Research

The early medieval town of Dzhankent (Kazahkstan): from initial hypothesis to new model

by Heinrich Härke (heinrich.haerke@uni-tuebingen.de), Irina A. Arzhantseva (arzhantseva@rambler.ru) and Azilkhan Tazhekeev (azik8484@mail.ru)

Dzhankent (or Jankent) is a deserted early medieval town in the delta of the river Syr-Darya (the Jaxartes of Classical Antiquity). Its pahsa (adobe) walls enclose a broadly rectangular area of 16 hectares; a prominent citadel takes up the northwestern corner of the town, but otherwise there are no remains of buildings or other structures visible above ground (Fig. 1). The site has attracted the attention of military topographers, antiquaries and historians since the 18th century, not least because it has long been assumed to be the town of Yengi-kent mentioned by 10th century Arab geographers as the seat of the yabgu (khan) of the Turkic Oguz nomad empire (Agadzhanov 1969; Bartol’d 1963). Early fieldwork since the 19th century did not produce much in the way of tangible results, with the exception of the sporadic work by members of the Khorezmian Expedition of the Academy of Sciences of the USSR between the 1940s and 1960s (Arzhantseva and Gorshenina 2018). They took an aerial photograph (1946), produced a topographic plan (1963; Fig. 2), and collected surface finds which confirmed a 10th century date for settlement at this place, but also hinted at one or more earlier phases (Classical period and/or from 7th century AD; Tolstov 1947, 1962). More systematic excavations within the walls of Dzhankent were started by Russian and Kazakh institutions in 2005, focused on a couple of trenches dug without scientific dating, although pedological research was undertaken as early as 2006 (Kurmankulov et al. 2007).

Fig. 1 View of Dzhankent from southeast (drone photo, Martin Goffriller 2018)

Fig. 2 Plan of Dzhankent (Khorezmian Expedition 1963) with hypothetical identification of features

A new international project with local expertise, wider cooperation and geoarchaeological methodology began here in 2011; this was reported at the time in TEA (Arzhantseva et al. 2012). The explicit starting hypothesis was based on the written sources and on the state of archaeological knowledge in 2010: The arrival of the nomadic Oguz on the steppes north of the delta in the late 8th century AD led to their control over local sedentary societies, resulting in state formation; this, in turn, resulted by the 10th century in the urbanization of the previously non-urbanized region, meeting the economic, administrative and socio-political needs of the incipient steppe state; the urban model was imported from the south via trade links to Khorezm (ancient Chorasmia, medieval Khwarazm), the Iranian river oasis civilization on the river Amu-Darya (Oxus); Dzhankent benefitted economically from its location on the interface of sedentary civilizations and steppe nomads, and its population mix (as reflected in pottery finds) shows that; the town was abandoned in the 11th century for environmental (shift of river channel) or political reasons (disintegration of the Oguz state). While it would be an overstatement to call it a ‘Hodges Theory’ (Hodges 1982) adapted to the East, elements of our working hypothesis were certainly informed by early medieval debates since the 1980s on the links between urbanization, trade and state formation in western and northwestern Europe.

A decade of fieldwork and new insights in the course of the project have meanwhile led to substantial modifications of the initial hypothesis. Several carefully placed trenches have revealed small houses of Central Asian (Khorezmian) type, with sequences of up to four superimposed phases in the 10th century alone (Arzhantseva and Tazhekeev 2014). A fragment of decorative wall painting was uncovered in the same part of the town, suggesting a wealthier element in the population outside the citadel; but large representative buildings have not been found so far (by the end of the 2019 season), neither in the lower town nor in the citadel. The corner citadel wall was of Khorezmian type; and the 10th century southern town wall, at the point of its characteristic ‘kink’, was found to stand on top of an occupation layer with 8th century pottery, implying the existence of an open settlement before the building of the walled town. A trench in the lower town (shahristan) has been taken down to natural, producing sherds of Afrigid pottery of 6th/7th century date from Khorezm. Finds from stratified contexts have confirmed the identification of three main pottery styles originally suggested by stray finds: a regional tradition rooted in earlier sedentary societies (Dzhety-Asar Culture); a hand-made, heavily decorated style thought to belong to the nomad element of the population (Oguz); and wheel-turned wares from Khorezm. Special finds include a bronze bowl from a house inside the west gate, opposite the citadel; a 10th-century pot with three chicken eggs bearing Arabic lettering, found against the outside of the northern wall of the citadel, i.e. outside the walled town area altogether; and about half a dozen vessels with Arabic graffiti from several trenches. The latter finds, together with Samanid coins of the 10th century and Khorezmian pottery types, bear witness to the links of Dzhankent to the, by then Islamic, civilization to the south and southeast.

Key evidence for our rethinking of the initial hypothesis has been supplied by the application of geoarchaeological methods and scientific dating. Non-invasive prospection (magnetometry, electric resistivity, electrotomography, georadar; directed by I. Modin, Moscow) has supplied an almost complete plan of the final phase(s) of Dzhankent (10th-11th centuries), showing a dense arrangement of houses in most parts of the interior, with a regular chequerboard arrangement in large parts of the western half, suggesting centrally planned construction (Fig. 3). Electrotomography has helped our understanding of the stratigraphy of the citadel, showing it to be placed on top of a natural hillock or earlier tell, not (as frequently in Central Asia) on a clay platform.

Fig. 3 Geophysical prospection results at Dzhankent (as of 2018)

The geomorphologist on the project (A. Panin, Moscow) initiated a change in fieldwork strategy after we had realized the sheer depth of occupation layers across the site. In 2019, he laid a grid of two dozen coring points across the site and used a lorry-based mechanical drill to obtain cores from these points, through a depth 3 to 8 metres (depending on location) of cultural layers down to natural (Fig. 4). More than 100 samples from these cores are currently being radiocarbon-dated (CEZA, Mannheim). These dates, together with the pedological analysis of the cores, will provide us with an outline history of occupation of the interior and of selected points outside. Initial observations during coring have already changed a number of earlier hypotheses and tentative identifications of features (see Fig. 2): a large complex against the inside of the northeastern town wall is not an urban ‘estate’, but a massive clay platform with a smaller, rectangular building on top; the bowl-like depression in the southeastern quarter is not contemporary with the town, but apparently a reservoir created in the 20th century for rice cultivation then practised next to the monument; the hump outside the east gate is not the remains of a ‘caravanserai’, but an island in a prehistoric river channel (OSL-dated); the ‘workshop area’ north of the town, suspected there after initial geophysical prospection, does not appear to have seen any occupation; and the ‘inlet’ about 5 km south of Dzhankent is not a proper bay of the Aral Sea affording waterborne access to the town, but probably just a series of shallow lakes in the Syr-Darya delta.

Fig. 4 Mechanical coring in the interior of Dzhankent in 2019

Soil scientists (M. Bronnikova and A. Ivanova, Moscow) involved in this project have also found evidence of a wet environment (today the delta area around Dzhankent is almost completely dried out) and of local grain cultivation (inferred from spelts). One of their most intriguing findings was that of keratin-eating microfungi in the empty northern annexe of the town, suggesting it as a place where animals or humans had been penned up (with the latter being more likely because of the low levels of phosphorus; Ivanova et al. 2014). One of the archaeozoologists contributing to the project (A. Haruda, Halle-Wittenberg) has identified, among the masses of animal bones from the site, most of the skeleton of a cat which interdisciplinary teamwork has now shown to be the earliest domestic cat of the region, C14-dated to the late 8th century (Haruda et al. 2020). As cats belong to urban, or at least sedentary, environments, not to nomad contexts, this spectacular find implies that the cat was brought to Dzhankent via trade links from the south or southeast; and it suggests that the predecessor settlement of the walled town already had a (proto-)urban character in this early phase.

Radiocarbon dating has supplied some of the most important evidence since the beginnings of the project. Starting in 2011, several series of dates have put the beginnings of settlement at this place in the delta back to the 7th or even late 6th century AD, some three centuries before its first mention in written sources. These dates confirm a perceptive hypothesis by Levina (1971) who had suggested this early date on the basis of regional pottery types among the surface finds from Dzhankent, but without any supporting dating evidence being available to her at the time. Most importantly, the new dates place the origins of Dzhankent in a rather different historical context: in the time before the appearance of the Oguz on the steppes north of the Syr-Darya (historically dated to the late 8th century AD; Golden 1992), and close to the beginnings of the Northern Silk Road operating along the Syr-Darya (from late 6th century; Kovalev 2005). There are so far no C14 dates later than the 11th century; a few glazed pottery sherds from the uppermost levels may imply some limited continuity into the 12th century, or a temporary re-occupation in the 14th century when the Golden Horde town of Montobe was built nearby.

Analytical work is continuing, and another fieldwork season is planned for 2021. But it is already possible (and indeed, necessary) to revise the initial working hypothesis about Dzhankent because it has been falsified by new evidence. The distinct pre-Oguz phase necessitates an explanation of its origins independent of the presence of the nomads. Even in the upper layers of the site, possible evidence for nomads is scant, limited to a few belt fittings (some of them Early Turkic) and weapon parts, and a pottery style tentatively ascribed to the Oguz. Within the town, there do not seem to be any open spaces for seasonal nomad habitations (although these may have been pitched outside the walls); and the citadel lacks the monumental buildings known elsewhere from nomad central places and Central Asian state capitals (Rogers et al. 2005; Siméon 2010). While the date of the walled phase of Dzhankent (from late 9th/early 10th to late 11th/ 12th centuries) would still coincide with the presence of Oguz nomads in these parts, the historical record need not indicate true state formation within the Oguz realm (pers. comm. P. Golden, Rutgers University, and A. Khazanov, University of Wisconsin; Kradin 2014 and pers.comm.), contrary to the preferred interpretation in Kazakh historiography and archaeology (e.g. Bajpakov et al. 2012).

Here, then, is our new provisional model of Dzhankent. The open predecessor settlement (Dzankent I) was founded as an undefended fishing and farming village around AD 600 by a displaced regional population during the final phase and collapse of the Dzhety-Asar Culture further upriver. The settlement, located as it was on the newly operative northern branch of the Silk Road, and close to the river Syr-Darya as well as the coast of the Aral Sea (as it was then), attracted the attention of Khorezmian merchants who founded a trading post here. By the late 8th century when the Oguz nomads arrived in the region, the settlement was already of proto-urban character. When occupying the delta (an ideal pasture for nomads), the Oguz elite probably became aware of the potential of the location and, at some point in the 9th century, instructed Khorezmian architects and builders to reconfigure the native settlement with its trading post into a walled town with winter quarters for the Oguz ruler (Dzhankent II). It is probably not coincidental that the first evidence for a lively trade from the Baltic via the Volga to Central Asia dates to the 9th century, indicated by the start of the ‘silver flow’ of dirham coins to Scandinavia (Jankowiak 2017; Kilger 2008). Dzhankent thrived for over a century on the crossroads of the two corridors, operating as a transshipping point (land – river – sea) for the east-west trade along the Northern Silk Road, and as a transit point for the north-south slave trade (Fig. 5). When the latter faltered in the later 10th century (judging by the weakening coin flow into Scandinavia), the ‘boom town’ (we owe this label to C. Kilger, Gotland) of Dzhankent declined and was eventually abandoned with the disintegration of the Oguz polity (mid-11th century), like the two neighbouring, less well-explored sites of Kesken-Kuyuk-Kala and Bolshaya Kuyuk-Kala.

This new model provides pointers to several features which remain to be tested against further evidence and will also be addressed in the final fieldwork season of the current funding cycle, in particular features relevant to the urban character of the settlement, the ‘slave-pen’ function of the annexe, and the nomad elite presence. It is also a starting point for the interpretation of this curious cluster of early medieval towns in the Syr-Darya delta, a previously non-urbanized region far away from contemporaneous towns, by putting it in the wider context of trade and urbanization in Eurasia at this time. The change of the model over the course of the project is also an illustration of the potential of interdisciplinary geoarchaeological approaches, and of the constant need to adjust working hypotheses and fieldwork strategies of ongoing projects.

Fig. 5 Map of Central Asia with location of Dzhankent and Silk Road (dotted line; after R. Girshman)

Acknowledgements

The beginnings of the Dzhankent Project in 2011 were supported by an International Collaborative Research Grant from the Wenner Gren Foundation (ICRG-104). The current project phase has been funded since 2017 by the Deutsche Forschungsgemeinschaft (DFG), project no. 389351859 (http://gepris.dfg.de/gepris/projekt/389351859) while based at the Universität Tübingen (Germany; https://uni-tuebingen.de/de/169047). We are grateful to these foundations and institutions for their significant financial and organisational support of our work, and to Korkyt-Ata State University of Kyzylorda (Kazakhstan) for additional institutional assistance. Part of the research for the Dzhankent Project has been carried out within the research plan of the Institute of Ethnology and Anthropology, Russian Academy of Sciences, Moscow (Russian Federation).

References

Agadzhanov, S. G., 1969. Ocherki istorii oguzov i turkmen Srednei Asii IX-XIII vekov. Ashkhabad: Ylym. Arzhantseva, I. and S. Gorshenina, 2018. The patrimonial project of Dzhankent: constructing a national symbol in the longue durée. Ancient Civilizations from Scythia to Siberia 24: 467-532.
Arzhantseva, I.A. and A.A. Tazhekeev, 2014. Kompleksnie issledovaniya gorodishcha Dzhankent. Raboty 2011-2014 gg. Almaty: Arys.
Arzhantseva, I. A., M.S. Karamanova, S.A. Ruzanova, A.A. Tazhekeev, I.N. Modin and H. Härke, 2012. Early medieval urbanization and state formation east of the Aral Sea: Fieldwork and international workshop 2011 in Kazakhstan. The European Archaeologist, 37: 14-20.
Bajpakov, K.M., D.A. Voyakin and R.V. Ilin, 2012. The towns of Huwara and Yangikent, the old and new capitals of the Oghuz state. Bulletin of the International Institute for Central Asian Studies 16: 22-44.
Bartol’d, V.V., 1963. Oguzÿ (Turkmenÿ) do obrazovaniya sel’dzhukskoï imperii. In V.V. Bartol’d, Sochineniya, vol. 2: 559-573. Moscow: Nauka. Golden, P., 1992. An Introduction to the History of the Turkic Peoples. Wiesbaden: Harassowitz.
Haruda, A.F., A.R. Ventresca Miller, J.L.A. Paijmans, A. Barlow, A. Tazhekeyev, S. Bilalov, Y. Hesse, M. Preick, T. King, R. Thomas, H. Härke and I. Arzhantseva, 2020. The earliest domestic cat on the Silk Road. Scientific Reports 10: 11241. https://www.nature.com/articles/s41598-020-67798-6#citeas
Hodges, R., 1982. Dark Age Economics: The Origins of Towns and Trade AD 600 – 1000. London: Duckworth.
Ivanova, A.E., M.A. Bronnikova and A.A. Gol’eva, 2014. Kharakteristika selitebno-khozyajstvennoj deyatl’nosti na territorii gorodishcha Dzhankent po geokhimicheskim, mikrobiomorfnym i mikologicheskim dannym. In I.A. Arzhantseva and A.A. Tazhekeev, Kompleksnie issledovaniya gorodishcha Dzhankent. Raboty 2011-2014 gg.: 141-160. Almaty: Arys.
Jankowiak, M., 2017. What does the slave trade in the Saqaliba tell us about Early Islamic slavery? International Journal of Middle East Studies 49: 169-172. Kilger, C., 2008.
Kaupang from afar: Aspects of the interpretation of dirham finds in Northern and Eastern Europe between the late 8th and early 10th centuries. In D. Skre (ed.), Means of Exchange: Dealing with Silver in the Viking Age. Kaupang Excavation Project Publication Series, Vol. 2 – Norske Oldfunn XXIII: 199-252. Aarhus: Aarhus University Press.
Kovalev, R.K., 2005. Commerce and caravan routes along the Northern Silk Road (sixth to ninth centuries). Part 1: The western sector. Archivum Eurasiae Medii Aevi 14: 55-105.
Kradin, N.N., 2014. Nomads of Inner Asia in Transition. Moskau: URSS.
Kurmankulov, J.K., I.A. Arzhantseva, E.D. Zilivinskaya, S.A. Ruzanova and J.T. Sydykova 2007. Materialy Dzhankentskoj arkheologicheskoj ekspeditsii: Arkheologicheskie raboty na gorodishche Dzhankent v 2006. Kyzylorda: Inzhy-Marzhan.
Levina, L.M., 1971. Keramika nizhnei i srednei Syrdarya v pervom tysyacheletii n.e. Trudy Khorezmskoi Arkheologo-Etnograficheskoi ekspeditsii, VII. Moscow: Vostochnaya Literatura.
Rogers, J.D., U. Erdenebat and M. Gallon, 2005. Urban centres and the emergence of empires in Eastern Inner Asia. Antiquity 79: 801-818.
Siméon, P., 2010. Hulbuk: architecture and material culture of the capital of the Banijurids in central Asia (ninth–eleventh centuries). Muqarnas: An Annual on the Visual Cultures of the Islamic World 29: 385-421.
Tolstov, S.P., 1947. Goroda Oguzov. Sovetskaya Etnografiya 1947 No. 3: 57-102.
Tolstov, S.P., 1962. Po drevnim del’tam Oksa i Yaksarta. Moscow: Vostochnaja Literatura.

Authors

Heinrich Härke, Honorarprofessor, Universität Tübingen, Germany; Professor, Higher School of Economics University, Moscow, Russia; Visiting Research Fellow, University of Reading, Great Britain. heinrich.haerke@uni-tuebingen.de

Irina A. Arzhantseva, Principal Researcher, Institute of Ethnology & Anthropology, Russian Academy of Sciences, Moscow, Russia; Associate Professor, Higher School of Economics University, Moscow, Russia. arzhantseva@rambler.ru

Azilkhan Tazhekeev, Director, Research Centre for Archaeology and Ethnography, Korkyt-Ata State University of Kyzylorda, Kazakhstan. azik8484@mail.ru

Go back to top