Did specialized nomadism play a pivotal role in the evolution of steppe confederacies during the late Iron age in southeastern Kazakhstan?

A case study from the Talgar region of southeastern Kazakhstan.

Introduction

For most scholars of the Eurasian steppe societies, the origins of horseback riding was the key innovation that led to the evolutionary transformation of steppe society from mixed herding and farming communities to specialized nomadic confederacies and states (Khazanov 1984: 94; Anthony, Telegin, and Brown 1991). The origins of horseback riding now can be traced back as far as the Copper Age of the Sredni Stog culture (4300 BC to 3500 BC) at Dereivka in the Ukraine (Anthony, Telegin, and Brown 1991). Once horseback riding was invented, its diffusion throughout the Eurasian steppe was rapid and transformative. As a result of the introduction of this innovative technology, changes within the Eurasian steppe communities occurred -- these changes included increased trade and communication, the necessity for agricultural communities to defend themselves against outsiders, and the penetration of these horseback riding cultures into the deep steppe in areas previously unexploited (Anthony, Telegin, and Brown 1991:100). The rapid development of the nomadic confederacies of the Iron Age (first millennium BC) known as the Saka/Scythian cultures, however, took place three millennia after the earliest evidence for horseback riding. This suggests that steppe cultures underwent a slow and stable rate of change from the Copper Age through early Iron Age. Horseback riding was only one of the many factors leading to the rapid development of nomadic confederacies during the first millennium B.C.

We hypothesize that one fundamental factor led to the rapid development of nomadic confederacies, states, and empires -- the necessity for tribal groups to maintain control over agricultural as well as pastoral production. This assertion is hardly new or even revolutionary since both Soviet and Western scholars frequently refer to the existence of agricultural practices on the steppe. In the desert oases of Central Asia both rainfall and irrigation agriculture were important during the Bronze Age, in the Semirechye (7 rivers area of the Southeastern Eurasian mountain-steppe region) large-scale irrigation canals have been found at medieval sites (ca. 7th to 14th century AD) along the Silk Route (Hiebert 1993; Baipaikov 1990). The interdigitation between cereal farming systems and animal husbandry in Southeastern Kazakhstan was the crucial component of Iron Age economies in the Talgar region. Animal herds could be grazed in areas away from cereal crops. After the harvesting of such crops these same herds could feed off the stubble or fallow fields while manuring the fields and providing natural fertilizer for the next year's crop. We believe that the dual herding and farming economy combined with other food getting strategies such as hunting and foraging made up the economic base of Iron Age settlements. The social inequality and hierarchy within nomadic confederations became more marked when this multiple resource base needed to be defended or protected from outside groups.

By incorporating the preliminary data collected from our own survey reconnaissance and test excavations in the Talgar Region of southeastern Kazakhstan, we hope to shift attention away from "nomadism" toward a clearer understanding of the dual economies of herding and farming on the steppe. We propose the following evolutionary scenario. Along the fringes of the steppe in southeastern Kazakhstan, a dual herding and farming economy was the foundation for the development of complex, stratified societies during the second half of the Iron age. Specialized nomadism was a strategy used for the militaristic purposes of subjugating other farming and herding communities, and defending one's own economic resources. Both long-distance and vertical transhumance took place on the southern borders of the Eurasian steppe, but these forms of pastoral transhumance were "ecological-economic adaptations" within a fully integrated herding and farming economy and not the "specializations" that led to the evolution of nomadic confederacies. Our scenario is borrowed from Herodotus' claim (Book IV) that the Scythians had three estates: (1) the military elite; (2) the shamans and soothsayers; and (3) the herders and farmers. In other words, the nomadic confederacies in Semirechye arose from the stratified steppe communities that depended upon a multiple resource base of herding, farming, hunting, and foraging. During the Iron age the inequality among occupational groups of people and a warrior elite was most likely intensified by the competition for rich resources found in the circumscribed pockets along the fringes of the Eurasian steppe. The "reputed nomads" of southeastern Kazakhstan were not the poor herdsmen grazing sheep, goats, cattle, horses, and the occasional camel on the steppe, but the elite who used weaponry, horses, and luxury items as the political and ideological means for defending their own valuable resource base and for conquering other communities.

The goals of this essay are to: (1) use preliminary field data collected from survey reconnaissance of the late Iron age through medieval period sites in the Talgar area for constructing a settlement model for how nomadic confederacies evolved; (2) show how environmental zones were used for herding, foraging, and agriculture; and (3) provide future directions for understanding the rise of Eurasian steppe confederacies, states, and empires using archaeological data.

The Study Area

The Semirechye area, the geographical corridor that extends from the Central Asian oases to the northern edge where the mountains and foothills border the vast southeastern border of the Eurasian steppe, is a natural passageway between China, Mongolia, Middle Asia, the Near East, and India. The Talgar Region can be described as a circumscribed alluvial apron which extends northwards from the glacier peaks of the Zaiilisky Alatau, an extension of the northern flanks of the Tien Shan Mountains of southeastern Kazakhstan. The Talgar River forms the source of the broad alluvial apron as it flows north from the foothills of the Zaiilisky Alatau. Our study area has been defined as starting at the foothill region just above the alluvial apron and extends north about 15 km to the semi-arid desert steppe where the soil particles become smaller in size and more clayey in composition.

The study area can be defined as consisting of these geomorphological units: the foothills, the upland alluvial valleys, alluvial terraces, mud slides, and gravel deposits. All along the northern Tien Shan where the foothills meet the steppe, there is a steep environmental gradation and a relatively short distance between mountains and arid steppe. For example, Zone 1 of our study area, the conifer forests and mountain found at at 1300 m in elevation is only 15 km from Zone 5, the desert steppe at 675 m in elevation. In terms of human use of the Talgar Region this is a crucial variable: prehistoric people could travel short distances while exploiting a number of different environmental and vegetational zones. Ancient people could live with a circumscribed area 15 to 20 km region while participating in such diverse economic activities as animal herding, hunting of birds and animals, fishing, and cereal and fodder farming. Furthermore, the vegetation in the Talgar Region is particularly lush because of the Talgar River and its many tributaries that flows north from the foothills and high mountains onto the steppe. In addition, the rainshadow effect of the Tien Shan Mountains often causes frequent summer rains in Talgar, thus making rainfall agriculture possible, even in the semi-arid desert steppe zone.

During 1994 a vegetation survey and plant inventory was made for the Talgar study area outlining five different environmental zones (Figure 1: 1:100,000 scale map showing the five major zones). Our initial intention was to use this environmental base-line to design a series of surface surveys within each zone that would allow us to identify different site types: winter camps, summer camps, mixed herding and farming settlements, urban centers, and rural outposts from the late Iron Age (ca. 600 BC through the end of the medieval Mongol period (1300 AD). Due to logistical constraints the actual survey reconnaissance only took place in Zones 1 and 2, although we do have excavation data from the medieval Silk Route city of Talgar in Zone 3, and excavation data from the late Iron age site of Tuzusai located in Zone 5. The following environmental zones are: (1) mountain coniferous forest and meadow zone on dark forest soils (1300 to 1100 m); (2) the mountain decidous forest/meadow zone with chernozem-like soils (1100 to 1000 m); 3) the herb-bunch grass steppe with chernozem soil zone (1000 to 850 m); (4) the dry bunch grass steppe with dark chestnut soil zone (850 to 750 m); and (5) the desert steppe with light chestnut soil zone (750 m to 675 m) (Envirc 1995) .

One method for designing a settlement model for prehistoric and historic settlement of the Talgar region is to examine the ethnographic patterns of Kazakh land use. Contemporary Kazakh pastoralists graze their herds of sheep, cattle, and horses across these zones. In 1995 and 1996 we observed that the first two foothill zones of coniferous and deciduous forest and meadows were used by herders grazing their sheep, cattle, and horses on the upland terraces during the summer months. Other herders had taken their herds to higher elevations for jailaq (summer pasture). However, our survey reconnaissances identified a set of five or six Kazakh winter houses (zemovka) from the pre-collectivization period (before the 1930s) in both upland zones. This suggests that it is incorrect to adopt a rigid notion of vertical zonation for dividing pasture areas into summer versus winter grazing areas.

Some broad generalizations can be drawn from our own ethnographic observations of Kazakh mobility: (1) the steppe zones are most likely used as winter grazing areas; (2) the foothill zones are transitional between winter and summer grazing areas; and (3) the mountain meadows at elevations above 1300 m are summer grazing areas. The factor that complicates a simple model of vertical zonation for defining seasonal grazing is the cultivation system. Zones 1 and 2 are also very rich agricultural areas today where wheat, vegetable gardens, and orchards are grown. They are protected upland areas that have permanent water sources (streams and springs) and warm microclimatic temperatures. Often transhumant herders must choose their grazing areas in zones that are not currently planted in crops. This may be the reason why Zones 1 and 2 were chosen for Kazakh winter places and not as summer grazing places during the historic period. In Mediterranean pastoral systems we have frequently observed situations where winter and summer grazing areas exist in all the "empty niches" where cultivation is currently not practiced (Chang and Tourtellotte 1992; Chang 1993). Other alternatives exist: (1) herders could plant ceral crops and then move to higher elevations for grazing, and (2) during the winter the herds could graze in steppe areas off stubble or fallow fields. Thus cereal cultivation in the Talgar Region need not require year-round permanent settlements and could fit quite nicely with the short-distance vertical transhumance of animal herds.

The Site Data

The following site data is based on test excavations, block excavations, and surface survey reconnaisance done in 1994, 1995, and 1996.

In Zone 1, surface survey and test excavations were conducted at the upland valley or Orman, an upland area of approximately 12 sq.km and ranging in elevation from 1550 to 1300 m. Six surface artifact scatters were found during the spring of 1995. The sites were identified as two Saka period sites (600 BC to 300 BC), one Wusun period site (300 BC to 300 AD), a medieval period site (700 AD to 1300 AD), a Cossack site (19th century AD), and a Kazakh zemovka (winter place) that probably dates to the pre-Collectivization period of the Soviet Era. In 1996 several Bronze age graves was identified (ca. 1800 BC to 1100 BC). A series of modern and historic irrigation canals were also found at Orman. There is some depositional evidence that the canals could have been used as water diversion channels as far back as the medieval period. These site data indicate that Zone 1, was used for both upland farming and transhumant herding from the Bronze Age through the post-medieval period. The two Saka period, one Wusun period, and one medieval period sites were probably outlying settlements.

In Zone 2, a text excavation was conducted at an upland valley below Orman at an elevation between 1200 and 1300 m. The test excavations at Talgar #3 show two distinct ceramic sequences: the Wusun period (300 BC to 400 AD) and the early medieval period (700 AD to 800 AD). Foundations of five Kazakh zemovka (winter places) were also identified in this upland valley. The upland valley is formed by a series of alluvial terraces. No other surface sites were found in this upland valley. Once again the upland valley could have been used for upland farming and as grazing areas for herders during the winter months or on their way to or from summer grazing in the higher mountain meadows. Talgar #3 was a small, hinterland site that was politically tied to the large urban center of medieval Talgar. The site could have been a seasonal herding camp or an outlying farm stead. Animal bones of sheep and cattle were found at Talgar #3, indicating the existence of an animal herd component at the site.

In Zone 3, the medieval Silk Route city of Talgar is located at the elevation of 920 to 940 m in the rich chernozem soil zone. A 30 ha area was occupied by city inhabitants and a 9 ha area was enclosed within city walls (Savelieva 1994). The block excavations made by Tamara Savelieva in 1994 and 1995 fit into the later period of site occupation (11th to 13th century AD). From these two field seasons we took the first paleoethnobotanical samples (2 flotation samples and 5 phytolith samples in 1994 were taken from a large block excavation in the area of the southwest tower of the city wall and 1 flotation sample and 1 phytolith sample was taken from the 1995 block excavation outside the northeast corner of the city walls). While the samples taken from hearths and ash dumps are too few to draw any valid statistical conclusions, the results from these first paleoethnobotanical analyses are so promising that more samples need to be taken in subsequent field seasons.

From the flotation samples Miller (1995) has identified seeds of wheat, barley, oats, and fodder. One sample had a variety of wild and weedy seeds indicating the presence of fuel dung. Since some of these wild and weedy seed species are not listed in our species inventory for the five vegetatonal zones, we suggest that herd animals producing this dung had been grazed outside the immediate zones (Chang, Savelieva, and Baipaikov 1995). The identification of wild and weedy seeds from fuel dung might be a very useful method for tracing herd mobility; identified species that were not present in the five immediate zones might indicate that animal herds were grazed in distant places on other plant communities. Rosen's (1995) study of the phytoliths (microscopic opal silicates of plant material) indicate a predominance of wheat and lower percentages of barley. In the one phytolith sample from Talgar 95 excavations, the dominant percentages came from Setaria (millet). These data indicate the existence of cereal cultivation in which wheat and millet were the dominant species, followed by barley, oats, and fodder plants. The irrigation canals found in Zone 1 at Orman which could be medieval constructions suggest the existence of both rainfall and irrigation agriculture.

Unfortunately zooarchaeological studies have not been conducted at Talgar. The following species identifications have been made: sheep/goats, cattle, horses, and camels. Residential areas outside the city walls consisted of rectangular sections of 12 to 14 houses, each house associated with a walled enclosure for herd animals. The size of these individual households appear to be from two to four times of the size of homesteads reported from the other medieval cities of Kazakhstan (Savelieva 1994). The pastoral orientation of some of the city dwellers is reflected in this kind of residential pattern.

According to our model of settlement, this zone should be the most productive agricultural area. The one drawback for agriculture in the vicinity of the medieval city is the the Talgar River, a turbulent, active river that deposits boulders and heavy gravels on the alluvial fan. During the colonial Russian occupation of Talgar, homesteaders built along the Talgar River therefore reserving the most productive lands for cultivation. The medieval inhabitants of Talgar probably also chose this settlement strategy, thus leaving the productive soils with no glacial or river gravel deposits available for cereal cultivation. In addition to rainfall and irrigation agriculture, short-distance vertical transhumance and long-distance meridonial transhumance could have been practiced by the medieval herders.

In Zone 4, no archaeological sites were surveyed or tested during our three field seasons. This steppe zone should be a productive area for both cereal farming and animal herding.

In Zone 5, the desert steppe, about 15 km due north of Talgar, is the late Iron Age site of Tuzusai (ca. 400 BC to 100 AD). The site, which has been excavated by the Kazakh Ministry of Culture since 1992 (Grigoriev 1992; 1993). Since 1994, we have spent three field seasons digging a total area of 6 m by 18 m excavation unit. The site was associated with 6 burial kurgans, only one still stands. The site appears to be a large village settlement. We are excavating in a peripheral area of the site. The complex stratigraphy suggests at least six different occupation levels: (1) a recent historic occupation (19th to 20th century AD) and (5) levels dating between 400 BC and 100 AD (based on ceramic typologies and a calibrated radiocarbon sequence of eight AMS and conventional dates run by Beta Analytic, Inc.). The architecture found in the large block excavations from five field seasons consists of more than 30 storage and trash pits, fire pits and hearths, a post mould, floor, and mudbrick or adobe walls. Some of the large rectangular pits may be semi-subterranean house structures although the two site directors, Chang and Grigoriev, disagree over this interpretation. Zooarchaeological analysis from the 1994 excavations of NISP (Number of Individual Species Present) yielded these results sheep/goats - 55 percent, cattle - 26 percent, and horse - 15 percent. From the 1995 excavations the (NISP) breakdowns were: sheep/goats - 49 percent, cattle - 31 percent, and horse - 13 percent.