ARCHAEOLOGY INTERNATIONAL

Jainism and local cults.6 These data have Landscape, and in ancient shed new light on inter-religious dynam­ ics, and on the development of exchange Julia Shaw networks between Buddhist and As spread into central and western India fr om its agricultural communities. centre oforigin in the central Gangetic Plain, how did thischange Subsequent field seasons were aimed at the ways in which the landscape was perceived and organized? refining and develop existing results, and testing earlier hypotheses with new meth­ In thisstudy of the regional settingof the great site of Sanchi and odologies. Two seasons concentrated on of other important sites in central and western India, religious, the history of local irrigation through a political, economic and agricultural changes are integrated in an combination of hydrological and archaeo­ holistic approach to archaeological landscapes. logical methods, including the collection of sediments from dams and reservoir beds for optically stimulated luminescence (OSL) his project was initiated in 1998 in the area or how they set up exchange dating and analysis of ancient pollen.7 with the objective of building networks with local populations. Other seasons have focused on systematic an integrated model of reli­ mapping and satellite remote sensing for T gious, economic and environ­ The Sanchi survey: research design testing the degree of concordance between mental history in both central The Sanchi survey was aimed at redressing already-identified archaeological remains and western India through the documen­ some of these problems, through a multi­ in the application area and levels of visi­ tation of key and settlement sites, phase archaeological survey carried out bility within a subset of different satellite and water-resource structures in their over an area of c. 750 km2 around the hill­ imagery.8 The latter produced variable archaeological landscape.1 Key research top complex (Fig. 2). Initial exploration results: in areas where forest clearance has questions include: How did the spread of between 1998 and 2001 resulted in the sys­ taken place, the high-resolution Quickbird the new (especially Buddhism) tematic recording of about 35 Buddhist satellite imagery served as a useful aid to between the third century BC and fifth cen­ sites, 145 settlements,4 17 irrigation dams5 on-site mapping, especially for hard-edged tury AD relate to other key processes such and over 1,000 sculptural and architec­ structures such as Buddhist monuments as urbanization, state formation, economic tural fragments associated with , (Fig. 3).9 However, dense vegetation cover change, and innovations in agriculture? How did incoming religious communities • city belonging to fi rst phase of urbanization (c. 6th century BC) establish themselves in new areas? How did they build up patronage networks with 0 city belonging to second phase of urbanization (c. 3rd century BC) @ Buddhist site site associated with the life or teachings of the local populations? How did they situate historical Buddha in the Gangetic valley themselves within the pre-existing ritual •Charsada e Buddhist site outside the Gangetic valley (c. 3rd century BC) landscape? The project's primary focus is the Sanchi survey study area oTaxila UNESCO World Heritage Buddhist site of Sanchi in central India (Madhya Pradesh), with some comparative investigations at other sites in central and western India such as Ramtek (Maharashtra), Junagadh and Devnimori (Gujarat) (Fig. 1). Sanchi's Kurukshetra• y<'j earliest monumental history dates from • Hastinapura s approximately the third century BC, and Indraprastha• is closely connected with the patronage �<%, . (Delhi) • � "kl Kap1lavastu of the Mauryan empire, whose expanding AtranJl 1era • @Lumbini Mathurao boundaries mirrored, in part, the early 'o Sravasti' @l.Kusinagara movement of Buddhist monks from the 0�_,0 :or>\ Vaisali{jl middle Gangetic plains. Both its distance <:>' G'�'-0 D Pawwaya @Sarnath oPataliputra fromthe cradle of Buddhism and its prox­ ®. Gaya Tumaho ¥f Kausambi Ra)ghat!" @ imity to the early-historic city ofVidisha,2 � •Rajgir c2 Bharhut@ just 6km to the north, made Sanchi a suit­ Vidishao Bodhgaya able case study for examining the socio­ Sanchi Eran . • Ujjain• o Trip ut economic and religious background of Nin a tr o 9t Buddhist propagation. The site is also an ]'lortllod � archetype for a variety of theoretical and A1ah anadi bunagadh �\ methodological problems in South Asian ".,0 B A Y Ajanta• archaeology in general, and the study of Sisupalgarho Buddhism in particular. For example, Bud­ 0 F dhist archaeology has mainly been focused DPratisthana on well known, grandiose monuments, studied in isolation from their surround­ B E N G A L ing landscape and the populations living ARABIA N in them. Thus, despite the large body of art-historical and epigraphical scholar­ S E A N ship at SanchV the site's relationship to other ritual or habitational settlements in the area had not been examined prior to the 0 km 500 current study. This obviously hampered t the understanding of how incoming Bud­ Figure 1 Ancient India, showing key Buddhist and urban centres, and Sanchi survey dhist communities established themselves study area in centre. I 43 _j ARCHAEOLOGY INTERNATIONAL

Figure 2 Th e Sanchi Survey study area, km N land over 440 m showing Buddhist sites, settlements and + -dam dams. Place names are either those of topo­ 0 + Buddhist silo graphical features (such as the hill at And­ her) or the names of the nearest modern + t 0 settlement site 0 villages to the various sites or site clusters. 0 (A nala is a seasonal river or stream.) 0 0 0 + -t across much of the hilly region ruled out 0 0 ' + +. -loB esnagar its effectiveness as either a mapping or 'l>e + primary site-detection tool. Further, in the 0 Vidisha+ 0 0 agricultural zone, sites such as settlement o mounds or dams that showed up clearly in 0 e + 0 the satellite imagery are already high up ao dO 0 + Ferozpur 0 cP the ground-based visibility scale: during Gulgaon 0 0 + 0 Satdhara' \ -I'Dargawan the primary reconnaissance phase, most of • 0 + 0 + these sites stood out as prominent earth­ ..r,8 + Karondih Sanchi + 0 0 � + I 0 works from surrounding hilltops. Finally, 0 0 Naroda 0 Nagauri o- --f 0 0 • many sites that rarely show up in satellite + -ho o imagery, such as hillside settlements, rock­ . 0 0 Dhakna 0 0 1 0 o Devrajpur shelters or springs and natural shrines 0 - o 0 More) + + 0 0 0 (small piles of rocks worshipped as a mani­ 0 0 o.., '-khurd 0 0 + festation of the divine) are more readily 0 + +o 0 detectable using traditional ground-based 0 0 o 0� 8 o 0 0 o o 0 survey methods and, more particularly, 0 Dabar n } 8 Andher+ a a Chandnao+ 0 o 0 through a sensitivity to the currency of #0 MJl.ei 0 � '{" 0 archaeological sites within the present­ of 0 0 r:JKala + day ritual landscape.10 However, remote­ 0 0 + sensing data derived from satellite images have been useful for generating survey con­ 0 tours that, together with sophisticated land­ based survey techniques, have enabled

/\ dam 0 monastery walls --- reservoir boundary

Figure 3 Monastery remains on Ferozpur hill: high-resolution Quickbird satellite imagerywith superimposed GPS-derived mapping data.

44 ARCHAEOLOGY INTERNATIONAL

Buddhist propagation, the state, and patronage networks Although these three processes - religious change, imperial expansionism, and urban­ ization - were evidently linked, there are major problems, which the Sanchi survey sought to redress, concerning the precise nature of this relationship, and its wider ramifications in rural hinterland areas. A major causal factor here is the lack of hori­ zontal excavation at many early-historic city sites, with settlement histories often based on limited vertical sequences, and the traditional site-based focus of South Asian archaeology, which rarely looks beyond individual sites to patterns in the wider archaeological landscape. Further problems stem from outdated models that assume that the distribution of Mauryan edicts represent the extent of a unified political entity.13 There are also major questions surround­ Figure 4 Stupa remains on Andher hill, c. second century BC ing the second major phase of Buddhist propagation in the post-Mauryan period, accurate calculations to be made of the and imperial expansion can be tracked between the late second century BC and the probable original areas and volumes of through the distribution of Ashoka's edicts first century AD. Whatever the role of the ancient reservoirs in the area. carved on pillars and rocks, fromAfghan­ state in the formal establishment of Bud­ istan to southern India;12 many of these, dhism in central India, it is during this later Sanchi and Buddhism in central like the Ashokan pillar at Sanchi, stand period of stupa building that Buddhism India within Buddhist compounds. This is also becomes firmly rooted in the rural land­ Sanchi is one oflndia's best preserved and when urban culture, already prevalent in scape. Inscriptions reveal that this work most studied Buddhist sites, with a con­ the Gangetic valley for several centuries, was funded not, as before, by state patron­ tinuous constructional sequence from the spreads westwards, as is attested by the age but by extensive programmes of col­ third century BC to the twelfth century AD. archaeological sequence at Vidisha and lective patronage supported by powerful Its earliest history was closely connected other central Indian city sites such as Eran, families and guilds.14 By the middle of the with the Mauryan empire, the first pan­ Tumain and Pawwaya (Fig. 1). The earliest first millennium AD, the donations of land Indian polity, whose westwards spread sculpture and architecture associated with and villages recorded in inscriptions indi­ from its epicentre in the Gangetic valley orthodox Hinduism also dates to this cate that the sangha (the collective Bud­ was responsible, in part, for Buddhism's period; Vidisha and other central Indian dhist order) was involved in sustainable transformation from a regional cult into a cities were important centres ofthe Bhaga­ exchange networks with the local agri­ pan-Indian phenomenon. Prior to the third vata cult, an early form of , cultural communities. Many Buddhist century BC, Buddhism was confined to the whose influence reached an apex some­ scholars are reluctant to acknowledge the middle Gangetic Valley, in modern-day what later, by the middle of the first mil­ existence of such links during earlier peri­ Bihar and Uttar Pradesh. Although the pre­ lennium AD. ods, mainly because ofthe traditional view cise dating is still debated, the life and of the ideal as one engaged solely in teachings of the historical Buddha are spiritual, rather than material pursuits. usually placed between the sixth and fi fth centuries BC, with most of our information coming from texts said to have been com­ posed during the lifetime of the Buddha. Although the four pilgrimage sites associ­ ated with the major events in his life have been located (Lumbini, Bodh Gaya, Sar­ and Kusinagara). no distinctly Bud­ dhist remains datable to the pre-Mauryan period have been identified.11 By contrast, many of the major cities in which the ear­ liest Buddhist narratives are set (e.g. Rajgir, Kausambi, Sravasti, Vaisali), have been identified archaeologically, and belong to India's first phase of early-historic urban­ ization (Fig. 1). It is not until the middle of the third century BC, following the Bud­ dhist conversion of the best-known of the Mauryan emperors, Ashoka (who reigned from 273 BC to 236 BC). that Buddhist archaeology appears on the scene, in the form of shrines and stupas (domed struc­ tures built to house relics of the Buddha Figure 5 Th e base of a large monastery at More] khurd, originally surmounted by brick and important followers) (Fig. 4). The over­ and stone superstructure. A heavily damaged stupa is visible in the foreground. Both are lapping processes of Buddhist propagation datable to c. second century BC.

45 ARCHAEOLOGY INTERNATIONAL

Buddhist sites in Vidisha's hinterland.

• Their position within the wider archaeo­ logical landscape warrants their being viewed as part of a cultural package that accompanied the spread of Buddhism, urbanization and the development of centralized state polities during the late centuries BC.

• Similarities with intersite patterns in Sri Lanka, where monastic landlordism is attested from about the second cen­ tury BC onwards, support the suggestion that the Sanchi dams were underlain by a similar system of exchange between Buddhist monks and local agricultural communities. Recent attempts to develop and assess these hypotheses have included detailed surveys based on satellite remote sensing and ground-based mapping (discussed earlier), and a programme of geological Figure 6 An ancient dam at Devrajp ur, with spillwayin foregro und. dating and pollen analysis that focused on two major dam sites: Sanchi and, some Consequently, mendicancy is viewed as to have acted as inundation tanks for 14km to the east, Devrajpur. This con­ the prevailing subsistence strategy, with upstream irrigation, whereas dams built firmed the suitability oflocal sediments for sedentary representing a later across deeper valleys, as found in the east­ optically stimulated luminescence (OSL) deterioration of genuine Buddhist values. ern part of the study area, were used for dating/ as well as affirming our working Closely related to this issue are received downstream irrigation. Some of those in hypothesis that the dams were built, along notions regarding the history and chron­ the latter category, such as Devrajpur, show with the earliest Buddhist monuments in ology of organized monastery architecture, evidence of spillways and sluice gates central India, in the late centuries BC. the prevalent view being that "there are no (Fig. 6). monastery remains clearly datable before Analysis of the surface remains, local The study of irrigation in ancient the Gupta period".15 present-day hydrology and ancient pollen India Archaeological patterns as documented sequences has led to various hypotheses The Sanchi dams have also provided an during the Sanchi survey present a direct regarding the dams' chronology and func­ empirical basis for re-evaluating current challenge to this view, particularly the tion, their associated crop usage, and their models regarding the administrative and identification of various monastery build­ relationship to the urban sequence at Vidi­ management infrastructure of irrigation. ings that evidently belong to the late cen­ sha and the history of Buddhism at Sanchi The development of advanced irrigation turies BC (Fig. 5). The scale and number of and neighbouring sites. These may be sum­ systems in ancient India is usually seen as a these post-Mauryan monastic centres in marized as follows: major factor in the rise of complex urban the Sanchi area predicate an integrated • The earliest dam construction occurred societies. However, some major questions sustainable system of exchange between between about the third and second cen­ regarding the history and chronology of monastic and non-monastic sections of turies BC, coinciding with the earliest irrigation technology and its role in the society, although some monks no doubt monuments at Sanchi and neighbouring wider economic, political and religious continued to beg for a living. Buddhist sites. landscape, remain unanswered. The tradi­

• They were built to provide irrigation, tional view, based mainly on readings of Water management and religious principally for rice, as a response to the problematically dated texts such as the change increased population levels suggested Arthashastra,16 is that the building and Further insights into the economic back­ by the distribution of habitational and management of such irrigation works was ground of Buddhist monastic ism were pro­ vided by a group of 17 dams, all in close proximity to Buddhist sites in the Sanchi area (Fig. 2).5 Surviving as pronounced earthworks up to heights of 6 m, the dams have earthen cores faced with stone slabs, and are laid across valleys up to 1400 m in length (Figs 6, 7). Immediately to the south of Sanchi Hill are the remains of a 350 m­ long dam, which, together with a second dam to the west, would have created a res­ ervoir about 3 km2 in area with a storage capacity of about 3. 6 x 106m3. Two smaller reservoirs at Karondih and Dargawan in the shorter valleys to the west appear to have been designed to maintain water levels in the main reservoir as part of an upstream irrigation system (Fig. 8). Similar reservoirs, with volumes of 0.03-4.7x 106m3, have been found through­ out the Sanchi area. Those built on gradu­ ally sloping terrain, as at Sanchi, appear Figure 7 An ancient dam at Ch andna, reconstructed in recent years.

46 ARCHAEOLOGY INTERNATIONAL

Buddhist settlement /\ctarn

--- reservoir boundary

Figure 8 Quickbird satellite imagery of Sanchi Hill and its archaeological landscape: Buddhist sites, dams and reconstructed reservoir areas. dependent on centralized state administra­ between the sangha and local agricultural tion. Marxist-inspired models, as exempli­ cults. Gupta-period accounts of Chinese fied by Wittfogel's "hydraulic civilizations pilgrims in eastern India show that naga of the Orient", have led to similar notions shrines were propitiated by monks, often regarding Asian economic systems as a within monastic compounds, because of whole.17 These have undergone major revi­ the nagas' ability to ensure adequate rain­ sion following studies of more devolved fall and agricultural success. In the Sanchi systems of irrigation management in Sri area, the positioning of naga shrines on Lanka and Southeast Asia, involving village dams conforms with this model, particu­ councils and religious institutionsY In larly since the sangha appears to have India, however, the traditional model has played a role in the management of local until recently remained unchallenged, in irrigation. Naga was part of Bud­ part because of the paucity of archaeologi­ dhist practice because its effects were in cal research on irrigation. harmony with the sangha's economic con­ cerns with water harvesting and agrarian Buddhism and local agricultural production. This hypothesis, based on the cults relative configuration of dams, settlements Another major focus of the project has been and monasteries in the Sanchi area, as well the temporal and spatial distribution of as similar patterns in western India and Sri over a thousand sculpture fragments across Lanka, forms part of an active model of the study area. Of particular interest is a religious change, which indicates that group of serpent (naga) sculptures datable monks moved into new areas with a set of variously to between about the first cen­ motives for local communities to extend t_ury BC and fifth century AD, and closely their economic support to the monastery. connected with ancient Indian agricul­ tural and fertility cults (Fig. 9).6 In most Conclusion and looking forward cases they are located on top of, or close to The results of nearly eight years of system­ the dams discussed above, and thus pro­ atic and non-systematic survey in the vide latest-possible dates for the construc­ tion ofthe dams. They also shed important Figure 9 A naga sculpture on Nagauri light on theories of religious change, hill overlooking the main Sanchi reservoir especially those regarding the dynamics (location indicated on Fig. 8 above).

47 ARCHAEOLOGY INTERNATIONAL

Sanchi area have provided an empirical 2000); "The archaeological setting of Bud­ & Ruprecht, 1991); and R. Conningharn, basis for re-assessing the economic and dhist monasteries in central India: a sum­ "The archaeology of Buddhism", in environmental implications of religious mary of a multi-phase survey in the Archaeology and world religion, T. Insoll change in ancient India. Steps have also Sanchi area, 1998-2000", in South Asian (ed.), 60-95 (London: Routledge, 2001). Archaeology: proceedings of the 16th 12. F. R. Allchin & K. R. Norman, "Guide to been taken to situate these patterns within international conference of the European the Asokan inscriptions", South Asian a broader regional framework. In 2002, a Association of South Asian Archaeolo­ Studies I, 43-50, 1985. study was conducted at the Sudarsana dam gists, C. Jarrige & V. Lefevre (eds), 665-76 13. For alternative views, see G. ·Fussrnan, site in Junagadh, Gujarat (Fig. 1), where a (Paris: Editions Recherche sur les Civili­ "Central and provincial administration in series of well known inscriptions mention sations, 2005); and Buddhist landscapes ancient India: the problem of the Mauryan the dam's construction and various bouts in central India: Sanchi Hill and archae­ empire", Th e In dian Historical Review14, of repair between the third century BC and ologies of religious and social change, 31-72, 1988; and N. Sugandhi, "Context, the fifth century AD. The dam, which had c. third century BC to fifth century AD content, and composition: questions of intended meaning and the Asokan not hitherto been subjected to archaeologi­ (London: Society for South Asian Studies, the British Academy, in preparation). edicts",Asian Perspectives 42(2), 224-46, cal scrutiny, offered interesting parallels to 5. J. Shaw & J. V. Sutcliffe, "Ancient irriga­ 2003. those at Sanchi, in terms of both its design tion works in the Sanchi area: an archaeo­ 14. V. Dehejia, "Collective and popular bases and its relationship to nearby ritual (both logical and hydrological investigation", of early Buddhist patronage: sacred mon­ Buddhist and Hindu) sites. Plans are now South Asian Studies 17, 55-75, 2001; and uments, 100 BC to AD 250", in Th e powers under way for follow-up investigations at "Water management, patronage networks of art: patronage in Indian culture, Sanchi and Junagadh, as well as the Bud­ and religious change: new evidence from B. Staler-Miller (ed.), 35-46 (Delhi: dhist site of Devnimori in Gujarat, where the Sanchi darncomplex and counterparts Oxford University Press, 1992). similar dam sites are known. Finally, sur­ in Gujarat and Sri Lanka", South Asian 15. The Gupta period is dated to between fourth and sixth centuries AD. vey work will be conducted at Ramtek in Studies 19, 73-104, 2003; "Ancient darns, c. settlement archaeology and Buddhist K. Trainor, "Constructing a Buddhist Maharashtra, which offers interesting par­ propagation in central India: the hydro­ ritual site: stupa and monastery architec­ allels to Sanchi from evidence that water logical background", Hydrological ture", in Un seen presence: the Buddha management played a major role in the Sciences 48(2), 277-9 1, 2003; and and Sanchi, V. Dehejia (ed.), 18-35 administration and land policy of the "Ancient darns and Buddhist landscapes (Murnbai: Marg Publications, 1996). Vakatakas (c. fourth-sixth centuries AD). in the Sanchi area: new evidence on 16. For a critique on the traditional dating of this time within the context of orthodox irrigation, land use and rnonasticisrn in this text, see Kautilya and the Arthasastra: Brahmanical religion rather than Bud­ central India", South Asian Studies 21, a statistical investigation of the author­ dhism.19 1-24, 2005. ship and evolution of the text, T. Traut­ 6. J. Shaw, "Naga sculptures in Sanchi's rnann (Leiden: Brill, 1971). Notes archaeological landscape: Buddhism, 17. Oriental despotism : a comparative study 1. Project website www.britac.ac.uk/insti­ Vaisnavisrnand local agricultural cults in of total power, K. A. Wittfogel (New tutes/SSAS/projects/sanchi.htm. I would central India, first century BCE to fifth cen­ Haven: Yale University Press, 1957). like to thank the Archaeological Survey of tury CE", Artibus Asiae LXIV(1), 5-59, 18. R. A. H. L. Gunawardana, "Irrigation and India, and the Madhya Pradesh Directo­ 2004. hydraulic society in early medieval Cey­ rate of Archaeology, Archives and Muse­ 7. OSL refers to the levels of luminescence lon", Past and Present 53, 3-27, 1971. ums (Bhopal) for supporting this project, emitted on exposure to light as the result 19. This component of the project is being with special thanks to S. B. Ota, 0. P. of released energy accumulated in crystal­ carried out in collaboration with Hans Misra, K. K. Mohammed, and S. K. Verma line minerals through the action of ioniz­ Bakker (University ofGroningen) who has for help during fieldwork. Since 2003, ing radiation fr om natural (background) already published extensively on aspects research has been funded by the Society radioactivity. When a sediment is of Vakataka religious and political his­ for South Asian Studies, the British Acad­ exposed to sunlight, the OSL acquired over tory: Th e Vakatakas: an essay in Hin du emy Board for Academy-Sponsored Insti­ geological time is removed, and the lumi­ iconology, H. T. Bakker (Groningen: tutes and Societies (BASIS), and Merton nescence clock is thus set to zero. If the Egbert Forsten, 1997). College, Oxford. Recent fieldwork was sediment is then re buried in, for example, carried out in collaboration with the Mad­ the body of an ancient darn,an OSL signal hya Pradesh Directorate of Archaeology, will accumulate over time in the crystal­ Archives and Museums (Bhopal], with line minerals at a steady and measurable additional input by John V. Sutcliffe, rate, which provides the basis for dating Anthony Beck (Durham University) and the time elapsed since the previous expo­ Lindsay Lloyd-Smith (University of sure to light; for further information about Cambridge). Thanks to Meera and Ishwar the OSL dating technique see: Dass in Bhopal; Dharmacaitanya and www.rlaha.ox.ac.uk/lurnin/lumindx.htrnl Pushpa Jain, and Ashutosh Vyas in Vidi­ For applications of this method (also of sha for many kindnesses. pollen studies) in this project, see J. Shaw, 2. The Vidisha Research Group (supported J. V. Sutcliffe , L. Lloyd-Srnith, by the Society for South Asian Studies), J-L. Schwenninger, M. S. Chauhan, with which the current project is closely with contributions by E. Harvey & connected, has other studies in the Vidi­ 0. P. Misra, "Ancient irrigation and Bud­ sha area: dhist history in central India: optically www.britac.ac.uk/institutes/SSAS/ stimulated luminescence and pollen groups/vidisha.htm. sequences from the Sanchi darns", Asian 3. Th e Bhilsa tapes, A. Cunningham (Lon­ Perspectives 46(1), in press. don: Smith Elder, 1854); Th e monuments 8. J. Beck & J. Shaw, "The application of sat­ of Sanchi, J. H. Marshal!, A. Foucher, ellite imagery to the Vidisha Research N. G. Majumdar (Calcutta: the Survey of Project, Vidisha, India", in preparation. India Offices, 1940). 9. The offset of approximately 20 m between 4. J. Shaw, "Sanchi and its archaeological the satellite imagery and site plan in Fig­ landscape: Buddhist monasteries, settle­ ure 3 is a cumulative effect of inbuilt GPS ments and irrigation works in central errors and those related to satellite India", Antiquity 74, 775-6, 2000; "The imagery projections. sacred landscape", in Buddhist reliquar­ 10. See n. 4: Buddhist landscapes in central ies fr om ancient India, M. D. Willis (ed.), In dia, J. Shaw. l- 11. Th e dating of the historical Buddha, 38 [London Bdti'h Mu""'m '"'" H. Bechert (ed.) (Gottingen: Venderhoech - !

48