13th July 2013
H D Sankalia’s 1969 advice on how Indian archaeology could progress
This is another remarkable writeup by Sankalia. Worth a thorough read.
Problems in Indian Archaeology, and Methods and Techniques Adopted to Tackle Them
by H. D. Sankalia (World Archaeology, Vol. 1, No. 1, Recent Work and New Approaches (Jun., 1969), pp.29-40
Archaeology in India has a multiple role to play. Not only has it to unravel or reveal the country's long prehistoric past, which lies mostly unknown, and wherever known uninterpreted, but it has also to be a handmaid of, and corrective to, a past with a long tradition of unwritten literature. This literature, known as Vedic and Puranic, continued to be handed down orally even in early historic times. Some of it, particularly the bardic (Puranic and epic), naturally got inflated. Thirdly, archaeology has to show how far the pre- and protohistoric cultures can be related to the preliterates who continue to survive in many parts of India.
The Deccan College Research Institute, since its inception, has had two main aims. First, to reconstruct the past with the help of archaeology in all its aspects, including inscriptions. Second, to link history with prehistory through protohistory.
Before I939 scholars made some attempt to study the Puranas and epics. Dynastic lists as well as accounts of culture – customs, manners, dress, ornaments, and such aspects of material and spiritual culture – had been prepared. These accounts relied heavily on internal evidence only. To this extent, these works were uncritical. For it was forgotten or never realized that this literature, except the Vedic, was often interpolated, and so the epics and Puranas did not refer to one particular culture or age, but to several cultures or ages.
Some more objective standard was necessary to measure or judge the accounts – particularly the span and time factor of events, objects and personalities mentioned in the Puranas. Critical editions of various old texts were not enough. What was needed was hypercriticism.
Archaeology though not so well developed some thirty years ago was the only discipline which could supply this norm. The VJyu Purana was then believed to be the oldest Purana according to internal evidence. This Purana was selected for such a critical treatment. Its contents were minutely analysed, card-indexed, and then regrouped into various topics. These were then compared on the one hand with well-dated archaeological objects, such as coins and sculptures, and on the other with equally well-marked stages of literature. It was thus possible to stratify the Vayu Purana with the help of archaeology (Patil 1946: I4).
(i) The Archaic Survivals, which are coeval with the similar material found in the Vedic Literature, 500 B.C., may be safely considered to be the lowest time-limit of this material.
(ii) The Ancient Material of the Purana aligning itself with the early Dharmasutras, the early Buddhist and Jain canonical literature, the Arthas-astra of Kautilya, the Manu Smriti, and the earlier portions of the Great Epic Mahibhirata. Broadly speaking, the beginning of the Christian era may be supposed to be the lowest chronological terminus of the material.
(iii) The Accretions or the mass of material that has been incorporated into the body of the text. Most of this material falls in line with the Smriti works, the later portions of the Great Epic, etc. It should be noted that this material does not generally cross the chronological line demarcated by the date A.D. 500.
It might also be mentioned that before undertaking this work the lower limit of the various incidents mentioned in the Puran.as was fixed by classifying the contents of the Gupta inscriptions, which happen to be well dated individually and collectively (Patil 1940: 148-65).
The Mahabharata, one of our two great epics, was similarly approached. Out of its numerous aspects, only those referring to the description of gods and goddesses (iconography) were selected for study. This work, entitled Iconographical Elements in the Mahabharata (Chapekar 1958: 232), takes us very much further than the critical text of the epic prepared by the Bhandarkar Oriental Institute at Poona. Chapekar was able to demarcate three strata or phases in the Mahabharata.
The pre-epic phase is mostly derived from the Vedic and the post-Vedic literature; the first phase covering the early period of the Mahabhdrata composition may be brought down to 300 B.C., and the second immediately follows it and sometimes may overlap it; the third begins approximately in the first century A.D. and may extend to the Gupta- Chalukyan period. (c. A.D. 300-700).
When ancient places or sites referred to in the Puranas and epics are scientifically excavated it becomes possible to say on the strength of the stratigraphical data when certain events in the Puranas could have taken place. In fact we are able to visualize how and when the stories in this bardic literature got inflated from time to time. Thus our excavations at Dwarka and Tripuri, though on a limited scale, not only proved the antiquity of the former site, but showed that the traditions about its submergence in the sea were well founded, though not so old as popularly believed (Sankalia i964a, i966a). The second excavation showed how the city of Tripuri was founded in the Early Iron Age in the sixth to seventh century B.C., how it became Shaiva in the third to fourth century A.D., displacing Buddhism and Jainism and how these incidents were represented in purely Shaiva Puranas in the seventh to eight century (Sankalia I966b).
Likewise, the account of the foundation of Takshasila and Pushkalavati given in the Ramayana refers not to the earliest foundation of these cities as believed by Wheeler (I962: I5), but to the towns laid out later in chessboard pattern by the Indo-Greeks in the first to second century B.C. (Dani I967: 25). To this extent the Ramayana account is late. Here again archaeology affords striking proof of the traditional view and the textual critics who had maintained that the Baalakanda of the Ramayana was composed later.
Inscriptions engraved in the 2,000-year-old Brahmi script and its later variants on stone and copper plates also serve as important evidence. Hitherto these records have been primarily used to reconstruct dynastic histories and to throw light on a few aspects of culture, such as religious tendencies and economic conditions. But they were rarely used for reconstructing the ancient territorial and administrative units or for understanding the composition and growth of ethnic elements in the ancient Indian society. Since many of these inscriptions are land grants to Brahmanas and are dated in the then current calendars, these records provide an invaluable source for such socio-political or administrative studies.
Realizing the importance of these inscriptions, the Deccan College has during the last twenty-five years worked out the territorial units in Gujarat, Maharashtra, Northern Mysore, Madhya Pradesh, parts of Uttar Pradesh, Rajasthan and Madras, ranging over a period of I,6oo years (c. 300 B.C.-A.D. 1300). So far the studies on Gujarat are published (Sankalia 1949), those on Maharashtra and Madhya Pradesh are partly so (Sankalia I967b). Large wall maps (12 ft x 8 ft) in the archaeological museum of the college also depict these reconstructed units against the physiographic background of Gujarat, Maharashtra and Mysore. These studies further provide good data for understanding the composition of the Brahmanas and other classes of society, as well as for understanding the changing customs of naming a child owing to the impact of new cultural forces. For it should be remembered that in India personal names were significant and were indeed a good index to a person's caste as well as religion. Inscriptional names are again of great linguistic significance (Sankalia 1965b).
In the field of pure archaeology the aim has been more and more precision in the methods of acquiring the data as well as their interpretation. This may be illustrated by a few examples.
During the last twenty-five years, mainly on the initiative of Deccan College, relics of the Early Stone Age have been found almost everywhere in Central and Western India and Orissa, whereas systematic collections have been made from several districts of Andhra and Mysore (Sankalia I962-3). These surveys have left little doubt of the wide distribution of the earliest food-gathering people who used, among other stone tools, handaxes and cleavers. Uncertainty exists, however, regarding the formation of the deposits in which these tools were found, and hence about the age of the tools and about the environment in which Early Man lived.
For an understanding of these problems it may be recalled that as far as the Himalayan foot-hills and the Potwar plateau of Western Panjab are concerned, we know from the work of the Yale-Cambridge expedition that the Stone Age tool-bearing periglacial or non-glacial deposits can be directly linked up with the glacier and inter-glacial deposits of the lower Himalayas. These studies have not only dated the various lithic industries on sound geological principles but have also well established the changing environment of North-west India during the pleistocene. As opposed to this, in Peninsular India the artefacts normally occur in a coarse gravel that rests on the basal rock or clay and is capped by a thick deposit of silty sand. The Narmada has three such gravel and silt deposits, while the other rivers have only two, the later ones usually exhibiting finer gravel and a different coloured silt. As a rule the artefacts are found in the gravels and rarely, if at all, in the overlying silt.
As a result of the observations at Nevasa (Sankalia I956), it was ascertained that the finer middle gravels yielded a stone industry different from that found in the basal and in the top loose gravels. These industries were initially designated as Series 1[, Series II, and Series III respectively, and are now assigned to Early, Middle and Late Stone Ages. Geologists and prehistorians (De Terra and Paterson I939: 300, Paterson and Drummond I962: I7 and Zeuner I950: ii and I963: 14) attribute the deposition of the coarse gravels to a wet climate followed by a drier phase, first causing erosion and later aggradation. Two, three, or four such cycles of wet and dry climates were thus postulated. Our systematic surveys of the Narmada and other rivers in Madhya Pradesh (Joshi et al. I965), the Banas and its tributaries in Rajasthan (Misra i967), the Goadvari and Pravara (Corvinus-Karve and Rajaguru I967), the Krishna (Pappu I966), Mula-Mutha (Mujumdar and Rajaguru I966a: I3I) in Maharashtra, and the Rallakalva (Murty I966) in Andhra have, however, shown that such deposits can be traced up to or within a few miles of the source of almost all the peninsular rivers.
We have therefore begun to question the simple theory of erosion and aggradation due to a wet and dry climatic cycle. While eustatic causes might be responsible for aggradation and erosion in the lower reaches, as for instance in the Narmada (Wainright I964), what brought about such uniform formation of deposits in the middle and upper reaches of the peninsular rivers? To understand these problems my colleague Shri Sr. N. Rajaguru has undertaken a geomorphological study of the Mula-Mutha, whereas Dr Corvinus-Karve has been entrusted with a similar study of the Godavari-Pravara, and a joint expedition of the Archaeological Survey of India and the Deccan College has been working on the Narmada.
From the morphometric analysis of gravels of the Mula-Mutha, Rajaguru has been able to throw some light on the depositional environments of the gravels during the late pleistocene, and his analysis of the associated soils and silts will help elucidate the problem of fluvial sedimentation in the past. Both Rajaguru and Corvinus Karve after their preliminary geological studies in the Godavari-Krishna valleys in Maharashtra have come to one important conclusion: that this part of India, so far supposed to be tectonically stable, had experienced tectonic disturbances of a moderate nature throughout the pleistocene. The main focus of such activities was in the Godavari basin, because of which a heavier sedimentation has occurred in the Godavari as compared to the Krishna. They are now trying to find out the exact role played by the pleistocene climatic changes in the history of these rivers.
Similar detailed studies have been planned around Maheshwar (Madhya Pradesh) where, unlike Hoshangabad-Narsingahpur, the stratigraphy is clearer. Preliminary petrological study of the gravel components in the pleistocene alluvium of the Narmada at Maheshwar has shown that the river in the upper pleistocene times had a rmuch wider channel, about twice the present channel, and the pleistocene deposits are more of alluvial origin than the colluvial ones.
Associated with an understanding of the formation of these early gravel deposits is the assessment of the nature and age of the contained stone industry. So far the collections of artefacts have been small and haphazard, hence incapable of statistical treatment. Various theories are therefore current. First, at Mahadeo Piparia on the Narmada one witnesses a development of the handaxe-cleaver industry from a true Oldowan-type pebble industry (Khatri I966: 98). Secondly, it has been shown that the handaxe in- dustry itself shows two phases – earlier Abbevillian and later Acheulian (De Terra and Paterson I939: 322; Sen and Ghosh I963: i6).
We tested these theories by excavating at Mahadeo Piparia, and the result has been most unexpected (Supekar I965 and I967). Though the 110 ft cliff is composed of horizontally bedded layers and there is no apparent sign of any disturbance, still the pebbly or bouldery gravel, at a depth of I5 ft from the talus surface, showed a mixture of three different lithic industries: pebble tools and large flakes; ovates, handaxes and cleavers, and between 2:z and 8o % of tools that are today classed as Middle Stone Age (fig. 7). Thus Supekar's small but careful excavation on the Narmada is indeed an eye-opener. Not only is there no pure-Oldowan or Mahadevian pebble tool industry in the basal gravel, but there is also a combination of pebble tool, advanced handaxe and the Middle Stone Age industries. Thus, surface observations and so-called in situ collections from river gravels are suspect.
Our other quest has been the search for the home of Early Man, and particularly for his skeletal remains. The first step is to discover primary sites, sites where man would live even temporarily and make tools from the easily available raw material. Of three such sites, Lalitpur in Madhya Pradesh discovered by Singh (Singh I965), Anagwadi in Mysore discovered by Pappu (Pappu I966), and Chirki-Nevasa discovered by Corvinus-Karve, the evidence from the last has been most convincing. Here within an area of 20 ft x I5 ft, 700 tools all in mint condition, made from local coarse reddish basalt as well as dyke basalt, have been found as they were made and left by their makers, along with the debris (Corvinus-Karve I967 and plate 5). Only a part of this 'camp' has been exposed, and it is hoped that further work in another season will lay bare the remains of man himself.
Another regional survey, by Dr M. L. K. Murty (I966) around Renigunta in Andhra Pradesh, revealed the existence of a blade and burin industry comparable in many respects to the Upper Palaeolithic of Europe. Further intensive work has yielded some 600 tools, all stratigraphically and typologically belonging to a period between the Middle and Late Stone Ages (plate 6). Excavations have therefore been planned at the site which, it is hoped, will give some other evidence about the blade-and-burin culture.
While the large Nevasa collection is under study, Dr Misra (I967) and Dr Pappu (I966) have given us an up-to-date statistical account of the Early, Middle, and Late Stone Age industries found in the Berach basin in South-west Rajasthan and the Krishna basin in Karnatak or Northern Mysore respectively. Dr Misra has employed Chi Square and 't' tests to measure inter-industry and inter-culture variability in terms of the frequencies and measurements of various tool types in the Early, Middle, and Late Stone Ages in the Berach basin in South Rajasthan. The use of statistical methods on a bigger scale, will, it is hoped, help in a better understanding of the dynamics of culture change during the Stone Age.
Our excavations at traditionally historic and prehistoric (protohistoric) sites like Nasik, Nevasa, Maheshwar, and Ahar (Udaipur) have not only shown that all these sites are old (of course, probably not so old as the orthodox scholars and laymen believed them to be), but that every one of them represents a distinct group of cultures which flourished in Maharashtra, Madhya Pradesh and Rajasthan between c. I800-T700 B.C. and I000 B.C. These excavations, preceded and followed by planned explorations, have helped fill on the one hand the gap between the Early Historic and protohistoric, and on the other hand between the prehistoric (Late Stone Age) and the protohistoric. We are now engaged in finding out the identity of the authors of these prehistoric cultures. Out of the several possibilities two or three are most likely, as suggested by the clues offered by a few distinctive pottery shapes (Allchin I96I) and designs as well as shapes (Sankalia i964b), and by the study of human skeletal remains (Ehrhardt and Kennedy i965; Kennedy and Malhotra I966; Nagaraja Rao and Malhotra i965). The first is that these Chalcolithic cultures were in part at least, if not wholly, brought about by the migration of food-producing communities or their wares, or by the diffusion of ideas from Western Asia. The other theory is that the aboriginal inhabitants of Mysore and Maharashtra (and also Madhya Pradesh and Rajasthan?) were self-supporting pastoral cum-food-producing people, each producing a specialized type of pottery, but otherwise not much different economically and perhaps racially. These people, who do not seem to be racially much different from the Bhils and other aboriginals or preliterates, were probably driven out to the hills and forests by the later Sanskrit-speaking peoples.
To substantiate these theories, much more evidence is necessary than is available at present. It is also essential to carry out archaeo-ethnographic studies of the preliterates who still live in or around the excavated sites. To achieve the first objective we excavated horizontally Navdatoli, a site opposite Maheshwar, then Ahar, Tekkalkota and Sangankal, the last two in the heart of the Polished Stone Axe cultures. Later Miss Nagar studied the Bhils and others residing around Ahar (Nagar I966) and Malhotra studied the Boyas (not published). The most positive gain has been the understanding of the Neolithic way of life in Andhra-Mysore (Sankalia I964c) and the Chalcolithic and Copper Age settlement pattern in Madhya Pradesh and Rajasthan (Sankalia i967a: 26z; Sankalia, Deo and Ansari i967-8). On these firm bases we shall be able to proceed.
Another problem that was (and is still) occupying the mind of scholars in India was whether there was a hiatus between the Chalcolithic and the Early Historic cultures or between the former and the preceding Stone Age culture. Wheeler's (I948) work at Brahmagiri showed that there was an overlap between the Polished Stone Age and the Megalithic culture. However, our excavations at Nevasa (Sankalia et al. 1960) District Ahmednagar revealed the existence of a weathered soil horizon between the Chalcolithic and the Early Historic. This observation was later supported by the late Professor Zeuner when a much larger area at Nevasa was exposed for inspection.
The only way to confirm these observations was to analyse chemically the entire excavated profile by taking samples at every xo cm. This was done by our archaeological chemist Dr G. G. Mujumdar. His analysis has shown that the weathered layer is a true natural soil developed on the debris of the Chalcolithic habitation. As the formation of such a soil requires an uninterrupted period of at least some hundreds of years, it also proves a definite time gap between the Chalcolithic and the Satavahana culture at Nevasa. Further, the comparison of this weathered layer with the virgin black soil occurring at the base of the Chalcolithic habitation, indicates a deterioration in the climate proved the hiatus at Nevasa, Dr Mujumdar has now enlarged his field, and soil samples from several Chalcolithic-Neolithic sites in Maharashtra and Madhya Pradesh and Mysore have been studied. He has made a comparative study of the black (fossil or buried) soils on which the Chalcolithic-Neolithic people came and settled and the modern soils occurring in these areas. This study has enabled him to postulate better conditions, both as regards climate and vegetation, prevailing in the Chalcolithic- Neolithic period. These fossil soils were found to be richer and more fertile than their modern counterparts and were suited for a better growth of cereals (grasses), which can be taken as an indirect evidence of food-production. These conditions must have subsequently deteriorated to yield the modern soils and vegetation.
Similarly Mujumdar and Rajaguru's study of the weathered or buried soil layer in an ashmound at Kupgal, District Bellary, Mysore has shown the existence of a wet climatic phase between the occupation of the Neolithic, or the ashmound, people and the earlier stone-using people (Mujumdar and Rajaguru 1966b).
The microlithic site at Langhnaj, first dug by the writer in I94I-2 and subsequently worked on for several seasons (Sankalia I965a), did not yield any datable artefacts from the lower layers, with which were associated many of the human skeletons, animal bones and microliths. Nor is it possible to place this entire complex in a true stratigraphical context, because no more than two layers are visible to the naked eye and the one detected by the late Professor Zeuner (namely, a buried soil layer indicating a humid climatic phase within the main arid phase), is of no practical value. For unless we 'see' the layer, it cannot be dug out. As a last resort, in the absence of charcoal, it was decided to obtain a relative date of the entire 6 ft deposit by analysing the fluorine, nitrogen and uranium content of bones found at various levels. Currently Dr K. P. Oakley is engaged on this work, while the few charred bones have been sent to the Tata Institute of Fundamental Research for C14 dating.
Our large-scale horizontal excavations of the Chalcolithic site at Navdatoli on the Narmada in Madhya Pradesh had yielded some 23,000 chalcedony blades and cores, after rejecting 7,000 pieces on the site. This vast collection contains cores in various forms, parallel-sided blades (used and unused), parallel-sided blades with one end slightly curved and pointed and hence hereafter described as 'penknife' 'blades, serrated blades, a few trapezes, lunates and awls. While excellent statistical accounts could be prepared on this material, it was thought desirable further to ascertain the uses of the various types of artefacts by performing various functions such as cutting wood, vegetables, stalks of grains and shaving; also to relate the artefacts to the excavated remains of each structure, and then to ascertain whether the latter was for human habitation or storage. Further, by plotting the tool types stratigraphically as well as horizontally – as far as practicable relating to the exposed structures – it has become possible to find out the needs of each house in its period. Such a detailed plotting revealed that the lunates, though very few, were required by every house at each of the four phases ranging from I500 B.C. to 1200 B.C. Nor were these mere survivals from the preceding Late Stone Age period. To answer the problem what the lunates and trapezes were required for and why their number was so small compared to the parallel-sided blades (55 %), an examination of the edges of all the blades (including lunates and trapezes) was undertaken. This showed that unlike the parallel-sided and penknife blades, very few lunates had their
38 H. D. Sankalia edges used. It was therefore inferred that the lunates and trapezes served as barbs of harpoons and arrow-heads in fishing and hunting. Unlike the parallel-sided blades, which were used in the home, these would normally be lost once an arrow was shot and hence irretrievable. Their number would thus get smaller and smaller, while that of the parallel-sided blade would remain constant. This and similar inferences enabled us to say something more positive on the socio-economic condition of the inhabitants of Navdatoli from a study of the blade industry alone (Sankalia i967: 262-8). Pottery and other finds from horizontal excavations could be similarly studied. Instead of having a large pottery section, there should be one devoted to every 'house' of each phase or period. We would thus know how many vessels, and of what type, each house required, and from that the probable size of the family might be inferred. Such a study for a site like Navdatoli would be extremely rewarding.
Our knowledge of metal technology is still very imperfect. When, therefore, in our excavations at Ahar, Udaipur, Rajasthan copper slag as well as copper tools were found, a systematic examination of the entire problem was undertaken. Samples of copper ore from Khetri, Rajasthan, as well as portions of a copper axe and sheet from Period 1, Phase b, of the Ahar excavation, were spectrographically and chemically studied by Dr K. K. Hegde of the M.S. University of Baroda (Sankalia, Deo and Ansari i967-8). These studies enabled us to say that Ahar was a copper-smelting area. This fact, coupled with the total absence of lithic blade industry from the site, further indicates that copper smelting was probably a major economic activity of the inhabitants. Hegde's study also tells us that the copper axes were cast in a crude unventilated sand, or earthy, mould and were left in a cast condition.
In archaeology of the historic periods, among several problems the one that has not been systematically tackled is the one which relates to the early temples, their style of architecture and the images which decorate their walls. For the last fifty years the latter have been used for preparing general corpuses on iconography, and identifications have been attempted very often with the help of texts of a much later period. This is anachronistic to say the least. However, the wider problem is whether the temple and the attendant images follow any particular text or school or whether the latter were written afterwards and embodied in the existing style of architecture and iconography. A group of well-dated temples belonging to a well-defined geographical area was selected for intensive analysis, and the results were compared with the texts believed to have been composed in the period. This study has shown that from the iconographic point of view, the correspondence is rather limited. However this is just the beginning, and a further probe is now in progress regarding another group of temples (Mankodi 1966).