Most archaeologists would agree that craft specialization is an integral component of increasing socioeconomic complexity. The occurrence of specialized craft production has served as a perennial favorite for discussions involving social stratification, economic exchange, and ultimately state-level organization (e.g. Brumfiel and Earle 1987; Childe 1950; Wright 1986:323–324). In fact, craft specialization could be characterized as the workhorse of archaeological investigations into complex society. Specialized production, as exemplified in an almost bewildering variety of material manifestations, has been used to monitor administrative influence over production (Feinman, Kowalewski, and Blanton 1984; Spence 1986), the development of interand intra-regional trading networks (Rathje 1975; Wright and Johnson 1975), and various characteristics of the producer/consumer relationship, including the degree of producer competition (Feinman et al. 1981) and the existence of elite consumers (Rice 1981:223).

Given its central role in interpreting social complexity, it is not unreasonable to also consider how craft specialization is viewed archaeologically. What methods have been advocated to identify craft production? What interpretive models have been advanced to meld these archaeological data into an understanding of the past?

The present work employs ceramic production data to address these questions. Ceramics were chosen for several reasons: (a) their positive correlation with sedentism and complex society (e.g. Rice 1987:190; Skibo et al. 1989:126); (b) their almost exasperating quantity as archaeological data (e.g. Willey 1961:230; Sullivan 1988: 23); and (c) their potential as vehicles for interpreting past production organization (e.g. Arnold 1985; Kramer 1985; Rice 1987).

In the nineteenth century, prehistorians in the Old World were already much concerned with issues of cultural chronology. Inspired by the success of paleontologists and geologists in their use of the “index fossil” concept (Daniel 1964: 33–4), prehistorians began to make use of certain stone tool types and pottery types to indicate the age of deposits in the same way (see Childe 1956: 59; Rouse 1972: 126; Brown 1982: 181). The use of type concepts therefore has a long history in the Old World (cf. Gorodzov 1933; Clark 1957: 134–8; Chang 1967: 8; Klejn 1982: 38–50). In the New World, however, it was generally supposed that the native inhabitants were fairly recent immigrants from Asia, and hence that there was no significant time depth in American culture history. Consequently, there was more interest in the spatial than in the temporal variations of culture, and for this kind of ordering type concepts were not particularly important. N. C. Nelson in the Galisteo Basin was probably the first archaeologist in North America to make systematic use of type concepts for dating purposes (Nelson 1916), and the practice did not become general until a decade later.

Once they had become aware of the possibilities, however, Americans soon made up for the deficiency in their earlier use of type concepts. The late 1920s and the 1930s witnessed an enormous proliferation of arrowhead, pottery, and other typologies; indeed, more than half of the typologies in use today were probably formulated originally during that period.

Matson's definition of ceramic ecology included an emphasis on “the functions in [the potter's] culture of the products [the potter] fashions” (1965:203). The suite of possible product functions will not be considered here. The ceramics produced in the Tuxtlas are primarily used for cooking and serving; this discussion presents information on the use and use-lives of this utilitarian pottery.

This section begins by considering variability in the frequency and proportion of ceramic vessels within the sample of fifty households. Differences in these measures are the result of two regulating forces. Assemblage diversity, or the occurrence of different pottery forms, is strongly associated with the techniques of food preparation practiced in the household (also Nelson 1985; Rice 1984a:245–246). Within households whose corn is ground mechanically, the containers used to boil corn are more likely to be metal. Those households processing corn by hand (using manos and metates) are more likely to have specific pottery vessels used in corn preparation.

A second factor conditioning assemblage size is access to replacement vessels. Access comprises several variables; among them are the number of potters within the community making the vessel type, the distance to the market, and the price of the vessel. Access to ceramics is used to investigate the interaction between household populations and assemblage size. The ethnoarchaeological literature suggests a poor correlation between the number of persons within the household and the frequency of pots within the assemblage (e.g. Kramer 1985:91–92).

The convergence of three long-standing interests – a fascination with Anatolia, the phenomenon of nomadism and commitment to the progress of analytical archaeology – have come together in producing this book. Scattered observations during travels in Iran and Turkey during the mid-1970s had convinced me that pastoral campsites provided an ideal laboratory in which to pursue certain strands of emerging middle range theory, particularly with respect to occupation floor models and site structure.

What might have been a detailed empirical study, carried out under controlled conditions in carefully selected sites, changed direction in the field. Confronted with the realities of ethnoarchaeological research in the modern Near East, I was not only forced to be more eclectic in terms of the campsites studied, but was also encouraged to diversify, taking an interest in Seljuk history and local history, ruined caravanserais and the accounts of nineteenth-century travellers, together with neo-Marxist anthropology and quantitative techniques of spatial analysis. While I had planned for a more in-depth coverage of a single region or group of nomads, there were compensations in the form of greater geographical range and diversity in the nomadic and semi-nomadic groups visited.

The beginning of my ethnoarchaeological fieldwork coincided to the day with the October 12 military coup in 1980 which put certain constraints on my activities. Yet compared with other countries in the region, conditions in Turkey were quite favourable.

Nomad camps have been likened to ‘mobile villages’ (Barth 1961, p.25; Stauffer 1965, p.285). The analogy holds up to a point. Camps, like villages, are composed of individual dwellings, each containing a household unit. Tents are generally clustered together for reasons of sociability and security. The residents of both settlement types form a face-to-face community. However as our attention shifts to residential patterns, community organization, property relations and territory, the analogy begins to break down. Failing major catastrophe, there is a certain inertia in the physical layout and membership of a village, which is absent in a nomad camp. Other things being equal, the villager will tend to remain where he is and any change in residence will involve a major decision. For the nomad such decisions are routine and the composition of a nomad camp can persist even over short timespans only through constant reaffirmation (Barth 1961, p.25). The size and composition of nomad residence groups is therefore highly responsive to decisions taken at the household level.

Nomad social organization

Central to this residential flexibility is the nomad facility for reinforcing social relationships through physical distance.

This study has been built around the proposition that nomadic sites constitute an analytically distinct settlement category requiring different methods of survey, recording and excavation from those commonly employed by archaeological fieldworkers in the Near East. Having arrived at an understanding not only of the structure, location and contents of nomad campsites, but also the kinds of methodological and theoretical problems that are best pursued in relation to this kind of site, let us now reconsider what archaeological evidence is available.

Tepe Tula'i

It is significant that the only site securely identified as a nomadic – or at least pastoral – campsite was detected not through intentional survey but in the course of rescue operations on a small Sassanian site prior to an agricultural development scheme in Khuzistan (Hole 1974–75, 1975). Tepe Tula'i lies near the northern edge of the Khuzistan Plain about 4 kilometres east of the Karkheh River and 17 kilometres west of Dezful, placing it on a possible seasonal migration route from Khuzistan to summer pastures around Khorammabad in the Zagros Mountains.

The site first appeared as a low mound dissected by old irrigation canals, the surface of which was covered by loose, rocky fill containing Sassanian sherds. These were partially mixed with prehistoric sherds from a layer of hard-packed fill, disturbed by burrows, which extended well beyond the confines of the mound itself.

Along Turkey's southern Mediterranean coast, in the area known to classicists as Rough Cilicia (Fig. 7.1), small numbers of nomadic pastoralists continue to set up camp for the winter. During two successive autumns – 1980 and 1981–I was able to observe the comings and goings of a single household and carry out a detailed analysis of their abandoned 1980–81 campsite.

The region

The small village of Karakegi, which served as a base for much of my fieldwork, lies strewn among citrus groves, overshadowed by the battlements of a Crusader castle and the ruins of what must have been a sizeable city in Hellenistic, Roman and Byzantine times. Limestone strata rising abruptly from the sea climb tier upon tier towards the Taurus Mountains in a karst landscape heavily dissected by seasonal watercourses, faulted by massive sink holes up to 200 metres deep and drained by subterranean systems of swallow holes. Unable to retain surface water, this stony landscape has little agricultural potential except where small patches of alluvium or swampland occur at the mouths of streams. Straggly stands of winter wheat and heavily irrigated citrus orchards clothe the lower slopes. Precipitation occurs between December and May along with moderate winter temperatures, and snowfalls are almost unknown, while summer temperatures hover around 35°C.

In the Hellenistic period the region lay at the boundary of the kingdom of Olba (see Mackay 1968) whose tall stone-built watchtowers can still be seen in the hills overlooking the sea, spaced at roughly 2 kilometre intervals.

As I approached Ali's winter camp all illusions about the simplicity of my task began to crumble. The tent – if it could be called that – was fashioned out of a mixture of rough limestone boulders and dressed blocks from a nearby Byzantine complex (Fig. 6.1). Its walls were nearly 1.5 metres high, and it was entered via a short passageway and a rickety wooden gate held by lengths of twisted wire. The structure was surmounted by wooden beams and a mixture of tentcloth; straw and polythene plastic. The interior of the dwelling contained the now-familiar array of rug-covered platforms though its proportions were less elongated than the norm. In the front wall, adjacent to the doorway, was a recessed fireplace beautifully constructed out of Byzantine building rubble and capped by a large limestone block covered by floral bas-reliefs. In summer Ali would be up in the Taurus Mountains living in a black tent with only the barest of stone footings. The following year at about the same time (early September) I found him camped some 200 metres away in a small beige-coloured tent whose interior was almost identical to that observed in the previous winter, but which was devoid of any foundation other than a cleared earth floor and a small hearth consisting of three stones.

The term nomad architecture may appear to contain an inherent contradiction. In fact there is a sizeable literature on the subject (Fagre 1979; Andrews 1981; Ferdinand 1959–60; Feilberg 1944; Edelberg 1966–67.)

We have seen that the low archaeological profile of nomad sites cannot be attributed to the absence of material remains alone. But while nomad material culture and architecture may be comparable in variety and durability to that of more settled communities, the distribution of these remains occurs over a much wider range and in greatly reduced densities. Moreover the manner in which fluid patterns of residence group formation and fortuitous shifts in migration tracks are translated into archaeological assemblages will vary according to time and place. It is possible nevertheless to make some broad generalizations about what sets a nomad settlement system apart from a sedentary one and how nomad campsites differ fundamentally in their internal structure from sedentary communities.

Nomad settlement systems

The term ‘settlement system’ has been carefully chosen as against ‘settlement pattern’ in order to distinguish a system of controlling variables (Flannery 1976b, pp.162–3) from a physical pattern of settlement (Chang (ed.) 1968; Trigger 1968). The latter approach applied to nomad camps would yield a picture of segmented settlement units located in relation to each other according to patterns of kinship, tribal affiliation and economic cooperation.

From the modern highway crossing the Taurus Mountains, Sanaydin Yayla appears as a collection of small beehive-shaped tents in a narrow hanging valley. Closer investigation will reveal stone and thornbush corrals scattered amongst the tents, a number of mud-roofed huts and well-beaten paths (see Fig. 9.1). What at first appears to be a stone-built, roofed corral roughly in the centre of the camp soon reveals itself as a ruined Seljuk caravanserai (Erdmann 1961; Özergin 1965), and out of the stone rubble lining the banks of a seasonal watercourse the practised eye will discern the remains of a paved road and arched stone bridge. The paved track winding its way from the valley floor to the Taurus watershed, at around 1,600 metres, is part of an ancient trade route running from Silifki on the coast to Karaman, Konya and ultimately Ankara on the Anatolian Plateau, mentioned in sources of the early Seljuk Dynasty in the thirteenth century (Turan 1971, pp.346, 512).

The setting

Having climbed the 2,000 metres or so from the shores of Lake Van to the rim of the crater of Nemrut Daḡ, the traveller is treated to an awesome sight. Below stretches a huge basin surrounded on all sides by precipitous walls. The interior is a tumbled chaos of conical hills, lava flows, depressions and jagged outcroppings of rhyolitic rock and obsidian, its western half drowned by the icy waters of a large semicircular lake. Nemrut Daḡ is a collapsed caldera, the remnant of a lofty volcano which emerged from the floor of the plateau to dam up the waters of Lake Van to the east. It is one of a chain of volcanic eminences lying to the north of the folded ridges of the Taurus Mountains – Ararat and Süphan to the east, Bingol, Erciyas and Hasan Daḡ further west.

The future of nomad archaeology in the Near East remains uncertain. With large regions closed to archaeological research through war and political instability or heavily restricted, opportunities for the kind of surveys and excavations necessary to uncover nomad sites on a large scale may not be forthcoming. But neither is this desirable. After all nomad sites should form but one component of a wider regional perspective along with many other kinds of sites. The principles outlined above should provide guidelines in designing representative surveys, recognizing the distinctive features and organization of pastoral settlements and detecting the presence of nomadic or composite dwelling forms.

Theoretical perspectives

This study, I hope, has gone some way towards providing, at least for a certain class of sites, what Binford has referred to as:

Following recent advances in intra-site spatial analysis to decipher the meaning of item distributions (Whallon 1984; Hodder 1976; Orton 1982), efforts are now being made to tackle the problem of relating such patterns to features. In Chapter 7 of In Pursuit of the Past Binford explores this interaction between site ‘frameworks’ and item patterning, but his observations remain tied to particular activity spaces such as hearths, specialist work areas, etc.

The defiant reply of the Scythian chieftain Idanthyrsus to Darius, King of Persia, contains a number of points of interest to any student of nomadic pastoralism as an anthropological and historical problem. The ceaseless cycle of mobility, the lack of fixed assets and the military advantages which flow from both – all are common themes in the anthropology of Near Eastern nomadism. Historians receive the occasional glimpse of nomadic cultures and life ways, such as the above.

While the quest for the origins of nomadic pastoralism is not the central theme of this work, the issue must be raised if only because it has formed the focus of so much archaeological work. Yet the origins of nomadism may not be of quite the same significance as other origins – such as the origins of food production and of urban life. In the latter instances we are dealing with major systems transformations which require extremely complex explanations and models and which are part of a general trend towards organizational complexity and increased energy capture. The advent of nomadic pastoralism has its place in all this but if; as suggested earlier, it is peripheral to the main line of cultural evolution, then it may not require the same level of explanation. Perhaps nomadic pastoralism cannot be said to have had a precise origin, or perhaps it has had many origins in many places and at different times. If, as argued by some (Khazanov 1984; Gilbert 1983; and also in this work), nomadism is an unstable and fluctuating phenomenon involving a heavy commitment to pastoralism, then we would be far better off studying those sets of conditions under which it emerges or declines than in searching for origins or building causal models.