A,,"¿\noeø\oV7 a, n c! k-lt^ vloqvÇvLa.c u\.tT/ -t \"\(,J?\p;;¿ì"f'utì't onq.

ád\\-/k2
thillrasún Steppe of Soutbezstcrn Kazakhstan / I O
FOL 3740
8 1994).I1lc Saka, an early Iron Age nomadic culrure, have been characrerized
as horse-riding populations who practiced nomadic, semisedentar¡ and scden-
tary v¡ays of life, depending upon their adaptations ro rhe diverse conditions of
The Grass Is Greener On the Other Side
the deserr-oases and outlying sreppe regions (Yablonsþ 1995: ZZ9). These pre-
A Study of Pasroral Mobility on the Eurasian Steppe historic reconstructions of pastoral nomadism have been based upon marerials
of Sourheastern Kazakhstan excavared from burial mounds.
Litde is known abour che general secclemenr parrerns or lifeways of che 're-

.LAUDTA .HANG ää1ïî'i:;.ï:::':ï,ï;:::ï::ï::::i":äåi:il::ï:l:?î:î,i;

in rhese pastoral nomadic cultures are inferred'rather than documenred. Two
central cultural historical quesrions have guided the archaeology ofthis region:
( t ) How dld the pascoral nomads of this are a contribute co large-scale migrations
The main obje crive of rhis chapre r is ro addre ss the issue of pastoral mobility
in
th. r.-p.r"t. grassland environmenrs of rhe norrhern Tian Shan Mountains of culural grouPs that reputedly took place in both periods? and (2) w'ere the
of Iron Age pastoral nomads such as the Saka and wusun the catalysm for change
,ooth."rr..n K"rakhst"rr. In parcicular, erhnographic observarrons on contem-
porary Kazakh pasroral *obiìity are used ,o piouìd. a set ofworkng srraregies and innovation over the vast grasslands of Eurasia? The goal of my research has
and methods fo, ,..o.rr,rrrcring prehistoric pastoral lifeways from survey and been to quesdon and reinvestigate these models ofpastoral nomadism.
excavation dara colle cred by the Kazakh-American Tälgar project. I also evaluate
some of the assumptions pur forrh by Sovier and post-Soviet archaeologists on
the origins and developmenr of nomadic steppe cukures. These archaeãlogists
Tr¡B SruoyAnEl
have characterized the formative stages for che evolurion of pastoral nomadism The srudy area is 25 to 80 km east ofAlmat¡ che largesc ciry of rhe Republic of
in this region of the Eurasian steppe as the Bronze Age (t700 BC to 900 BC) Kazakhsran, along the foorhills and alluvial fan areas of the Zailiisþ Alacau,
and Iron Age (ca. 800 BC to AD 500) (Akishev 1990; Alexeev l99l). a northern range of the Tian Shan Mountains. \Øe specifically chose two dis-
I examine archaeological notions of pastoral mobilit¡ how mobility can be tincc environmen al zones for survey, (l) the Talgar fan, an alluvial fan or apron
found in the archaeological record, and the relationship berween mobiliry and formed from the Tälgar River rhac exrends l0 to 15 km from the edge of the
population density of pastoral adaptations. Indeed, rhe erhnographic and eth- Talgar foothills; and (2) the Tirrgen/Asi upland valleys, a series of broad alpine
noarchaeological observations on pastoral adaptations in rhis region of Eurasia river valleys nesded between the high glacier peaks of the Zailiisþ Alatau.
suggest that certain forms of pastoral mobiliry should be extant in rhe archaeo- The Tälgar fan ranges in elevation from 550 m to 1,200 m and consisrs of
logical record. Most notabl¡ patterns ofvertical transhumance-where herders grasslands, forested areas, cultivated cereal fields, orchards, and urban and. rural
managing sheep and goars, catde, and horses move berween the lowland sreppe housing development and infrastrucrure. fhe upper Turgen valley (known as
areas (ca. 1,100 to 550 m in elevation) for fall rhrough spring grazing lands and OiJailau) and rhe Asi River valley range in elevation from2,200 m ro 2,600 m
the upland alpine meadows of the Tian Shan foothill regions (ca. 1,800 ro 2,600 and are vegetated wich meadows, steppe grasslands, and conifer foresrs on rhe
m in elevation) for summer grazing lands (July through Sepcember)--should northern exposures of rhe mounrain slopes. These upland valleys are situared in
be apparent in the archaeological record. Yet the models of pastoral mobility a national forest and thus have been set aside for recrearional and pasroral land
put forth by the Soviet scholars have been skewed toward documenting the dis- use. Herd management has been collectivized since the 1930s; Kazakh herders
tribution of burial kurgans and graves, while overlooking sites o. p1".., used as keep mixed herds of sheep and goats, cattle, and horses in borh environmental
pastoral loci (such as habitations, campsites, herding facilides, and shrines). zones. IJsually rhe contemporary herders pracric€ short-distance vertical crans-
The Andronovo culture of the Bronze Age has be en described as a sreppe- humance, moving between che upland pastures of Asi and Türgen (areas of sum-
based nomadic pastoral adaptation that brought innovarion and change to the mer pasture) and the alluvial fans (areas of winter pasture) of che Tälgar, Issyk,
agriculrural se¡dements of the desert-oases of Cenual Asia proper (Hieberr Turgen, and Chilik fuvers.

ØL ð? qO n''
" ï i,iå,$'Jl'oï li'J':'r'J;';
BIBLIOTECA
lE6 / LIaLdta Cba,ng l'4'tor¿l 1YlootLrry on t0t Lurann òtcllc uJ òuøLtrcbtér rt t\4.Ærr)t4fl / lo /

CoNrEup oR¡.Ry K¡z¡.xrr Pasron¡.r Mosrrrry landscapes of northern Greece and southeascern Kazakhstan usually involves
two 6xed residential places as well as a series ofplaces to which herders travel on
tVhen the topic of mobility is addressed by 'Western archae ology, examples are logistical forays, in search ofgrazing land and water. Ifherders travel far from
usually drawn from contemporary hunter-gatherers or horticulturalists thar their encampments, rhey settle at night with their herds at a corral or place that
continue to practice mobility (Kelly L992; Hard and Merrill L992).Themodels can be protected from wolves.
used for examining mobiliry generally rely upon Binford's (1980, 1982) distinc- The single most important factor in choosingwhere to move a flock or herd
tions between residential mobiliry and logistical mobiliry. For a clear illustra- is population pressure . The herder musc conside r carrying capacity-the number
tion ofthese discinctions and their pertinence to hunrer-gatherer populations, ofanimals that can be supported on a given area ofgrazing lands (Barth t96I).
I quore Kelly (t992:44): Too many animals on any given grazing rerritory result in degradation of the
pastures and ultimately reduce the carrying capaciry of the land. Most herders
Collectors move sidendally to ke y locations (e.g. ware r sources) and use
re
realize this, especially if rhey milk their animals, because the quality of milk
long logistical forays to bring resources Eo cemp. Foragers 'map onto" a
declines as the forage qualiry declines (Koster L977).The Kazakh keep cattle
region's resource locations. In general, foragers do nor srore food; rhey
and horses that they milk in order to produce cheese or kurnlrs (maret milk), so
make frequent residential moves and short logistical forays. Collectors
individual herders note the impact ofpoor forage on the quality and quantity of
store food; they make infrequent residenrial moves but long logistical
milk produccion.
forays.
Kazakh herders prectice wh*Kel\y
(1992: 45) describes as territorial or long'
Does the distinction between "foraging" and "collecting" epply ro pastoral term mobilitl.Building on Binfordt (1982, 1983) deÊnition, Kelly describes ter-
adaptarions? Kazakh herders, like all mobile herders, move rheir'food" or ac ritorial or long-term mobility as cyclical movements of a group utilizing a set
least their livestock to them. In the "raw sense" offood-gettinglogistics, pastoral ofterritories over a long period, such as a decade. \Øe have observed Kazakh
nomads are more like coilectors than chey are like foragers. Yet pasroral mobiliry herders who use the upland plateaus ofTurgen and Asi, returning every summer
is contingent upon moving the animals to adequate grazing terrirories and to to che samef ilau (surnrner pasture ) inJune through August. The location of the
available wate r as well as avoiding conflicts wirh othe r herders who also compete upland grazing territories might occasionally shift, although such changes have
for rhe best grazing and water resources. Thus most animal husbandry systems consequences. Herders who attempc to stake out a new grazing territory (closer
require a kind of "mapping on' srreregy, by which the herders lay claim to rer- ro the dirt track, for example) may discover that they are invading someone elset
ritories by moving from their camps (usually in a fixed place) on shorr logistical territory and must therefore compete with others for available pasture. In the
forays. Of course, the major difference betwe en hunter-gatherers and pastoral- summer of 2002, when rainfall was plendful in che upland Asi Valley (cz-2,200
ists is that pastoralists always move with their food source but musr maximize m in elevation), several households shared a broad valley along che Asi fuver ap-
the general health and condirion of rheir herds and flocks by establishing some proximately I km long that had been previously occupie d by a single household
means for claiming the best grazing lands and wacer sources for themselves. and its herd. The son of one of the "newcomersi a herd owner of over six hun-
Tianshumant herders in northern Greece lay claim to grazing areas by attempr- dred sheep and goats, catde, and horses, informed us thar his father intended to
ing'co pack their flocks" in a given territory marked by the locarion of their return to his previous territory the following year'
animal folds (Chang and Tourrellorte 1993). Herders in southern Greece, when Although cyclical or long-term mobility is practiced, thè more rypical pat-
confronted wirh en invade r attempting ro usurp grazing lands in the village com- tern of mobiliry involves rerritories thac are 6xed between rwo known points:
munal lands, mighr resort to violence or fold burnings (Koster 1977). the summer yrrt and the winter residence in a small town, village, or collective.
My echnographic observations of contemporary Kazakh herders (who main- Importandy, however, summer pesture territories involve more flexible patterns
tain sheep and goats, catde, and horses on the same landscapes as did their ofuse-rights, because upland grazing areas are considered to be "open' territo-
ancient predecessors) suggesr rhar they employ a mobility srraregy similar to ries or common pasture lands.
Binford's "collectors." Contemporary Kazakh herders pracrice shorr-discance Usually the main facilities-the felt ¡rrt, a corral for holding the animals dur-
vertical transhumance (wintering in lowland areas and summering in upland ing the nighl a herding household's
and milking facilities--mark the location of
areas) (Akishev 1990). Pastoral rranshumance as I have observed ic on rhe rural grazingterritory. The system of herding requires the separation of the mixed
188 / Cl4ud.ta Chang lastoral Mobility on the Lurasian Steppe oJ Southeasten Kazakbstan / 189

herd inco diffe¡enc grazing units by species. For example, rwenry horses may sampling designs. Until the mid-1990s (when we inrroduced che pracrice of pe-
be tended by one household member, forty cattle by another, and five hundred destrian surface survey) arrifecr scarlers of ceramics, animal bones, and grinding
sheep by ye t another. If the grazing territory is exceptionally rugged or inhabired stones found on the surfaceofplowed Êelds were overlooked and ignored. Even
by wolves, experienced herders are pur in charge of the flock. The composition the preliminary reconnaissance suryeys yielded information on arrifacr scarrers
of the Kazakh herding household may flucruate in rhe summer pasrure area. In that fall into the general chronological caregories of the Bronze Age, Iron Age,
1997 we m€t an older man who was spending his summer with his two daugh- or medieval period, based on ceramic typologies of the surface materials. The
ters-in-law and their children and rheir herd of a rhousand animals, while his local archaeologists have been able ro place rhe burial mounds into chronologi-
sons cultivated their agricultural holdings in rhe lowlands. The following year cal periods on the basis of dimensions (height and diamerer of burial mounds)
the sons and father returned to the same territory with their herd, bur without ¿nd surface features (srone circles, slab-lined cisrs, stone or soil matrix of rhe
their wives. These e rhnographic observations suggest the variarions with regard mound). Some archaeological fearures such aí the foundarions of sod-houses
to household labor, mobility parEerns, and camp locations. known zs zimouki (winter dwellings) have also be en idenrified, as well es srone-
coursed architecture from che Bronze and Iron Ages.
R¡sE¡.ncHrNG MoBrlrry rN THE ARcHAEor.ocrcAr, REcoRD A definitive phase designacion within broad chronological periods has yet to
be constructed for the Bronze Age or Iron Age ccramic sherds of rhe Semirechye
The current models of pastoral mobiliry adopted by the Soviet archaeologists region.'we can only make rough estimares thar place arrifacr scarrers or isolared
are based upon two systems of pasroral movemenc on the Eurasian steppe, The ceramic finds inro rhe following chronology: (l) Bronze Age (ca. 1700-900
first is a long-distance system of horizontal movement across the steppes, where BC); (2) Iron Age (ca.700 BC-AD 500); (3) Turkic Period (ca. AD 600-900);
summer pestures are locared in the south and winter pasrures are locared in the (4) Medieval Islamic Period (ca. AD 800-1250); (5) Mongot Period (AD 1250-
north; disrances can range from 100 km ro 1,000 km. The second is a short- 1500); and (6) historic Kazakh Period (ca. AD 1700 ro present). Obviously the
distance sysrem of vertical movemenr from the mounrains and foothills to rhe Iack of a more precise chronological framework wirhin chese broad labels limirs
Iowland valleys, where summer pasrurcs a¡e locared in rhe uplands and winrer our ability to define subphases and subsequenrl)'ro sorr our palimpsescs.
pastures are locared in the lowlands; distances between pastures can range from Archaeologists workng in rhe wesrern hemisphere have noted the method-
50 km to 100 km (Akishev 1990). ological and rheoretical problems with using surface survey daca to infe r ancienr
Contemporary pastoral transhumance between rhe upland valleys ofTurgen settlementpatrerns (Dewar 199l; Plog 1973; Rouse L972).Inperticular, Rober¡
and Asi (ca.2,200 to 2,600 m in elevacion) and the lowland steppe of Tälgar, Dewar ( t991) has commenred exte nsively on rhe fact that survey dara have been
Turgen, Issyk, and Chilik (ca. 1,100 ro 550 m in elevation) suggesrs rhat the cur- misused in setrlement-parrern analysis. He points our rhar secdemenr-patrern
re nt models ofpastoral mobiliry for the Bronze and Iron Age are inadequate and analysis often treats archaeological components within a single phase or period
lack sufficienc empirical evidence. Since 1994¡heKazakJt-America¡r Talgar Proj- es contemporaneous, akhough sites wirhin a given period or phase may not be
ect has conductèd surface surveys and excavations wirh the expecràrion thar the contemporancous or may even represent overlapping occuparional periods,
economies and land-use srrategies of the ancient nomads can be reconstructed While archaeologists
are fully aware that survey daca represenr remnanrs of pasc
from archaeological data. Our merhods, although standard for ìØestern archae- settlement-systems (Dewar I99L:604),rhey still use such daca ro derive popula-
ology, differ from rhe long rradicion ofRussian and Sovier period archaeological tion estimates and the spatial discribution ofsettlements across given landscapes.
research on che Eurasian sreppe, which has been based upon evidence drawn In the case ofour data, archaeological sites rhat conceivably span a rhousend
from ancient texts and from the archae ology of morcuary complexes. Our survey years but are placed within a single period (for example, the Iron Age) can hardly
and excavarions are "works in progress" that have their own limications. In the be considered suitable for detailed serdemenr-pactern analysis.
following se ctions I discuss how we designe d our research, rhe rheoretical frame-
work and methods we employed, and the preliminary resulrs of this research.
SunvBy Mr,rHoDor,ocy AND RxsuLTs
Ideall¡ an archaeological projecr designed ro research pastoralism-both
as an economy and as a land-use srraregy-would require fine-grained chrono- The Kazakh-American Talgar Project conducted pedestrian surveys from 1997
logical sequences and full spacial coverage ofrhe study area rhrough the use of through L999 on the Talgar alluvial fan, a broad delta formed by the north-
190 / Llaudla Chang

flowing Talgar fuver (approximarely 150 sq km). W'e walked ovet 287 trensects
in plowed Êelds and along stream cuts. Our goal was to cover as broad an area .lables.l.ChronologyofTuzusail(ExcavatedLgg2-|996),.|segankaS(Excavacedlgg8-2000),and
of the fan as possible, using the Talgar River as a natural boundary and dividing taldy Bulak 2 (2001)
the fan inco eastern and western sections. ]ùØe did so with a random sampling Radiometric date Calibrared result
StrarigraPhic sequence (2 Sigma,957o)
stretegy.
From 1994 through 2002 the Kazakh-American Talgar Project also exca-
VIII
vated four Iron Age sites on the Talgar alluvial fan and one Iron Age and one r40f70 BP CaIAD 1650 to 1950
Tuzusai, Occupation 5, Unit V-l I' Fire Pit I
multicomponent Iron Age and Bronze Age site in the upland Turgen and Asi
VIT
valleys. Most of the Iron Age sites show multiple occuPation levels based upon 650+50 BP' CaIAD 1275rc1410
Tuzusai, Burial l, Animal bone collagen
in situ raàiocarbon-dated conrexts. These excavated sires usually have four to From sheeP scaPula
eighc srrarigraphic levels, indicating repeared occuparions. Tuzusai (a small vil- VI
2ç29+40 BP- (oxford) Cal BC I00 to AD 75
Iage hamlet) has evidence of at least six different horizons and four occuPation Tuzusai, Pit 24 Âll
zo70+¡4} BP'(Oxford) Cal BC 180 ro AD 25
levels. At Tseganka to eight occupational and building sequences have been
8 six Tuzusai, Pit 17 6ll
documenred for archicectural features such as pit houses. The Taldy Bulak 2 VIa
2r90+80 BP' Cal BC 400 to 40
Tseganka 8, Unir V-10, on subsoil
sire has six differenr occuparional levels designated for different activity areas
and fearures. The meaning ofthese sequences ofsite reoccuPation is not entirely 2170+30 BP Cal BC 335 ro 290 and
Tuzusai, Pit 30 B ll5
cle ar. They could represent ( I ) shifring locations of hamle ts and small residential
BC230to
Cal BC 350 to 300 and
camps spanning a given occupational period; (2) repeatcd seasonal occuPetion Tscganka 8, Pit l3 bocom 2130+80 BP
Cel BC 2ZO rc 50
by groups of mobile pastoralists or mixed herding-farming groups; or (3) aban-
2280+40 BP' Cal BC 400 to 350 and
donment and then reoccupation by sedentary groups' Taldy Bulak 2, Fire Pit 2, Horizon 3
Cal BC 310 ro 2I0
\Øhat is particularly signiÊcant about the excavations et Tuzusai, Teganka 8, Cal BC 380 co 190
2230+30 BP
and Taldy Bulak 2 (all Iron Age sites from the Talgar dluvial fan) is the overlap- Tuzusai, Pit 29
zL70+608P Cat BC 380 to 40
Tuzus¿i, Unit V- 13, ash dePosic
ping radiocarbon sequence of Phases I-VI, spanning from 775 BC to AD 75 2130r40 BP- Cal BC 385 to I00
Tlegarrka 8, Pithouse L,FlootZ
(Chang et al. 2002). All three sites aPPear to have overlapping periods ofoccu-
IV
parion in Phase V spanning from 400 to 40 BC. These preliminary data suggest 2310+50 BP'(Oxford) Cal BC 415 to 345 and
Tuzusai, Pit 22
BC 3I0 to 210
thar the demographic expansion of the Iron Age, as represente d by burial sites
Z3ZO+40 BP. (Groningen) Cal BC 410 ro 260 and
and serdement sites, might have taken place during Phase V. For the regional Tuzusai, Pit 8-'92 Cal BC 230 ro I t5
cukural history of Semirechye rhese two periods of occupation are of special 1

interest, since the splendid Golden IØarrior tomb, located in Issyk (about 20 i
ilt C¿|BC775 to 370
2390!708P
km to the easr of the Talgar fan) dates from 400 co 200 BC. If indeed Phase V Tseganka 8, Pithouse 3, Floors 3alb
2400+70 BP Cal BC 780 to 370
Talãy Bulak 2, Unit D-8, Ash Pit' Horizon 4
does represent the peak of Iron Age setdement and demographic exPansion, this
lt
suggesrs rhat it is also the formative period when the height of Saka wealch and 2190140 BP Cal BC 350 to 310 and
T3eganka 8, Pithouse 3, Floor 4
status differentiation took place. Cal C 2I0 to 40

The single most important factor for testing these assumptions about de-
mographic expansion and the evolution ofhierarchy (as apparent from the ar-
I Tseganka 8, Storage Pit'98 2300180 BP' Câl BC 740 to 710 and
BC 535 co 80
chaeological remains of an extensive mortuery complex) is to develop a tight
chronological framework with phase designations r}rat can address the issues calibra¡ions were donc bY Beta Anal¡ic, lnc. (Stuivcr 1998)'
N¿¡¿: Thc rediometric deting and
of (l) frequency of populadon relocation and (z) length of phases (Dewar r AMS (Accelcraccd Mess Spccrrometry) darcs-obraincd
at Beta Analytic, lnc', unless sPeci6cd'

l99l: 605).If our assumption that the Iron Age populations of the Talgar and
Turgen/Asi aree were mobile or at least semisedentary is correct, however, the

s ffült1"uru sp
192 / Llaud.ta Lhan& l'astoral MoþtlrtJ on tbe Lurasla.n JtePPe oJ ;1utbeaster-n Kaz4khstan / ly5

problem of inferring setd€menr pacterns from survey and excavation data will sites (burial mounds) and settlement sites during rhese phases should be indica-
be even more serious. The paleoethnobotanical and faunal materials collected tive of tlre height of demographic expansion on the alluvial fans of Semirechye.
from Taldy Bulak 2, Tuzusai, and Tseganka 8 clearly demonstrate the presence It would also make sense that the increased population density found in these
of a mixed economy based upon cereal cultiva¡ion ofwheat, millet, and barley phases corresponds to increasingpatterns ofsocial stretification (noted from the
and a herding system reliant upon sheep and goats, catde, and horses (with very inventories found at the burial mounds) and greater reliance upon agrarian food
small percencages of camels) (Benecke 1999-2O00; Rosen ec al. 2000). Almost production (Akishev and Kushaev 1963: Chang and Tourtellotte 1998).
all of the Talgar fan Iron Age settlements currendy excavated show multiple
occupetion levels, chus indicating that a single location was used, abandoned, THp PnosrnM oF SITE oR PLAcE Vrsrsrr,rtv
and rhen reused over a period ofsix hund¡ed or seven hundred years. Does this IN THE ANCTT¿,BOI.OGICAL R.ECONO
represent long-term pastoral mobilit¡ seasonal mobility, or repeated sequences
ofuse, abandonment, and reoccupation? Such questions can be posed bur not Local archaeologists have conducted their own survey reconnaissance and
answered by our data. have documented and inventoried known archaeological sites, plotting these
Some sites such as Tseganka 8 appear to be occupied in a confined locale, locacions on a l:500,000 scale Arheologishaya Karta Kazakhstana (Ageeva et
where repeated floors and occuparion levels show dense concentretions ofarti- al,. 1960). Western-style archaeological surveys, however, have only been intro-

facrs. Orher sites such as Taldy Bulak 2 and Tuzusai are distributed over a large duce d since the mid-1990s by the Kazakh-American Talgar Project (Chang et al.
territory (up co I kmz). Does this suggest that Taldy Bulak 2 and Tuzusai were 1999). Surface surveys were made by pedestrian walking in upland areas where
temporary encempments that were reoccupied year after year, in the same man- site visibiliry (especially scone oudines o€houses, burial mounds, and graves) is
ner that the contemporary Kazakh yurts and encampments are reoccupied? \Øe high and on the alluvial fan in plowed fields and along river and erosion cucs.
have noted that the Kazakh encempments might shift 10 to 100 m from season Isolated 6nds of ceramics, grinding stones, and other artifacrs were recorded,
ro season in the upland placeaus of Asi and Turgen. Low-densitysires spread as well as artifact scacters and architectural features. Each cluster ofartifacts or

over large areas could represent seasonal temporary encampments with many architectural features within a 100 m2 area wes re corded as a locus. The results of
differenr activiry areas, while tightly packed pithouse sites wit-h repeated occu- our survey have been reported elsewhe re (Chang and Tourtellocte 2000). Here
pations in a single con6ned locale represent hamlets occupied on a permenent, I provide date on our preliminary resulcs from these surveys.
year-round basis. In Pithouse 3 at Tseganka 8 we noted the thick packing of one
floor level over rhe next and recorded over six to eight different flooring layers,
often with lictle or no Êll levels between some of the floor levels (3a and 3b). Täble 8.2. Survey Results from rhe Tâlgar Alluvial Fn (1997-1999) and rhe Turgen-
Could rhese repeated foor layers r€present episodes ofcontinuous occupations Asi ( 1997-2002) Surveys
or pe riodic remodeling of dwe llings that were occupied on a pe rmanen! or semi- Locus caregory Talgar alluvial fan Turgen-Asi uplands
permanent basii ?
Iron Age sitcs 59
Until we have excavated more sires and noted the variation in features, oc-
Probable Iron Age sices (sherd scartcrs) 29
cupation episodes, and cheir palimpsest nature, we cannot determine the settle-
Iron Age kurgans 182 60
ment patterns for these agro-pastoral populations. Yet the wide variety of fea- Loci with I sherd r00
cures, arrifact distribucions, and densities at the three excavated Iron Age sites Loci wirh 2 sherds r6
ove rlapping in time sugge sts a wide range ofvariation in the rypes of s€tdements Loci with 3 sherds 25
and episodes ofuse in any give phase or subphase. Setdement packing (ofboth Loci with grinding scone r6
people and animais) must have been a problem during the height of the Iron,A.ge Bronze Age sites 6
?q
occuparion of the Talgar alluvial fan (Phases IV through VI). The majority of Bronze Age kurgans
Mcdieval sites 2
burial mounds in Talgar, Issyk, and Bes Shatyr also date to this period (ca. 400
Kurgans oFunknown period 85
BC-100 AD), often labeled the Saka-ìØusun period by Soviet archaeologists
Totals 427 r87
(Moshkova 1992). Clearly the relarionship between high numbers of mortuary
t94 / Claudia Chang Pastoral Mobility on tbe Eurasian Steppe ofSoutbeastern Kdzakhst¿n / 195

An area of approximately 150 kmz in the Talgar alluvial fan yielded a total dung), a fe ature that might not preserve well in the archae ological re cord. Ironi-
of 427 Ioci, a density of abour 2.8 loci per km2. About 270 ofthese Locí (63Vo) cally, the Bronze and Iron Age mortuary sites in the uplands are more visible,
were idenriÊed as probable Iron Age sires. The remainingwere Bronze Age, me- because the surfece topography and geomorphology ofthe uplands contribute

dieval, or of indererminate period. In an area of approximaæIy 46 kmz in the ro high sice visibiliry.'ï?'e may have been unable to 6nd many settlement sites
Turgen,/Asi upland valleys, approximately 187 loci were recorded, a density of in the uplands because the artifact scatters associated with such buried sites are
about 4.1 loci per kmz. Abour 35%o of these loci are lron Age, and about l9o/o zre invisible in these upland grasslands. Bronze Age burials include cist-lined graves
Bronze Age. only six Bronze Age settlemenrs and seven Iron Age setdements and ossuaries pIaced inside reccangular stone wall structures, while graves from

were found; the remaining loci are kurgans (burial mounds). rhe Iron Age through che Turkic and medieval periods are marked by burial
Sice visibility was a far greater problem on rhe Talgar alluvial fan than in the mounds. The Iron Age to medieval burial mounds could conceivably be used as
upland valleys of Turgen and Asi.'\Øe occasionally found Iron Age artifacts in indications of population densit¡ particularly because it was the custom to place
the profile curs of srreambeds bur not on the plowed surfaces, so it is clear that only one or two burials in each mound.
sires on the Talgar fan were often de eply burie d under the wind-blown and river- In contrast, the alluvial fan areas north of the Tian Shan Mountains have
deposited loess soils. The loci thar yielded the highesr number of surface arti- been expose d Eo processes of rapid soil deposition by wind or watet The small

facts (sherds, grinding srones, and bones) wefe often the most disrurbed and village and hamlet sites of the Iron Age are deeply buried, usually under 0.5 to
destroyed by modern-day agriculrural acriviries. ìØe also discovered that the re- I.0 m of loess. Deep plowing and bulldozing of prime agricultural land since the
Iarionship betwe en surface Ênds and subsurface remains was skewed, as noted by 1960s have exposed many archaeological sites in the Tâlgar alluvial fan.'We con-
tinue to be puzzled, however, by the lack of Bronze Age materials found there.
Jack Nance and Bruce Ball ( l9s6) whe n developing tesr pir
sampling stracegies
for che discovery of burie d sire s. Nance and Ball (1986) conclude that sites with Is this due to the geomorphology of the alluvial fan (Bronze Age sites could be

more surface arrifacrs probably represenr a higher density of buried artifâcts. covered by I m or more of loess), or does it represent the lack of Bronze Age
These sites are more likely to be recovered by using tesr pit sampling than by seftlement on the lowland steppe areas ?

surface surveying. In our surface surveys, ic may indeed be the case that both
lowdensity sites (surface and subsurface) and high-density sites are present, but INsrcnrs rNTo PASToRAL MoBrr,rry oN THE EURAsIAN SrEppE
deeply buried sites will be invisible and cherefor. tror found on surface surveys
(Shon lg95; wandsnider and camilli 1992). Many ofour surface surveys on the From this preliminary research, it appears rhat the Bronze Age populations
Talgar fan were done on plowed Êelds, already indicating the skewed nature of of Semirechye were pastoral nomads, while the Iron Age populations wete se-
these artifact scatters. misedentary agro-pascoralists. First, the locations ofthe Bronze Age setdements
Michael Shott (1995: 478) nores: "But surface documenrs [arrifact scatters] and mortuary complexes in the upland plateaus of Asi and Kurgen but nor in
the fercile lowlands suggest that the populations of thac period we re practicing
afe nor me rely limited and slightly skewed samples of underlying records [bur-
ied artiFacr deposits]. They also can con[ain a strong random element, such thet some kind of vertical transhumance. The contemporary climate (rainfall and
temperacure) statistics show that Asi and Turgen are located in areas without
successive episodes of cultivation do not necessarily expose similar numbers,
distributions, or kinds of artifacts.' enough frost-free days to sustain the cultivation of crops, The concemporary
Our erhnographic observations of Kazakh corrals and campsites in the up- Kazakh herders who pasture their animals during July through Sepcember in
\ùØhether
these upland valleys certainly do not prectice any form ofcultivation.
Iand valleys of Asi and Turgen indicate thac the pastoral nomadic camPs, es-
presenr-day climatic conditions are a good indicator of the exiscing clirnate dur-
pecially summer camps, have a low errifact density and would probably not
be visible after abandonment. A prime pastoral location used successively over
ing the Bronze Age clearly needs further investigacion.

many years, however, should yield a higher arrifact density over time. But these The number of lowland settlements and burial mounds and their overâll
repeated occuparions might show up as low-density surface remains over a large density on the Talgar alluvial fan indicate a demographic expansion of popula-
area (as at Taldy Bulak 2) rather than as dense concentrations ofartifaccs within tions in the thousand-year period of the Iron Age on the fertile alluvial fan. The
a confined area (as at Tseganka 8). The mosr disdng;ishing element of the pas- paleoethnobotanical data at Tuzusai, Tseganka 8, and Taldy Bulak 2 all indi-
cate the presence of cultivated species of millet, wheat, barle¡ and possibly rice
coral sice is the corral (marked by an archicectural feature and deposits of animal
ly6 / LIauAta Løang Pastoral Mobilxl on the Eurasian Steppe oJ Soutbeastern Kazdkhstan / 197

(Ä.le.re Rosen, personal communication). Yec chese data may be a product of (b) 'Were these species herded over sers of rerritories in a system of short-
the Iimitations of our survey methods and results. \Øe have yet ro locate Bronze distance verrical ûa¡rshumance (radius of50 km or less) or in a system of
Age deposits on the Talgar alluvial fan. It is probable tÀat agro-pascoralists or long-distance horizontal transhuma¡ce (radius of 50 to 500 km or more)?
nomadic pastoralisrs of the Andronovo period settle d or urilized these alluvial (c) V'hat was the sysrem of spacial organization used by agro-pastoralists in
fans, which would have provided good grassland environments for pastures or exploiting the Tâlgar alluvial fan?
for foraging and for incipient agriculture. (d) Did the agro-pastoraliscs spend the whole y€er ar a given village or ham-
Our current paleoethnoboranical and zooarchaeological research on the Iet, or did they circulare seasonally or yearly over a set ofterricories ?
excavated Iron age sites ofTuzusai, Tseganka 8, and Taldy Bulak 2 indicates a (e) Do the repeated occupations at the Iron Age setrlemenr sites of Talgar
mixed cereal economy of millet, wheat, barle¡ and possibly rice and a herding represent abandonment and reuse over a long period or wichin shorr-
component of sheep and goats, cattle, horses, and camels. The upland Bronze term phases?
Age sites of Asi I and Asi 2 (occupied in the earþ Andronovo period, ca. 1600
As more surveys and excavarions are conducted on rhe Talgar alluvial fan and
to 1400 BC) show no evidence of culcivated plants but do contain evidence
in the upland Tirrgen and Asi Valleys, the issue of pastoral mobiliry needs ro be
for the herding of sheep, goats, and catde. The Iron Age site of Kizil Bulak 3
considered. A larger, regional coverage ofboth areas may allow us to posit the
in che Turgen Valley is a seasonal site with over 12 different periods of occupa-
existence ofa system ofeirher long-disrance horizontal or short-distance yerrical
cion thar shows evidence of sheep, goets, cacde, and horses and wild species of
transhumance.'$7'e musc look for broad pafterns of spatial organization and in-
plants. Some grinding stones found at Kizil Bulak 3 also indicate processing of
dicators of seasonaliry ar sites found in borh environmenral zones. For example,
gachered plants or agricultural grains brought from the lowland agriculrural
if short-distance vertical rranshumance was practiced during the Iron Age of
areas.
Semirechye, then there should be evidence for winrer occupation on the Talgar
From such data we can infer chat the upland valleys ofTurgen and Asi were
alluvial Fan and summer occuparion in the upland Turgen and Asi Valleys. In
for seasonal lranshumance, most likely during the summer months.
besc suited
the same vein, if semisedentary agro-pastoralism was practiced in the Iron Age,
That does not necessarily mean, though, that rhe Bronze Age populations of
permanent year-round villages and hamlets should be found on the Tâlgar al-
Semirechye did not praccice agriculture in other are¿s. It s'ggests chat we have
luvial fan, while temporary campsites or house srructures should be found in the
found the place where agriculcure would have been mosr productive in both
upland Turgen and Asi Valleys. Such inferences musr be drawn from a compari-
the Bronze and Iron Ages: on the alluvial fans and in the lowlands. The Iron
son of the excavated materiels found at Iron Age sites in both the upland and
Age steppe sites represenr a mixed cereal and animal husbandry economy and
Iowland zones in conjunction with surface surveys. Indicarors such as shared
a setdement-pattern that indicates a year-round or seasonal occupation during
ceramic styles, seasonal usage, and tool assemblages of sites in both zones might
the summer when agrarian activities took place. The Iron Age sites found in
then allow compararive analysis of secdement perrerns and overcome some of
the uplands were not agricultural sites; che paleoethnobotanical data show only
the problems created by site visibility and palimpsests.
evidence of wild plancs (Rosen, personal communication). Such sites represent
Perhaps the greatest single stumbling block for the archaeology of this re-
the use of the uplands for pastoral activities during the summer months.
gion is the assumption rhâr pasroral nomadism was che sole economic base of
The Bronze Age excavations at Asi 2 suggest the existence of anomadic pasto-
che Bronze Age and Iron Age popularions. The concept of pastoral mobility
ral economy. Wherher the Bronze Age nomadic pastoral populadons pracciced
be came a me ans by which the Soviec period archae ologists could avoid conduct-
agriculcure is still unclear. If Bronze Age settlements could be identiÊed on the
ing settlement pattern-analysis. I agre e wholehearce dly wirh Dewar (1991) end
Talgar alluvial fan, we might be able to test this. Sdll, beyond these generaliza-
others who question setdement-pattern analysis when lengrh of occuparions at
tions, neither the excavations nor the surveys of rhe Tälgar alluvial fan or the
given sites is not considered adequarely. Bur a necessary Êrst step ofany srudy on
Turgen/Asi upland valleys are sufficient to produce an accurate reconstrucdon
pastoral nomadism in prehistory is detailed analysis of the distribution of sires
of mobility patterns during the Bronze Age or Iron Age in Semirechye.
across e physical space, especially space that can be demarcated into differenc
The following questions should guide future research in tle Semirechye re-
environmental zones. 'We know from the many studies of pastoral nomadism,
gion of southeastern Kazakhstan.
semisedencary pastoralism, and sedentary agro-pastoral groups thar rhere is tre-
(a) V'hat types of species were predominantly herded by the early Bronze mendous v¿riation in che spatial organization ofplaces in a given landscape uri-
Agepopularions? lized by people who spend some portion of rheir lives herding and husbanding
Ð+¿to
S ffia/ ft/îA Ë/U
SP
t98 / Claudia Chang Pastoral Mobilìty on thc Eurasian Steppe ofSoutbeasten Kazakbstan / 199

animels (see chang 1992). Thus it seems mandarory rhet archaeologists consider Benecke, N.
the spacial distribudon ofarchaeological loci across rhese landscapes, even ifrhe 1999-2000 Unpublished Reports on the Faunal Remains from lron Age Sices in rhc
sites themselves represent at best remnants of the past settlement-system. Talgar Alluvial Fan. On Êle in the Dcparrmenc of Anthropology and Sociology,
Sweet Briar College.
Binford, L. R.
AcxNowrrDGMENTs
1980 'lf'illow Smoke and Dogs'Tails: Hunter-Gatherer Secdemcnt Systems and A¡-
The Kazakh-American Talgar Project has been generously funded by the chae ological Sice Formation . Americzn Antiquitl 45(l): 4-ZO.
'!7'enner-Gren Foundation for Anthropological
Research and rhe National 1982 The Archacology ofPlæe.Jountal ofAntbropological Arcbaeology l: 5-31.
Ge ographic Socie ry from 1994 to 1996. Field rese arch from L994 through}}}Z 1983 In Pursuit of tbe Pøl. Thames and Hudson, London.

was funded by rhe National Science Foundation Archaeology Program Grant Chang, C.

SBR-9603361, "The Social Evolution of Steppe Communities in Southeastern l99Z Archaeological Landscapes: The Erhnoarchaeology oFPascoral Land Use in rhe
Grevena Province of Grecce. In Spaæ, Time, ønd Archaeological Landscapes, ed,-
Kazakhstan and the fuse of Civilization." K. M. Baipakov, F. P. Grigoriev, and
ited byJ. Rossignol and L. W'andsnider, pp. 65-89. Plenum Press, New York and
A. N. Mariashev, all of the Republic ofKazakhsan, were ofinvaluable assistance
London.
during the Âeldwork. P. A. Tourrellorte directed the surveys described in this
Chang, C., and P. A. Tourtellorte
chapter and provided rhe maps and photographs. I wish to thank pei-Lin yu,
1993 The Ethnoarchaeological Survey ofPasro¡al TianshumanceSires.Journal ofFieU
Frédréic Selle t, and Russell Greaves for inviting me ro parricipare in the sympo-
Arc hae o lo g1 (3) : 249 -26 4.
sium where rhis work was Êrst presenred. I welcomed rhe commenr of the two
1998 The Role of Âgro-pastoralism in the Evolution o[Steppc Culture in the Semire-
anonymous reviewers who read this manuscript, who urged me to consider the
chye Area of Southern Kazakhsran during the Saka/Wusun Period (600 BCE-
problem of mobiliry within borh the ethnoarchaeological and archaeological 400 CE).ln Bronze,4ge andEarþ Iron,4ge Peoples of Central,4sid,edited by Victor
contexts of nomadic pastoral sociecies. H. Mair, pp. 264-2T9.Insriturc of Man Press and the University of Pennsylvania
Press,'W'ashingron, D.C., and Philadelphia.
R¡rrn¡Ncns 2000 The Kazakh-American Talgar Projec Archacological Field Surveys in rhe Talgar
and Turgen-Asi Are¿s ofSoutheastcrn Kazakhstan: 1997-1999.1n Kurgans, Rit-
Ageeva, E. L, K. A. Alishev, G. A. Kushaev, A. G. Maksimova, and T. N. Senigova (com- u¿l S¡ta, and, Settlements: Eurasían Bronze and lron Áge, edited byJ. Davis-Kim-
pilers) ball, E. M. Murph¡ L. Koryakova, and L. T. Yablonsþ, pp. 83-88. BAR (British
1960 Arbeologicheskaya K¿rta Kazaþbstan¿. Afme-Ate: Izdateltwo Akademenni Nauk Archaeological Reporcs) Internarional Series 890.
Kazakhskoi SSR. Chang, C., P. A. Tourcellorre, K. M. Baipakov, and F. P. Grigoriev
Akishev, K. A. 1999 The Kazakh-American Talgar Archaeological Surveys in 1997 and 1998 in the
1990 Les nomades à cheval du Kazakhstan Antique. ln Nomades et sédentaires en Asie Talgar Region. tia, nauk (219) : 168
1z u es o b s b e s tu e nn1 e -I8 4.
Centrale: Apports de archéologie et de I'ethnologie, edited by H.-P. Francfort, pp. 2,002 The Evolution of Sreppe Communiries from rhe Bronze Age rhrough Medieval
15-18. Edicions du Cenrre National de Ia Recherche Scientifique, Paris. Periods in Sourheesrern Kazakhsran (Zhecysu). A. Kh. Margulan Instirute oFAr-
Akishev, K.,{., and G. A. Kushaev chaeolog¡ Sweet Briar/Almaty.
1963 Dreunlala Kul'tura Saþ,ou n Usunei Dolini Reka lli (AncientSaka and'!ü'usun Cul- Dewar, R. E.
ture of the IIi River Valley). Alma-Aca: Nauke. l99l Incorporating Variation in Occuparion Span into Settlement-Pattern Analysis.
Alexeev, V. P-
l.mericanAntiquitl 56(4) : 604-620.
l99l A Bricf Cultural History. Lecrures given at Harvard Universic¡ transcribed by Hard, R. J., and tù(/. L. Merrill
Geraldine Reinhardt-!?'aller. Chaprer T,Bronze Age in Eurasia, lecrures 7 and 8, 1992 Mobile ,{griculturalists and the Emergence of Sedenrism: Perspecrives from
delivcred July 17 ¿nd 29, I99I, and chaprer 8, Iron Age in Eurasia, delivered Au- Northern Menco.'lmerian Anthropologist n.s. 94(3) : 601-620.
gust 5, 1991. Hiebert, F. T.
Barth, F.
1994 Origins of the Bronze Age Oasis Ciuilízation in Cennal Asl¿. American School o[
1961 Nomads of Soutb Persia.LiccIe,Brown, Boston.
200 -/ Claudra Lhang

Prehistoric Research Bulletin 42.Peebody Museum of Archaeology and Ethnol- 9

og¡ Flarvard Universit¡ Cambridge, Mass'
KelI¡ R. L.
From Atlatl to Bow and Arrow
l99Z Mobility/Sedenrism: conccpts, Archaeological Measures, and Effects. Annual
Reuiew ofÃntbroPology 2l 43-66' Implicating Projectile Technology in Changing Systems
Koster, H. A.
of Hunter-Gatherer Mobility
1977 The Ecology of Pasroralism in Relacion to Changing Patterns of Land Use in the
Northeast Peloponnese. Ph.D. dissercacion- University of Pennsylvania'
Moshkova, M. G. PEI-LIN YU
1992 Stepnaya polosa Asiaxleoi chasti SSSR u skfo-sarmatshoi uremi (Steppe Region of
che Asiatic Part oF rhe SSSR in the Scythe-sarmanti Time). Nauka Izdateltwo,
Moscow.
Nance,J. D., and B. F. Ball In regional culrure histories, the cransition from large, broad-based projectile
L986 No Surprises ?: The Reliabiliry and validity of Test Pit sam pIíng. American Antiq' points ro smaller, lightweight forms is often cited as a shift in the launching
uiry 5r(3)t 457-483. method from spearthrower (or atlatl) to bow and arrow (cf. Aikens and Higuchi
PIog, F. T. 1982 I09; Cabrera Yaldez 1984:279; Grayson 1994:250). Much research on
1973 Diachronic Anrhropology. In Researcb and Theory in cunent Archaeology, edircd the projectile transition has focused on intrinsic attributes of points in order to
by C. L. Redman, pp. l8l-198.John Wiley, NewYork' separare rhem into discrete types (Becdnger and Eerkens 1999:73L; Beck 1998:
Rosen, A- M., C. Chang, and F. P. Grigoriev 2l). In addition ro formal variatio:ì in stone points, the reduction method may
2000 Paleoenvironmenrs and Economy of lron Age Saka-Wusun Agro-pastoraliscs in differ. Analysis of one sample from northeastern North America showed that
Southeascern Kazakhstan',4ntiquity 74: 6ll-623. dart points are typically reduced from cores and arrow poincs from flakes (Nas-
Rouse, I. saney and Pyle 1999:251-Z5Z).
L972 Setrlement Patterns in Archaeology. ln Man, Settlement, and Urbanism, e dited by Projectile point cypes are often used as chronological markers or "guide fos-
P.J. Ucko, R. Tringham, and G.'W'. Dimbleby, pp.95-L07' Duckworth' London' sils" (Huckell 1996:32,6), although variation in form may also result from me-
Shott, M.J. chanically conditioned behaviors, such as breakage, repair, and resharpening
l9g5 Reliability o[Archaeological Records on Cultivated Surfaces: A Michigan Case
(Huckell 1996:327).In a global survey, Pierre Cacelain (L997:232) found that
Scudy.Jounal of Field Archaeology 22(4): 475-490. projectile point form alone is not correlated with hafting contexts or means of
Stuiver, M., et al. launching. Charlotte Beckt (1998) analysis of examples from GatecliffShelter
1998 INTCALS8 Radiocarbon Age Calibration. Radiocarbon 40(3): l04l-1083.
indicates that neck width is acted upon by selective forces and is useful in dis-
'Wandsnider, L., and E. L. Camilli
tinguishing darts from arrow points. Statistical tests for many archaeological
l99Z The Character of Surface Archaeological Deposits and Its Influence on Survey
sequences show that small, Iightweight points replace or augment large, healy,
Accur tcy.Jo urtt a t of Fi e ld A rc h a e o I ogy 19 (2), I 69 - 188.
broad-based points (Shon 1997).
Yablonsþ L. T.
The projectile transition occurred ar different times, and at different rates,
1995 The Marerial Culture of the Saka and Historical Reconstruction, Chapter 14. [n
throughout the world. The transition to bow and arrow never occurred in Aus-
Nornads of the Eurasian steppes in the Earþ lron Ag4 edited byJ. Davis-Kimball,
tralia. Atlatls and bows and arrows were used in tandem in the recent past in
V. A. Bashilov, and L. T. Yablonsky, pp.20L-240- Zinat Press, Berkeley'
the Arctic, the North American Southeast, and parts of Mesoamerica. Recent
efforts to explain the variation in scope and timing of the proje ctile transition
have focused on distinguishing between in situ development and diffusion, es-
pecially in North America (cf. Betdnger and Eerkens 1999; Nassaney and Pyle
3]4o R)L 1999), then proce eding to test models for different modes of transmission. Rob-
S B D/MAË/ IJ,S P