4. The gap in the archaeological record: retreat or resistance? 

The Neolithic societies all over the Sahara underwent important changes towards “complexity”, as it has been often called by archaeologists and historians, but it would rather be described as “inequality” (Paynter 1989; McGuire and Paynter 1991; Price and Feinman 1995). As it is argued from anthropological data, the tending of animals, particularly cattle involves such an assiduous commitment that there is an ubiquitous tendency for the herds to be the private property of extended families inside the clans. This involves a major transformation in the social relations of production and the ideology of prestige (Ingold 1980). Archaeologically, the appearance of independent animal enclosures in the first farming societies has been interpreted as evidence of some kind of private property. In the Egyptian sites of Merimda and El Omari this feature goes together with the separation of grain silos in the huts, as opposed to the communal silos recorded at the earlier site of Fayum (Hassan 1988: 154-5; Midant-Reynes 2000: 116). In the Saharan area, Bovidian rock art scenes do change with respect to the previous Round Head style, with an under-representation of the female figure in the drawings that conforms to the emergence of a pastoral ideology that attributes a greater meaning to the masculine figure (Barich 1998; Gifford-Gonzalez 1998).

Even though no evidence of that kind –individual silos or enclosures– has been detected in the Khartoum Neolithic, several indications point to the inception of social inequality taking place also in this area. At the Neolithic cemetery of Kadero, both location and furniture quantity of the graves were used to differentiate the dead (Krzyzaniak 1991: table 1, figs. 2-3), most probably according to their social status (Binford 1971: table 4). The importance of men at the site is manifest, with 20 out of 28 adult graves whose sex could be ascertained, and six out of the eight richest tombs belonging to male individuals (Krzyzaniak 1991: table 2). Important socio-technic artefacts, such as porphyry maceheads were found in all cases associated with adult men graves (Ibid.: 523). The social elite was also ascribed grave furnishings of marine shells and malachite/amazonite objects traded from abroad (Ibid.: 531). The very fact that the two richest grave categories include children could even indicate that prestige was not merely acquired during life but inherited, just as it has been postulated an idiosyncratic feature of hierarchical chiefdom societies (Peebles and Kus 1977). Proof that this process was constantly proceeding in the Central Sudan comes from the cemeteries of Kadada, where only a few centuries later (Kadero is dated to 5900-5000 bp; Kadada to 4800-4600 bp) the graves show an extraordinary array of differences in furniture richness, including human sacrificial secondary burials (Reinold 2000: 70-1).

The site of Sheikh el Amin may be of relevance to this question. It was excavated rather intensively on 140 square meters, yet not any sign of Neolithic human burials was detected. Sherds from pottery vessels considered as prestige emblems and exclusively associated with graves, such as the flare-mouthed finely decorated beakers (Reinold 2002: fig. 4) weren’t found either. Even rhyolite gouges, whose socio-technic character has been occasionally denoted (Haaland 1987: 221) are infrequent in the site, where only 26 pieces were found. In the riverine Neolithic settlement of Shaheinab, 467 pieces were counted excluding those broken to less than half the original size (Arkell 1953: 31). This lack of evidence suggests a more egalitarian organization of the group than those living at the riverine sites, maybe connected to its economic adaptation, based on hunting-gathering with a small herding component (13.1% of identified bones, cf. Chaix 2003: table 12). Several authors have discussed the difficulties that modern hunters have in making the transition to food production, particularly to integrate herding (Smith 1990; Marshall 2000: 215), on the grounds of current evidence of hunting people living on the edge of pastoral societies (Smith 1998: 26), and the persistence of hunting as a component of generalized pastoralism (Marshall and Hildebrand 2002: 121).

Sheikh el Amin probably represents a short-lived kind of adaptation to the savanna ecosystem, based on multi-resource food procurement in small semi-sedentary villages. Shaqadud cave, a permanent post of hunter-gatherers without livestock until a very late date, could have corresponded to the same adaptation of groups far from the Nile at about the same period (Marks and Mohammed-Ali 1991).

Further cultural and economic changes are visible in the archaeological record several centuries later. Late Neolithic sites investigated during our survey, such as Rabob and Wad el Amin dated to after 4500 bp, present somewhat different features. Surface distribution of artefacts follows a model of “sheet midden” that has been interpreted as the result of seasonally reoccupation of the settlement, with people erecting their tents or temporary huts and choosing the waste zones in different places every year (Sadr 1991: 21-3). The site layout at Sheikh el Amin is different, with cluster midden mounds surrounded by empty spaces –cleared habitation zones– that could correspond to a permanent or at least a “medium” term occupation where people lived for a long time in the same structures (Ibid.). Thus a change to a more mobile economic system is noticeable at the end of the Neolithic period, which could correspond to the inception of a more intensive pastoral economy in the region as it has been argued by several authors (Krzyzaniak 1978; Haaland 1987; Caneva 1988b).

Ceramic analysis and seriation of the surveyed Neolithic sites demonstrate that there is a fairly substantial continuity both in the pottery manufacturing techniques and decoration types throughout the period (Fernández et al. 2003a: fig. 56). Similar models of archaeological type variation have been proposed elsewhere as evidence of uninterrupted cultural and demographic permanence (for the Epipalaeolithic in the Maghreb, see Lubell et al. 1984: fig. 3.4). Thus it seems that, contrary to the aforementioned evidence for the beginning of the period, there is not proof of cultural and/or demographic gap during the Sudanese Neolithic as it has been postulated by Haaland (1987, 1992). It does appear more probable that the same Butana Early Neolithic groups from Sheikh el Amin and other sites as El Lahamda (Fernández et al. 2003a: section 7), progressively abandoned their sedentary economy adopting a more mobile herding economy. The resultant enlargement of their annual territories induced more frequent contacts with groups of the Eastern Sudan, as it is reflected in the increase of simple impression and scraping decorations on their pots all through the period.

The progressive intensification of arid conditions in the area (Wickens 1982: 44-7; Hassan 2002: 323) was most likely one of the main causes of the higher mobility of the savanna groups. But the riverine areas were still very apt for human sedentary living and yet they appear almost completely devoid of archaeological remains from this period. Later on, a time comes from when there are not known archaeological sites both in the savanna and the river, with the possible exception of some burial mounds (Caneva 2002). The period roughly corresponds to the interval between 4000 and 2500 bp (c. 2500- 700 BC calibrated) the latest date marking the beginning of the Napatan-Meroitic periods in the first millennium BC. This “disappearance” of Late Neolithic cultures in Central Sudan has been related to the demise of the Nubian A-Group, both of them probably induced by changes of power balance in Egypt and Northern Sudan, namely the emergence of the Egyptian state and the Kingdom of Kush at Kerma (Caneva 1988b: 371). An external origin has been also alleged by Haaland (1987: 224-231, 1992: 58-61) who presents an interesting scenario of competition between Nilo-Saharan speaking groups with multi-resource adaptation (Khartoum Neolithic) and Cushitic-speaking specialized pastoralists (Butana-Khashm el Girba tradition). The first would have migrated towards the south, preserving their way of life in more humid ecosystems while the Khartoum region was occupied by the second groups coming from Eastern Sudan. The recent finding in the Khartoum region of tumuli, dated to 3220 bp and with some cross-hatched pottery similar to that of the Nubian Pan-Grave culture in Northern Sudan and the Mokram group in Eastern Sudan (Caneva 2002), ancestors of the present-day Beja Cushitic-speakers (Sadr 1990), appears to be a further support for the hypothesis of culture contact in Central Sudan.

The results of our survey do not endorse, however, such conclusions. Pottery from the Late Neolithic sites bears only some similarities to the eastern wares, and the substantially gradual variation of decoration types during Early and Late Neolithic times seems to continue without important breaks even until the richly decorated hand-made wares of the Meroitic period (Fernandez et al. 2003a: section 7). In addition, the language that was written in the Meroitic script at the end of the I millennium BC and the first centuries AD, probably spoken in much of Central Sudan, was not Cushitic but belonged to or at least had some relations with the Nilo-Saharan phylum (Bender 2000: 56).

Archaeological investigations of Sudanese Neolithic graveyards, such as Kadruka and Kadada reveal that a nearly “pre-dynastic” stage was achieved both in Nubia and the central Sudan at roughly the same time, the second half of the 6th and the first half of the 5th millennium bp, i.e. at the end of the 5th and during the 4th millennium BC in calibrated dates (Reinold 2000: 58-85). Yet the transition to state organization was completed only in the northern region, with the onset of the Kerma kingdom. Significantly, the Nubian Nile Valley, like the Egyptian, meets one important specific pre-condition which Robert Carneiro (1970) argued in his well-known theory on the state origins, namely the “environmental circumscription”. This circumstance is met when a growing population lives in a confined area, delimited by mountains, jungles, deserts or seas (Claessen and Skalník 1978: 13). Extreme deserts did not restrain the Central Sudanese groups and the availability of close suitable and nearly uninhabited lands made it possible for them to avoid the transit to an unequal social organization.

It is known that hunting and simpler farming communities ubiquitously follow customary strategies to keep to social equality (Clastres 1978). One of them could have been in this case the shift to a more mobile economy, approaching specialized pastoralism. Instead of adopting agro-pastoral strategies in the alluvial plains as the Egyptian and Nubian communities did, groups in the Sahelian Nile chose to flee from the river and wander the increasingly arid savannah. (Could this circumstance be a main reason for the time lag of agricultural practices in the region?) African nomadic communities have often been considered as intrinsically egalitarian polities, being frequently without centralized authority and endorsed with a “democratic” ideology (Bonte and Galaty 1991: 23-4). The material basis of this condition lies on the fact that livestock wealth, self-reproductive and mobile, cannot be easily monopolized (Ibid.) and that it is difficult for pedestrian herders to exploit the labour of others on a mass scale (Goldschmidt 1979: 23). One of the best-known historical examples of “pastoral democracy”, geographically close to our area, is the Oromo people of southern Ethiopia and northern Kenya (Legesse 2001). Although originally egalitarian, Oromo became progressively autocratic in modern times (Hultin 1979), expanding through the conquest of a large extension of southern Ethiopia and incorporating quite a number of simpler societies to their kingdoms (Hassen 1994).

Some of those egalitarian societies belonged to a larger group of Nilo-Saharan societies in the Ethio-Sudanese borderlands that have been mentioned above (Grottanelli 1948; Cerulli 1959; Murdock 1959: 170-80). Historically, peoples in this area have resorted to their cultural traditions to avoid subordination and cultural assimilation, as an example of “cultures of resistance” or “deep rurals” (Jedrej 1995: 3). Emphasis on traditional material culture and rejection of innovations has been documented in other known cases of conflation of identity and resistance (e.g. Levi 1998). Though in a different context, the European Megalithic cultures, construction of big tombs has also been interpreted as a means of resistance against social division (Criado 1989: 91-2). As a collective endeavour strongly related to the symbolic domain, the burial mounds in Central Sudan could have also operated as a peculiar version of the potlatch ritual, i.e. a copious consumption of co-operative workforce for the benefit of the whole group.

A state organization became eventually established in Central Sudan at the time of the Napatan and Meroitic kingdoms. The same kind of social system prevailed throughout the periods of the Christian Kingdom of Alwa in the Middle Ages and the Funj Muslim Sultanate at Sennar up to the Egyptian conquest in 1821 AD. But these polities were mainly riverine systems, and the results of previous archaeological surveys in the Butana (Hintze 1959) and our own data reveal very few settlements, the main archaeological sites from those periods being again burial mound grounds (Fernández et al. 2003a: section 8).

Some linguistic and historical data on the region attest the existence of population movements and contacts across the Butana plain and the Blue Nile river, connecting the Central Sudan and the Ethiopian escarpment. First it is the ancient separation of Kunama languages, a dialect cluster today spoken in south-west Eritrea, and of the Koman languages, spoken in the central Ethio-Sudan border, from the old proto-Northern Sudanic and proto-Nilo Saharan language groups (Ehret 2000: 273-7). These may be interpreted as the first historical splits of Nilo-Saharan peoples from the main stock situated in the Saharan and Sudanese plains. Then we have the similarities observed between Meroitic and Barya (Nera), another Nilo-Saharan Eritrean language, possibly by the influence of the state level language over the people living in its frontiers (Trigger 1964; Bender 1981, 2000: 56). The Meroitic has been also related to the Koman languages (Shinnie 1967: 132, n. 7). Information from Arab travellers in the Middle Ages suggests that Kunama and Barya peoples were at that time installed nearer the core of the Christian kingdom of Alwa, from which a later displacement to their current position in the Highlands is deduced (Murdock 1959: 170; Pankhurst 1977: 3). Oral history from the Berta people, now living at both sides of the central border, indicates that they also moved to the Highlands from the southern part of Sennar kingdom in recent times (Triulzi 1981: 21-5).

All those frontier groups have preserved hunting-gathering practices until very recently, and though some of them tend cattle, those living in the forested escarpment are mostly hoe farmers (Cerulli 1956: 179). The current lack or insignificance of livestock, anyway, does not imply the same condition in the past, since the groups could have lost it owing to the pernicious effects on cattle of trypanosomiasis of forest barriers when arriving from the plains to the Highlands (Gifford-Gonzalez 2000: 119-23). Social organization of all these communities has been basically egalitarian until quite recent times (Grottanelli 1948: 311-5; Murdock 1959: 176).

What we are presented in all this information may perhaps be considered part of the historical processes of longue durée at the Eastern Sahelian region. The closeness of a forested and rugged mountainous region may have been a powerful attraction as a refuge area to small independent groups that combined an egalitarian social system and their general linguistic affiliation with the Nilo-Saharan language phylum. Their particular languages belonging up to six families of the phylum (Bender 2000: 44-6), suggests that these peoples reached the “sanctuary” in different historical periods. Even in a continent that has been characterized historically by mostly egalitarian economic and political systems (Coquery-Vidrovitch 1969; McIntosh 1999), or just because of this very condition, many traditional groups have resisted the inception of “complexity” and subordination with all their means, actually resorting to retreat at that intricate hiding band in the Highlands corner that Evans-Pritchard (1940) called the “corridor of death” (Johnson 1986: 219).

Later on, except for occasional slave raids (Pankhurst 1977), these Shankilla (black, slave) populations lived for centuries in an acceptably independent situation at the edge of the Sudanese and Ethiopian kingdoms, as historical data from foreign travellers to the Highlands suggest (Páez, Prutky, Bruce, etc.). Their inferior position in modern times (e.g. Donham 1986: 12) could be more a consequence of Abysissian expansion in the 19th century than the result of earlier enslaving practices. The Sudanese refugees newly settled at the Ethiopian side of the border due to the civil war attest the persistence of the process, which started in prehistoric times, up to the present day. In such cases as the T’wampa (Uduk), the whole ethnic group, some 20.000 people, has been resettled in the old refuge areas (James 1994), which are now called the Tsore, Bonga or Sherkole camps.