000 years” could be detected (Bellwood 2005, p.194). He and others have used this to attempt to calculate the age of different languages and some archaeologists have quoted some of their findings in an attempt to tie in language spread and the dispersal of agriculture (e.g. Muzzolini 1992, p.149). However, for very good reasons, not everyone is convinced that languages change at a steady rate over time, and even most of those who use it believe that this is useful only for the last 500 to 2500 years. Impacts on linguistic patterns in prehistory cannot be recreated so it is difficult to employ the glottochronological clock. 7.1.3 Prerequisites for Dispersal Once agriculture was established in the Near East, it dispersed into Europe, towards southeast Asia and into northern Africa by any one or a combination of a number of possible mechanisms. The following conditions would have to apply in any event:
If the physical dispersal model is favoured the following also need to apply:
If the ideas dispersal model is favoured, the following also need to apply:
Dispersals from southwest Asia went in three directions – Europe, southeast Asia and northern Africa. They were not the only unique inventions of agriculture to be dispersed, by different mechanisms, but they are the only ones examined here. In each case, dispersals may have been physical (by migration), conceptual (ideas and technologies being adopted by indigenous groups) or admixture (a combination of both). They took place in different geomorphologies and climatic conditions. The reasons for dispersal, whether by physical movement or indigenous adoption, could have been many and varied, but most proposals tally very much with those first proposed as suggested reasons for the adoption of pristine agriculture. However, additional processes were also probably very important – particularly involving boundary issues and new social complexities. 7.2 Mechanisms and Speed of Spread The mechanism by which agriculture spread from the Near East into surrounding areas has been the subject of extensive theories and discussions, and is almost as difficult to track as the original establishment of agriculture in the Levant. Over the period of only a few thousand years, farming reached as far west as Britain, as far east as southeast Asia and as far south as Egypt (beyond which the Levantine domestic package was not ecologically suitable). The establishment of agriculture outside its original area of establishment is usually referred to as “secondary” establishment (as opposed to “pristine” or “primary” establishment in the Levant). In discussions of the spread of agriculture in the Near East, which is often described in terms of an ever-moving front, the conceptual nature of frontiers between two types of subsistence strategy and social organization that is inevitably raised. – that between hunter-gatherer and agricultural communities - Differing mechanisms of physical migration of agriculturalists into new areas, cultural and conceptual adoption by local populations, and interaction between different populations are all factors that need to be considered. For example, the physical impact of agriculture on the landscape, local resources, social consolidation and complex internal organization, and population expansion could all be relatively important factors. Similarly, Zvelebil and Rowley-Conwy (1986) suggest that a delay in a shift to agriculture would be expected, and is visible in more sedentary and complex hunter-gatherer communities. How agriculture spread is in part a discussion of how agricultural communities impacted the landscape and the much wider territories of hunter-gatherer communities. “What the neighbours do can have a profound effect on what one’s community does” (Fellner p.120). The sedentism of one group could impact a neighbour’s activities. Similarly, the nature of existing hunter-gatherer communities would impact the possibility of change of use of territories and the likelihood of adoption of new methods by the indigenous communities: “Hunter-gatherer societies did not adopt farming indiscriminately, but selectively, to fit the local needs. These needs varied from region to region, and so did the factors which combined to bring about the shift to food production. It follows, therefore, that there is no single cause for the transition which would fit all situations” (Zvelebil 1986b, p.167) Zvelebil (1986) describes three main paradigms for the transition to agriculture by hunter-gatherer populations who are exposed to it:
Cavalli-Sforza and Ammerman (1984) suggest a gradual ripple-in-the-pond model of spread – widespread all-encompassing colonization of new areas, a suggestion derived from the human genetic record. Rindos (1984) also takes a gradualist approach, seeing humans altering the environment to which the plants ancestral to domesticates were indigenous – he believes that changes to plant morphologies and behaviours were caused by human alterations to their environments, and that human exploitation of plants within these environments caused these changes to be enhanced over the centuries. More complex models see a more staggered “pulse” like rate of expansion which do not completely eliminate hunter-gatherers in agricultural areas, and see pockets of agricultural activity in largely hunter-gatherer territories, as well as wider areas of replacement (Dennell 1985). This is similar to Gould’s (1989) “pulse” explanation of evolution in natural history. Bellwood suggests (2005, p.84) “As long as there are niches, hunters can of course survive for millennia amongst farmers, but only if the conditions are right”. Zvelebil and Rowley-Conwy (1986) see the adoption of agriculture as a series of regional phases, with local adaptations to some elements of farming often after a long period of proximity to neighbouring farming activities (p.85). Full adoption, in this model, would follow later. The model moves away from “simple frontier dichotomies” and instead emphasizes “the extended duration and complexity of the process” (p.86). Zvelebil describes three phases of transition to agriculture, which attempt to model the processes involved without attempting to explain the cause (1986), as follows:
A lot of space has been given to the colonization and admixture approaches. Another opinion has been put forward by Dennell (1983) who says that the idea of colonization “still tends to exercise a powerful influence on explanations of the Mesolithic-Neolithic interface. Mesolithic societies are still assumed to have played no part in the expansion of food production over their own economies, and the possibility that they developed indigenously into food-producing societies is rarely considered” (p.155). He offers yet another view of the way in which agriculture disperses, but with an emphasis on indigenous adoption rather than colonization. His three-phase model is as follows:
Bellwood (2005) discusses four conceptual approaches to agricultural origins and spread, which he describes as “zonal” (p.274):
Bellwood suggests that “the rate of spread of farming increased very rapidly as the degree of environmental difficulty declined and as technological capacity increased. Without a need for any adjustment to differing environmental circumstances, cereal-based cultures in spread mode could generate sufficient demographical impetus to snowball along very far and very fast indeed” (2005, p.277). He sees the patterns of dispersal as one of pulsation at intervals, often with long periods of stability between. He believes that it is possible that both language and early agricultural economies spread through areas occupied by hunter-gatherers primarily via processes of “population growth and dispersal, but with admixture” (p.278). Cavalli-Sforza has made a much-quoted observation (1996, p.53) that “the rate of advance of farmers from southwest Asia to Europe can be measured from archaeological data; it is 1km per year, on average, and somewhat higher along the coasts of the Mediterranean than the interior of Europe”. I would draw attention to the “on average” – Cavalli-Sforza is not suggesting here that this was an absolutely even rate of dispersal, temporally speaking, but he is pointing to its rate. Jones et al (1996) describe the spread as one that “would surprise a demographic biologist by its sluggishness rather than it speed” (p.95). In terms of what the effect of population expansion actually was, in terms of increased numbers of individuals, Hassan (1981) calculates that agriculture caused a rapid increase in world population from 10 million to 50 million. Hayden (1995) suggests a 1600% increase in human population from the Natufian to the PPNB in the Levant between 10,000 and 7500BC. Bellwood (2005) describes ethnographic examples of population expansion as frequently very rapid, particularly in “frontier situations” (p.14), and particularly where there are no indigenous populations or where the existing populations were not in a position to counter the impacts of the incoming populations. He also states, based on the discussion of three modern examples of ethnographic situations, that in situations where hunter-gatherer communities are living in rich areas, even though they might well have contact with agricultural communities, there is no tendency towards the adoption of agriculture, and the hunting and gathering lifestyle persists (p.26-28). He concludes that “it is certainly not suggested here that ancient hunter-gatherers could never have adopted agriculture from outside sources. But they would only have been likely to do so in situations where they had some demographic or environmental advantage over any farmers in the vicinity and where there would have been significant reasons why the normal hunter-gatherer disinterest in agricultural adoption should be overturned” (2005, p.42) Hodder (1990) in his book “The Domestication of Europe” begins with his speculations about Catal Huyuk and describes the adoption of agriculture, in his words “as a social symbolic process. The natural (wild) is made cultured (domesticated, agri-cultured). Animals, plants, clay, death and perhaps reproduction are all ‘natural’ phenomena which are ‘cultured’ and brought within the control of a social and cultural system” (p.18-19). It is important to understand that different processes will have operated in different areas at different times, depending on numerous variables, biological, ecological and cultural. For example, in highly fertile areas where high productivity was possible, the pressure on hunter-gathering communities would have been higher than in those areas where land was not as attractive for cultivars.
7.3 Tracking Spread in the Human Biological and Linguistic Records Zvelebil (1986) has examined skeletal data from Europe and claims to be able to distinguish two different skeletal morphologies which he argues are ethnically different. Populations displaying predominantly cromanoid or Palaeo-europoid traits he interprets as “indigenous Mesolithic” (p.45). Populations with leptodolichomorphic (particularly mediterranoid) populations from the 6th and 5th millennia onwards expanded from Greece towards Danubian, Basian and along Mediterranean coasts. This coincides with the appearance of Neolithic economies: “the spread of leptodolichomorphic Neolithic populations was no doubt accelerated by the existing Climatic Optimum, which increased the productivity and reliability of farming” (p.45). On the basis of the skeletal evidence Zvelebil suggested that farming was established by colonization in south and central Europe, but that less favourable conditions in northern Europe would have led to mixed economies to ensure that the risk of new economic practices was minimized, and colonization was inhibited, encouraging acculturation rather than displacement. Zvelebil suggests that this is backed up by the fact that the palaeo-europoid morphological traits survived down to the 2nd millennium bc in northern and western Europe. However, Sokal (1991, 1992) argues that bone data from Neolithic cemeteries are too unsubstantiated to discriminate between theories of migration or indigenous adoption of agriculture. Genetics and linguistics are becoming increasingly important, especially following the work of Cavall-Sforza, Paabo, Renfrew, and others: “It must be that linguistic change and, where appropriate, genetic changes (in the molecular sense) are all aspects and consequences of concrete historical events that it should be possible to situate unequivocals in time and space” (Renfrew 1996, p.76). A number of experiments have been carried out to try and derive chronological information from modern genetic data, most predominantly by Paabo, e.g. Paabo (1993). It is unclear to what extent it is possible to use modern genetic data, because as people breed and families exchange genetic material there will be a dilution of the original source ancestral material through time and space. The question is whether this can be tracked, especially given the impacts of altering processes like invasion, epidemic, etc. There are two types of DNA used in attempts to track the spread of humans – Mitochondrial DNA which can be passed down exclusively through the mother and Y-chromosome lineages. Particularly with respect to mtDNA, molecular clocks have been devised in an attempt to track nucleotide sequence mutation rates and calculate the distance of time over which they have been mutating. Geneticist Cavalli-Sforza (1996) suggested that “In the absence of adequate skeletal and other archaeological data, a more direct indication can be derived from modern genetic data, because the spread of agricultural settlers from an area of origin into regions inhabited sparsely by hunter-gatherers, who initially had a different genetic background, is very likely to generate a genetic gradient observable even today” (1996, p.53). He analyzed both blood proteins and gene sequences and created a “genetic map” with which he tracked the spread of gene frequencies from the Middle East through Europe. He saw a connection between the spread of genetic data with the spread of agricultural adoption/replacement, and suggested that Neolithic farmers migrated with their subsistence strategies, rather than exclusively passing on the knowledge to indigenous communities – a theory that became known as the “wave of advance” model. His conclusion was that both main hypotheses – the spread of farming by physical diffusion and by diffusion of farming ideas and technologies, were correct “if they were not, the observed genetic gradient would not form” (1996, p.53). Bellwood (2005) suggests that the data are by no means conclusive: “by 2002 the supporters of both cultural and demic diffusion had fired many salvos, and the debate is far from over” p.260. He goes on (p.261): “the debate over demic vs. cultural diffusion for Neolithic Europe still remains very lively. The dust has not settled yet and there is no obvious neat and tidy genetic reconstruction of the whole of Europe that can be presented in a nutshell”. Renfrew has used linguistic data to suggest that there is a “major a correlation between the distribution of the world’s language families . . . on the one hand, and that of farming dispersals . . . on the other.” He suggests that the spread of Indo-European languages spread with Near Eastern farmers into Europe, bringing economic and cultural elements with them, at around 10,000bp, at the beginning of the Neolithic period. He believes that “farming dispersals, generally through the expansion of populations of farmers by a process of colonization or demic diffusion, are responsible for the distribution and areal extent of many of the world’s language families”. This challenged more conventional views that the first Indo European languages were probably established in Europe at around 5000bp. Bellwood believes, on the basis of linguistic studies, that Turkey may have been the Indo-European homeland (2005, p.188). Marija Gimbutas has suggested a different model, whereby Indo-European was spread not by Near Eastern colonization, but by an incursion from the Kurgan period in Russia in three phases. Renfrew contests this on a number of grounds, but significantly because there is absolutely no archaeological evidence to support it. Cavalli-Sforza has suggested that instead of an either-or model, it is possible to see two separate movements – an earlier one from the Near East and a later one by their descendents, from Russia. He believes that this would meet the challenges posed by archaeological, genetic and linguistic data (Cavalli-Sforza 1996): “if this fusion of hypotheses from Renfrew and Gimbutas is correct, the Proto-Indo-European language reconstructed by linguists on the basis of modern Indo-European languages must be much closer to that which was spoken in the Kurgan area some five or six millennia ago, than to the Pre-Proto-Indo-European spoken by Anatolian farmers” (p.66). Bellwood (2005) agrees with Renfrew’s arguments against Gimbutas because there is no Kurgan horizon visible in Europe. He believes that Proto-Indo-European (PIE) was most likely to have originated in Anatolia. Words in the vocabulary include terms for, for example, mountain, domestic animals, grain, the plough, weaving and wool: “This listing is sufficient to demonstrate the possibility that PIE could have been an early farming society – it evidently was not a hunter-gatherer one” (p.205). He quotes recent studies which suggest dates for PIE of around 7000BC or 8100+/-900BC. The linguistic and genetic models proposed by Cavalli-Sforza, Bellwood and Renfrew agree very well in many ways. However, a team from Oxford looking at mitochondrial DNA in European and Middle Eastern populations suggested a different model. Although there are some concerns about the rates of mutation in mtDNA, and therefore the reliability of it as a temporal/spatial indicator, Sykes and Richards proposed that unlike previous models, only a very few migrations occurred from the Middle East at c.10,000bp, and that what spread through Europe were not populations, but language, knowledge, skills and ideas. They suggest that migration is not necessary for any of these elements to be transferred to new populations. However, as discussed above, Bellwood (2005) points out that it is very rare for language to change unless total integration between the “owners” of the language and those who adopt it occurs. Two language groups are relevant in the discussion of North African archaeology - Afro-Asiatic (AA) and Nilo-Saharan. 7.3.2.1 Afro-Asiatic Afro-Asiatic includes:
There are two schools of thought about the origin and dispersal of North African languages. The first view is that they had a homeland in northeast Africa in pre-agricultural (but not necessarily pre-herding) societies. Support comes from the fact that five out of six of the subgroups (excluding Semitic) come from northeast Africa. However, there is no supporting archaeological data. A second view is that the languages had a southwest Asian (Levantine) origin. A number of different pieces of evidence back this view. Vocabulary reconstruction suggests that the Proto-languages lack various terms for agricultural items but include plants and animals of southwest Asian and not African origin (e.g. sheep, goat, barley and chickpea). Proto-Semitic was certainly of Levantine origin and has a full “agropastoral” vocabulary. Although a disputed technique, glottochronology, when married to other types of evidence may provide useful information: used on ancient Egyptian and Semitic it suggests that Proto Afro Asiatic (PAA) dates to between 10,000 and 7000BC, which would make it older than PIE. Archaeological data from Merimde Beni Salama and the Faiyum Depression suggests that there may have been a spread of Neolithic components into Egypt c.5500BC or before, combined with PPNB herders’ occupation of western side of the Arabian Peninsula in the early Holocene. Two routes were possible – through the Southern Levant into Egypt via the Sinai Peninsula (with the full compliment of domesticates) and through western Arabia to Africa (with only the pastoralist components including sheep and goat). As Bellwood states “it would also explain why the Egyptian Neolithic as known at present started after the PPNB, with the Pottery Neolithic” (Bellwood 2005, p.210). Renfrew also suggested that the Afro-Asiatic language family may be the result of a farming dispersal from southwest Asia: “there is general agreement that it is in this region that the goat and possibly the sheep were exploited in North Africa, as well as wheat and barley, were domestication” (Renfrew 1996, p.79). A number of writers have suggested affinities between northern Africa and southwest Asia, distinguishing the northern African languages from those in sub-Saharan Africa (e.g. Escoffier et al 1987). Renfrew suggests that “it is possible, as several scholars have argued, to see indigenous African origins for farming. But that does not account for the southwest Asian origin of sheep and goats, and then of wheat and barley” (1996, p.80). 7.3.2.1 Nilo-Saharan Nilo-Saharan is a diverse language group that extends over a very wide area, including sub-Saharan areas. Ehret (1997, 2003) suggests that Proto-Nilo-Saharan in the Middle Nile area at the end of the Pleistocene. Groups speaking the North Sudanic branch of Nilo-Saharan are considered in Ehret’s model to have adopted cattle herding and the manufacture of pottery by 9000BC. Plant cultivation and the presence of domesticated sheep and goat were present in the central Sudan by c.7000BC. After 5000BC populations spread south in keeping with increasing aridity, reaching Lake Turkana c.3000BC. Both archaeological and linguistic data agree very well. 7.3.2.2 Archaeological and Linguistic Correlations Bellwood (2005) believes that language and archaeological correlations do suggest that agricultural expansion and language advance are related and do not occur at random. It is difficult to prove but worthy of much more work. The earliest evidence for cereal cultivation on the Indian subcontinent comes from Mehrgarh in northwest Pakistan. Some writers have suggested southwestern Asian influences in the material remains, although there are obvious differences as well, both archaeologically and linguistically. 7.4 Tracking Spread in the Archaeological Record 7.4.1 Post-Glacial Environmental Opportunities The spread of agriculture from the Near East into Europe, Eastern Asia and Egypt is not completely understood in the archaeological record. However, after 14,000bp the ongoing climatic amelioration would have allowed barley and einkorn to expand out of their glacial-maximum areas. There are indications from southern and northern Greece that “barley perhaps other cereals were present in southeast Europe before the appearance of ‘village farming’ communities around 6000bc” (Dennell 1983, p.160). However, emmer wheat is not likely to have distributed itself outside the Near East at this time. Zohary (1969) notes that even though it adapts well to numerous climatic conditions, it dislikes disturbed ground, prefers limestone and basalt and, archaeologically, appears to have a very restricted distribution confined mainly to the Near East. It is known from Jarmo (late 7th millennium bc) in northern Iraq, Cayonu in southeast Turkey (8th millennium bc) and Abu Hureyra in Syria. Legumes from the end of the last glaciation include pea (Pisum) and lentils (Lens). It is by no means certain where the ancestors of domesticated forms were distributed, and identification is not always straightforward because there were two possible wild ancestors for pea and three for lentil, distributed over a fairly wide area. Difficulties of identification lead, as Hodder (1990) points out to some circular arguments: “since agriculture in southeast Europe is assumed to have originated from the Near East, Neolithic finds of Pisum and Lens are assigned to a Near East type; this can then be taken as evidence that agriculture developed in the Near East and spread outwards. As legumes have been found in early Holocene contexts at Franchthi and also in southern France, it is at least as likely that they pulses used in Neolithic Europe had a local origin” (1990, p.161). Agriculture was introduced into Cyprus during the PPNB, c.8500BC. It is not disputed that the introduction of agriculture into Cyprus was as a result of physical transference. Source sites were probably in the Levant – possibly in Syria, but those sites will have been submerged by expanding seas (Bellwood 2005, p.71). The agricultural occupation of Cyprus included both plant and animal species:
Hodder (1990) also supports this model, adding that incoming groups would not have had the local knowledge required to know where the best lands were located. At the same time that agriculture was being established in Cyprus, it was also emerging in Greece and Anatolia. However, there is a much clearer blend of Anatolian and Near Eastern features, rather than a simple replacement mechanism. 7.4.3 The Establishment of Agriculture in Europe Dennell (1983) counters the colonization model from the Near East by making a number of points:
Dennell suggests that as climate improved cereals were likely to have expanded beyond refuge areas beyond the Near East, into the Aegean basin and northern Greece prior to the appearance of farming at c.6000bc. Cereals include wild barley and einkorn, but not emmer, which had a very restricted distribution (Dennell 1983, p. 159-161). He sees similar issues regarding legumes (p.161-163). Dennell also believes that sheep may have been indigenous to Europe before the Neolithic – for example, some sheep remains in southern France date from the early Holocene. He therefore goes on to suggest that “the development of morphologically domestic barley, einkorn, legumes and sheep could have occurred in a ‘non-centre’ that extended from parts of the Mediterranean basin to southeast Europe and the Near East . . . . If that were the case, the earliest village family settlements in southeast Europe could have developed most of their features without any Near Eastern influences (p.163). But he also points to problems with Bulgarian sites predating northern Greek ones which are themselves more recent than southwestern Turkish sites. He therefore suggests that it is worth thinking in terms of a centre for dispersal but without migrating colonists. Dennell points out that cereals are assumed to have migrated out of the Near East with humans because they could not adapt on their own – having lost the features that make them adaptable to the wild, for example. But einkorn has sometimes been a weed, and if the conditions were right it might have spread – perhaps as the result of having been consumed by animals or birds and deposited elsewhere. Exchange could be another system. Obsidian was clearly being traded, for example: “what was exchanged for obsidian is unclear, but at an early date, these may have been some of the commodities that passed in the opposite direction into southwest Turkey” (p.165). He sees three phases of non-colonizing developments (p.165-167):
He explains the location of settlements in areas perfectly suited for farming as an indication of prior knowledge, which would have been available only to indigenous communities. Hodder has looked at assemblages in the early Neolithic of the second half of the 6th millennium bc, tracing the spread of agriculture into southeast Europe and beyond. He observes that “while domesticated plants and animals occur widely on sites in the late sixth and early fifth millennia, the dependence of wild resources remains relatively high. For example, wild animals often make up 35-50 per cent of the animal bones from a site” (1990, p.48). However, there is huge regional variation in this. Some sites have much lower levels of a wild animal component. Similarly some sites were probably permanent, like Koros and Cris. Some, like Koros, are a linear scatter along a riverside and are thought to represent a drift of occupation over time. Most sites are less than one hectare, and there are very few house-type structures represented: “the small social units that often depended heavily on wild resources and were often only semi-sedentary are not associated with great symbolic elaboration of the house” (p.50). There were no links at this time between houses and death, or at least very few. Exceptions include Obre 1 where burials of children were made in former storage pits. Hodder argues (1990) that the increasingly complexity of the home or “domus” was due to an increasing sense of the wilderness outside and the ability of culture and society, based on the home, to tame it. It would follow, therefore, that increased challenges would lead to greater social and structural complexity. Hodder does not extend his models outside Europe. However, while in southeast Europe there are signs that this took place in settlement contexts, it is possible that a different version is visible in the Upper Egyptian model – the contrast between the desert and the Nile, and the extremes of different lifestyle sand extreme isolation may have been expressed in the funerary and decorative, rather than the domestic. Jones et al (1996, p.95) suggest that the picture painted of the adoption of agriculture by Ian Hodder (1999) during the LBK is one of environmental domination: “an image of early European farmers not as victors in a conquest of nature, but as simply another part of the general ecological diversification of the mid-Holocene landscape”. 7.4.4 The Establishment of Agriculture in Southwest Asia For Southern Asia Bellwood concludes (2005) that: “unlike Southwest Asia, a firm and fairly unitary homeland and expansion history for agriculture cannot be established. The reality has clearly a mix of external introduction form several sources, combined with some currently rather faint hints of independent internal development” (p.87). In a discussion of the rate of the spread of agriculture into Europe and south Asia, Bellwood (2005) says that latitude was a big factor: “along a given latitude . . . there could be relative swiftness, as in the case of the northern coastline of the Mediterranean from Greece to Portugal. But across divides . . ., the rate was reduced to a heavy plod” (Bellwood 2005, p.96). 7.4.5 The Establishment in northeast Africa (including Egypt) North African areas available for occupation include zones now so arid that they cannot be occupied by either human or most animal populations. However, at various periods in the past the Saharan region was much wetter than it is today with large areas of savannah occupying areas that are now desert. In sub-Saharan Africa the most important crops are native species including sorghum, millet and African rice. These are different from species native to southwest Asia, which also appear in Egypt and north Africa, including wheat, barley and lentils. It should be noted that Barich (1998) suggests that some faunal and botanical species which are conventionally believed to be Levantine are in fact local adaptations (e.g. barley and Levantine type sheep breeds). However, this is much disputed and not widely accepted – most authorities agree that the domesticated crops and sheep/goat and pig found in Egypt were not native. Cattle may be an exception. At around 20,000bp wild resources were exploited at Wadi Kubbaniya including a rich mix of animal varieties but few plant remains which include, however, tubers, dom-palm fruit, camomile seeds, millet and Scirpus maritimus. Terminal Palaeolithic Saharan pastoralism, a mobile economy based on cattle, and accompanied by distinctive pottery, is attested to over wide areas of North Africa during the early Holocene, prior to the establishment of agriculture in northern Egypt. The earliest claims for a pastoralist economy come from the southeastern Western Desert at sites like Napta Playa and Bir Kiseiba, which have been excavated since the 1960s – variously termed but referred to by its excavators, somewhat controversially, as the “Saharan Neolithic”, or more specifically as the El-Adam phase. The suggestion is that cattle at these sites could not have survived in these areas by themselves and that human intervention must have occurred to maintain them. Cattle were native to both African and the Near East, and genetic studies are currently attempting to clarify which species were the first to be domesticated in Egypt and northern Africa. There are other examples of pastoralism based on cattle with similar pottery types, all over northern Africa. Domesticated sheep/goat only appear much later (c.5500/5000BC). Comparing this archaeological data with linguistic data, Bellwood states that “linguistic reconstructions for the Nilo-Saharan family make early Holocene cattle keeping or management of some din dint he Sahara a very viable possibility” (2005, p.106). There were no signs of full agriculture (i.e. both plant cultivation and animal domestication) in northern Africa until 5500-5000BC, when a full package of Levantine type faunal and floral type species appear in Egypt at Merimde-Beni-Salama to the north west of Cairo, and in the Faiyum Depression, to the southwest of Cairo. These sites, separated by the Epipalaeolithic by several hundred years, are remarkably isolated, and it is possible that earlier or contemporary sites may be hidden under the alluvial silts of the Nile Delta, or under the urban sprawl of Cairo itself. If the Cairo area sites are indeed the earliest north African agricultural sites, it is most probable that they were introduced via the Negev and Sinai. This would perhaps account for the late date, when compared to both pristine domestication dates in the Near East and dispersal dates to Cyprus and mainland Europe – the Negev and Sinai were highly arid areas which were only occupied by relatively economically unsophisticated groups in a dispersed settlement pattern, and may well have been a barrier to dispersal. It is unclear at the moment whether southwest Asian component represent expansion or adoption, although some linguistic data, described above, would suggest expansion rather than exclusively indigenous adoption. Archaeological data also suggests connections between the two areas:
There are also significant differences, including architecture, an apparent absence in the earliest Lower Egyptian Neolithic levels of sedentism, and a continuation of the importance of hunting and fishing in the economy. Current data suggests the sudden appearance of this fully formed agricultural economy argues for an external introduction – but this may change if earlier evidence is found under Deltaic silts. It is unclear why any groups would have wished to traverse the Sinai, complete with a full agricultural kit, to occupy areas of northern Egypt – unless the changed conditions at the end of the Levantine PPNC were part of the equation. Another possible route into northern Africa could have been across the Red Sea or the Barb al Madab, although there is no archaeological evidence of it at an early enough date (linguistics suggests that there may have been two expansions – an early one via Sinai and a later one into eastern Africa via the Red Sea). Following its establishment in the north of Egypt, it is unclear how agriculture became established over the rest of Egypt, and how it spread west along the Libyan coastline where it appeared in Cyrenaica at Haua Fteah Cave in the 6th millennium bp. Sheep/goat herding appear with pottery, but without Near Eastern cereals, in Morocco and Algeria at around 5500BC, which is the same time that they appear in the Faiyum – a fact that requires explanation, but for which data is not currently available. There appears to have been a period of increasing aridity after c.4000BC and it is possible that this forced occupants of the Western Desert to abandon areas now fully desertic, in favour of oases and the Nile Valley itself, perhaps blending pastoralist knowledge with Nilotic traditions. It is certainly the case that the earliest agricultural activity took place in Upper Egypt at this time. As mentioned above, how the Levantine traditions adopted in Lower Egypt came to form a part of this blend is unclear, particularly given the gap between dates for the establishment of fully agricultural economies in Lower and Upper Egypt of around 1500 years (which could be an artificial result of hidden or destroyed archaeological data, but could also, of course, represent the actual situation). In the more tropical environment of the Sudan and in many areas of sub-Saharan central, eastern and western Africa, Levantine cereals would have been outside their favoured habitat and could not have been grown. In sub-Saharan Africa the native species were key in economic activities, including the sorghum, African rice and millet mentioned above, as well as yam. Domestication occurred late in Africa’s history, probably in part because the cross-pollinating character of native cereals makes them difficult for the isolation required for specific traits to be bred for preference. Bellwood (2005) points out that by 1500BC the economy of the southern half of Africa was still almost completely hunter-gatherer. The dispersal of farming was clearly not a straightforward process - “The transition to farming must be seen as a complex process resulting from several forces operating simultaneously at different chronological and spatial scales of resolution” (Zvelebil 1986b, p.167). This is reflected in all the literature, not just archaeological but linguistic and genetic as well. Bellwood’s survey of agricultural dispersal from the Near East concludes that “these early spreads of farming were clearly inexorable, even unstoppable in the long term, but they were not instantaneous . . . . However, the farming spread, it was neither by quick-smart adoption by clever hunter-gatherers, nor was it by express trains carrying hungry farmers ever toward the horizon, leaving astonished and landless hunters in their wake” (Bellwood 2005, p.96). However, most discussions come back to the two main models – indigenous adoption or migration. In favour of colonizers, who expanded into Europe and mixed with existing populations are Bellwood, Cavailli-Sforza and Renfrew. Bellwood states that indications do point to a population spread from the southeast to the northwest, with genetic data diluting steadily across this space, although the timing is disputed. Demic diffusion is the favoured model. Those who see an emphasis on adoption by indigenous populations include Dennell and Hodder. As Dennell summarizes it: “the ‘spread’ of farming after 6000bp over most of temperate Europe is best explained by the way that Mesolithic societies assimilated new resources and technologies, rather than by the expansion of farming pioneer” (Dennell 1983, p.176).
|
|||||
|
Copyright (text and images) Andie Byrnes 2005, unless otherwise stated
|
|||||