Sally and Alvin V. Shoemaker Chair

Department of Anthropology

University Museum of Archaeology and Anthropology

University of Pennsylvania

Teaching: Courses


Originally excavated by R.E.M. Wheeler in 1947, the archaeological site of Brahmagiri, Karnataka State, India, still defines the basic chronology and ceramic sequence for southern India, and thus constitutes the standard against which all subsequent archaeological constructions are compared.  As a model for south Indian history, however, the Brahmagiri analysis is seriously flawed.  These flaws are evident in terms of ceramic and lithic taxonomy, stratigraphic analysis, statistical treatment of artifact distributions, and incomplete reporting.  Work has begun on the first phase of a two-part restudy of the Brahmagiri artifact collections and excavation notes.  This work will clarify key stratigraphic, chronological, and taxonomic issues and will help set future research on a more secure analytic footing.  The second phase of this project will move the analysis of Brahmagiri beyond stratigraphic and chronological issues to a more nuanced spatial analytical perspective that addresses domestic architecture, ceramic and metal production, and issues of local and long-distance exchange. Reconsideration of the important site of Brahmagiri is a necessary preliminary to a long-term field project that will address issues of urbanization, population aggregation, agricultural intensification, and the establishment of institutionalized religions. 

    When R.E. (later Sir) Mortimer Wheeler was appointed Director-General of the Archaeological Survey of India in 1944, just three years prior to independence, the prehistoric and Early Historic periods of southern India were very poorly understood; little stratigraphic excavation had been carried out and, prior to the development of radiocarbon dating, there existed few ways to securely date most archaeological material.  Accordingly, Wheeler established a program to develop a South Indian chronology based on correlations with the Mediterranean.  At the site of Arikamedu, on the east coast, he found Roman pottery in association with an indigenous ceramic ware known as “Andhra ware” (now referred to as Russet-Coated Painted Ware).   Andhra ware was also known from the inland settlement at Brahmagiri, in Chitradurg District (then in the Mysore State).  Brahmagiri was also attractive because of the presence of an Aśokan rock edict mentioning the town of Isila.  Aśokan edicts date to the third century BC and are associated with the north Indian Mauryan empire.  Both the Roman (a Roman coin was subsequently located at the nearby site of Chandravalli) and the Mauryan “presence” at Brahmagiri presented Wheeler with chronological and taxonomic anchors for the stratigraphic sequence.  In particular, he was interested in dating the enigmatic Megalithic burials of the region. 

    As should be clear, Wheeler’s aims were almost entirely chronological.  He had little interest in elucidating the nature of local economic, political, or social organization, or in discussing the involvement of South Indian peoples with either the Mauryans or with the long-distance exchange networks that linked South Asia to the Mediterranean and island Southeast Asia.  Of the extensive excavations at the Isila “town site,” no plan view and only one stratigraphic section (Br. 21) was published, a section that would come to be seen as iconic for all of South India.  Phase two of this project is designed to address the lack of spatial, functional, technological, and organizational information from Brahmagiri.

    Even given Wheeler’s limited focus, however, his analysis is problematic.  Using the Br. 21 section, Wheeler identified a sequence of three “cultures” that he called the Stone Axe Culture, the Megalithic, and the Andhra culture.  These would now be called Neolithic, Iron Age, and Early Historic.  Conventional dates for these periods (based on limited radiocarbon samples) are ca. 2400-1500 BC, 1000-500 BC, and 500 BC to AD 500.  Although section Br. 21 clearly shows that ceramics associated by Wheeler with these three phases overlap substantially and that the “Megalithic” or Iron Age ceramics (found in Megalithic burials) never occured in isolation, he still viewed Brahmagiri as establishing a clear sequence of successive “cultures.”  In fact, he suggested that changes from period to period were accomplished by waves of migration.  One may question this reconstruction on a number of grounds, including the designation of individual strata, the construction of ceramic types, and the association of types with strata.

    Recent excavation and survey work my colleages and I have conducted in the Bellary and Raichur Districts of Karnataka, about 200 kilometers north of Brahmagiri, has suggested that the “Megalithic” is not, in fact, a distinct time period at all, but is instead a set of burial practices and associated grave goods.  Further, artifactual and other aspects of the Neolithic seem to persist long after they “ought” to disappear, and material assigned to the Early Historic can, in many cases, be more productively seen as evidence of non-local exchange rather than as simply chronological markers.  Clearly, Wheeler’s and others’ understanding of artifact and feature variation as influenced primarily by change through time is inadequate, not only for describing empirical patterns in archaeological material, but also for developing a more contextual view of sociopolitical and economic organization in southern India. 

    Methods of artifact analysis developed since Wheeler’s day could be fruitfully applied to his excavated collections; a contextual rather than simply stratigraphic approach to his excavation notes and plans may yield new understandings of this key settlement.  Brahamagiri, a place depicted by Wheeler as an outpost of the Mauryan empire, might be more usefully conceived as a local center that selectively deployed relationships with local and more far-flung trading partners.  At some point, Brahmagiri developed into a large town, one of the few such places in a region that, prior to the Iron Age or perhaps Early Historic, had never seen a settlement on this scale.  However, the published data do not allow us to say even how large Brahmagiri was at different periods in its history.  The eventual demise of Megalithic mortuary practices and the institutionalization of architecture and textual traditions associated with Hinduism, Buddhism, and Jainism are also problems that can be investigated with archaeological data, but only when those data are exploited for more than their presumed chronological indications. The most efficient and cost-effective way to revise our understanding of Brahmagiri, and hence of  South India between the mid-third millennium BC and the first millennium AD, is by analysis of the Wheeler collection.  That this project will also enhance our understanding of the history of archaeological research is an added inducement to tackle this project.

    This research project is structured in two parts.  The first part, described here, may be considered preliminary in that details of the research design of the second phase depend on both initial results and on the state of curation of the original notes and artifacts.  Wheeler was a very meticulous excavator and may be expected to have left detailed notes.  Artifacts from Brahmagiri are stored in the Purana Quila facility of the Archaeological Survey of India (ASI) in New Delhi.  I have begun working on the collections in Delhi and have begun systematic study of the excavation notes, maps, photographs, and drawings.

    My first objective is to evaluate the key stratigraphic sequence (Br. 21), re-analyzing all of the ceramics, using an attribute-based analysis rather than Wheeler’s typological approach.  As noted, the basic typological categories used by Wheeler are of dubious utility.  Since only type counts by strata are published, reconsideration of types as well as of stratigraphic units could radically change our view of the Brahmagiri sequence.  Attributes coded include; slip color, surface treatment, vessel form, rim diameter, rim form, paint type and color, paste composition, sherd size, thickness, and presence of post-firing modification.  These data will be analyzed statistically and compared with similar material from other archaeological sites. 

    In the second part of the analysis, I plan to work with spatial distributions of artifacts and features as well as stratigraphic relationships. Insofar as possible, I will recreate the excavation of both the town site and the Megalithic burials and re-integrate the new artifact analysis with this spatial reconstruction.  This second phase will involve much larger numbers of artifacts and analysis of more extensive excavated contexts than the first phase.