Social distinctions during the south Indian Neolithic: changing mortuary practices in a late prehistoric cemetery at Maski

Abstract Human burials have been recovered from a wide variety of intra- and extramural settlement contexts at Neolithic period sites (3000–1200 BC) in southern India, yet formal cemeteries remain virtually unknown from this period. Research at MARP-79 in the Raichur District of the south Indian state of Karnataka, near the type-site of Maski, documents a large Neolithic cemetery, now with the largest number of radiometrically dated burials of any archaeological site in southern India. The cemetery demonstrates considerable, previously undocumented variation in mortuary ritual, involving new materials, technologies and burial practices, which challenge culture-historical models, pointing instead towards long-term incremental developments that alter how we understand the emergence of Neolithic social differences.


Introduction
For more than 1000 years, beginning in the mid third millennium BC, agro-pastoral communities buried and memorialised their dead in a sprawling cemetery located in the shadow of a long, low inselberg in the Raichur Doab region of India's South Deccan. The cemetery at MARP-79 was discovered by the Maski Archaeological Research Project (MARP) during systematic archaeological survey of the area surrounding the Neolithic settlement site of Maski, in the southern Indian state of Karnataka ( Figure 1). Archaeological investigations at the site, including surface documentation and excavation, have recorded important patterns of change and continuity in mortuary practices during much of the south Indian Neolithic period (3000-1200 BC) and the early centuries of the Iron Age (1200-300 BC). The research we present here documents considerably more variation in mortuary practices than has typically been associated with the Neolithic period in the South Deccan and attests to the presence of significant mortuary distinctions much earlier than previously recognised in the area (cf. Wheeler 1948;Allchin 1960;Nagaraja Rao & Malhotra 1965;Nagaraja Rao 1984; Devaraj et al. 1995;Walimbe & Paddayya 1999). Moreover, based on the first radiocarbon assays of burials from a Neolithic cemetery in this region, we document both the emergence of a particular suite of burial practices and their gradual and incremental development into the megalithic mortuary practices associated with the south Indian Iron Age.
Earlier research on mortuary traditions in prehistoric south India has largely perceived an abrupt transition from the Neolithic period to the Iron Age. Explanations for this change, framed within culture-historical and socio-evolutionary epistemologies, have favoured theories of population replacement or of a punctuated evolution of social complexity (e.g. Allchin 1960;Leshnik 1974;Sundara 1975;Allchin & Allchin 1982;Mcintosh 1985;Moorti 1994). At MARP-79, we document a more gradual transition in mortuary practices, demonstrating that the source of these changes was in long-standing, localised south Indian Neolithic social practices and cultural activities (Johansen 2019;Bauer & Johansen 2020). Here, we argue that mortuary rituals are assemblages of materials, practices, people and knowledge that converge on the biologically and culturally inflected problem of death, creating localised, historical contexts that provide a venue for a wider politics through which social affiliations and differences may emerge.
By as early as the beginning of the second millennium BC, the diversification in mortuary rites at MARP-79 involved the assembly of novel materials, technologies and practices in a range of contingent configurations that appear to have created meaningful social distinctions, both in how the dead were buried and in how the dead were commemorated by and for the living. Novel materials involving innovations in technology and style (e.g. slipped and polished fine ware ceramic serving vessels and, later, copper and iron objects), together with new burial techniques (i.e. terracotta coffins, combusted organic coffins), commensal practices, and the use of stone as grave architecture, were increasingly mobilised as commemorative assemblages that included people-the living and dead. These changing and diversifying assemblages of mortuary rituals constituted both an opportunity and a venue for the Maski region's Neolithic agro-pastoralists to establish and negotiate relations of affiliation and difference within historically contingent social groups and networks, likely those of kin, locality or other cooperative associations.

The south Indian Neolithic period
The Neolithic period is perhaps the most extensively researched of any in south India's archaeological record. Initially defined in the late nineteenth century as associated with a diverse array of ground and pecked stone tool types distributed across the South Deccan (Foote 1979(Foote [1916), it later became synonymous with Wheeler's (1948) 'Stone Axe Culture', which included a predominantly handmade, coarse fabric ceramic industry most notable for its plain and burnished grey and dull red wares. Wheeler's initial division of the Stone Axe Culture into earlier and later deposits at Brahmagiri was expanded and modified by Allchin (1960) based on work at the site of Piklihal, which remains the most extensive and systematically published analysis of prehistoric south Indian ceramic data. This periodisation was further subdivided and modified by Fuller et al. (2007), who mobilised an expanding corpus of radiocarbon dates, macrobotanical remains and new ideas about site types, settlement patterning and economic activities (see Table 1).
Research on faunal and macrobotanical remains (e.g. Paddayya 2001;Korisettar et al. 2002;Fuller 2006) has identified cattle and sheep/goat pastoralism, together with the farming of regionally domesticated pulses and millets (the 'south Indian crop package'), as important agro-pastoral subsistence practices during the Neolithic period, alongside the use of wild food resources. By the turn of the second millennium BC, macrobotanical assemblages at several sites featured small but gradually increasing quantities of wheat and barley, previously cultivated for centuries in the northern subcontinent, followed by the introduction of African crops and north Indian species by at least 1600 BC (Fuller 2006;Fuller et al. 2007;Boivin et al. 2018; Table 1). These crop introductions appear to correlate with new forms of necked ceramic jars (Fuller 2005), and our findings at MARP-79, discussed below, also include new forms of serving vessels.
The increasing presence of non-locally domesticated crops points to expanding exchange networks and changes in foodways and agricultural practices, all with implications for the diversification of social relations (Fuller 2005;Boivin et al. 2018;Johansen 2019). Further evidence for expanding exchange networks comes from the presence, in small quantities, at Neolithic sites of beads made of carnelian and lapis lazuli, with origins within and beyond the northern subcontinent. This was a time of considerable change across South Asia as many regional exchange networks adjusted to the deurbanisation of the Indus Valley Civilisation (after 2000 BC).

Social distinctions during the south Indian Neolithic
These changes included the concurrent development of copper production, the farming of wheat and barley, and the performance of some mortuary practices similar to those of the north-western Deccan (i.e. the West Deccan Chalcolithic; Dhavalikar 1984;Shinde 1994).
Neolithic communities occupied small village settlements and grazing camps with circular wattle-and-daub houses, and extramural areas set aside for lithic manufacturing, animal penning and butchery, and, in some cases, featuring 'ashmounds'-large mounds of burned and vitrified cattle dung (Allchin 1963;Paddayya 2001Paddayya , 2019Korisettar et al. 2002). Ashmounds appear to have been focal points for periodic communal ritual practices involving the consumption of cattle products, including feasting, as suggested by excavations at Sannarachamma, Hiregudda and Budihal (Boivin et al. 2005;Paddayya 2019).
There are four Neolithic settlement contexts recorded within MARP's 64km 2 project area ( Figure 2). MARP-97 is a multiperiod site excavated by B.K. Thapar (1957), located 1.5km north of the cemetery at MARP-79, with Neolithic settlement and funerary deposits deeply buried beneath Iron Age and Early Historic strata. We have recorded three new Neolithic settlements: MARP-64 is a small settlement (0.3ha) constructed across several adjacent inselberg terraces, 6km north-west of MARP-79; MARP-155 (0.5ha) and MARP-203 (0.3ha) are both likely small hilltop herding camps, with low-density surface distributions of Neolithic lithic and ceramic artefacts, located north and south of MARP-79, respectively.

Neolithic mortuary archaeology in the South Deccan
Neolithic mortuary practices are best known from infant urn burials and relatively unelaborated, extended adult pit graves from intra-and extramural settlement contexts (Allchin 1960;  Nagaraja Rao & Malhotra 1965;Devaraj et al. 1995;Walimbe & Paddayya 1999). The bulk of published data on Neolithic burials in the South Deccan are of sub-adult urn burials, often associated with habitational contexts, although some children were buried in simple pits without any enclosing urn. Sub-adult urn burials contain the remains of one-occasionally two-individuals, with some variation in the number and configuration of ceramic vessels used to contain the body (e.g. single, double, vertical, horizontal, and with bowls or broken jars as lids). Typically, these burials contain no grave goods, but exceptions include Burials 6 and 9 at Budihal (Walimbe & Paddayya 1999). At most Neolithic settlements, adult burials are outnumbered by those of sub-adults. Excavations at several sites document site-based temporal shifts in mortuary practices between occupational phases (Nagaraja Rao & Malhotra 1965;Subrahmanyam et al. 1975;Indian Archaeology: a Review 1983;Devaraj et al. 1995). Burials at these sites are dated by association with surrounding stratified settlement deposits that include changes in the distributions of ceramic wares, and, in some cases, by a limited number of radiometric dates (e.g. at Watgal and Tekkalakota) (Indian Archaeology: a Review 1983; Devaraj et al. 1995). Most documented adult burials are extended pit interments associated with wellknown Neolithic ceramic wares (e.g. plain and burnished grey wares) and forms (e.g. everted rim and spouted jars), either in surrounding deposits, or in small numbers as grave goods. Variation among these burials is limited, yet differences do exist within and between sites, including: the presence of primary and secondary interments; the orientation of the body; the placement of stones above all or part of the body; and the absence or inclusion of ceramic vessels (e.g. 1-3 vessels, especially bowls and spouted jars) and chipped and ground stone tools (Allchin 1960;Nagaraja Rao & Malhotra 1965;Krishna Sastry 1979;Devaraj et al. 1995).
At two Neolithic sites, Tekkalakota and Ramapuram, less-well reported later Neolithic burials overlie earlier graves, documenting incrementally different, as well as novel interment practices (Nagaraja Rao & Malhotra 1965;Indian Archaeology: a Review 1984). At both sites, later graves include greater frequencies of artefacts that were dominated by slipped and polished ware ceramics (e.g. Black-and-Red Ware, and slipped and polished black and red wares); these are more commonly associated with megalithic burial practices and habitational deposits of the succeeding Iron Age. At Ramapuram, interment practices included the use of stone cairn packing, stone slab cists, the burning of coffins made of organic materials and the inclusion of as many as 29 ceramic vessels, including combinations of more 'classic' Neolithic pottery (e.g. grey burnished ware) and slipped and polished wares (e.g. Black-and-Red Ware); iron and copper objects were included in some of the burials (Indian Archaeology: a Review 1983Review , 1984Review , 1985. Similar kinds of mortuary contexts, with ashy, burned coffins and slipped and polished ceramic vessels were also identified by B.K. Thapar (1957) in strata located between Neolithic and Iron Age habitation deposits at Maski (MARP-97; see the online supplementary material (OSM)), though Thapar associated these with the onset of the Iron Age period at the site.
The relative paucity of well-published adult Neolithic burials from this region has led several archaeologists to speculate that most of the adult population were interred in contexts which have largely eluded archaeological detection (e.g. Korisettar et al. 2002); the small Neolithic cemetery at Nagarjunakonda's Site 68 is a notable exception (Subrahmanyam et al. 1975). Thus, while prior to our research at MARP-79, the existence of Neolithic cemeteries has been predicted, and in one case (Nagarjunakonda) investigated, there remains a dearth of research on this important context of Neolithic mortuary practices, and a complete absence of radiocarbon dating of burial sequences.

The MARP-79 cemetery
The cemetery at MARP-79 was first identified in 2012, during MARP's systematic pedestrian survey of the project's study area (Figure 2). Soil and gravel quarrying activities had exposed numerous burials in partial plan and section across several large quarry pits. We recorded 21 partially exposed burials, leading to the prediction that the cemetery may have once held hundreds (Figure 3), some of which may still lie in undisturbed portions of the site. Documentation began by identifying individual burials, typically located in pit wall sections but occasionally in partially destroyed horizontal exposures in the bases of quarry pits. Graves were documented, photographed and, where possible, profile drawings of exposed sections were made, followed by surface collections of cultural materials (e.g. ceramic vessels, stone and iron tools, wood charcoal) and human remains. Three of these burials were excavated in 2018 and 2019.
Thirteen radiocarbon assays on wood charcoal recovered from seven individual graves at MARP-79 (Table 2; Figure 4) document changing Neolithic funerary practices and track Social distinctions during the south Indian Neolithic the historical development of several mortuary assemblages across several centuries. Table 3 describes ten of the better-preserved graves. The radiocarbon record begins with Burial 12 (Figures 4-6) in the third quarter of the third millennium BC, during the Neolithic IB (Tables 1-3). Burial 12, identified in section, offers an example of a relatively unelaborated pit burial with modest grave furnishings-similar to extended pit burials from other excavated Neolithic sites in the region.
By the turn of the second millennium BC, while simple pit burials continued, as evidenced by Burial 7 dated to 1895-1756 cal BC (at 95.4% confidence), new mortuary assemblages had developed at MARP-79. These include terracotta sarcophagi burials, such as Burial 8 (1934-1700 cal BC, at 95.4% confidence), and combusted organic coffin burials containing fragmentary, excarnated skeletal remains. Burial 11, for example, provides four radiocarbon assays dating to the early second millennium BC (Figure 7; Tables 2 & 3). These burials now constitute south India's earliest radiometrically dated contexts with slipped and polished ware serving vessels, which became common in both mortuary and settlement contexts during the succeeding Iron Age (Table 3; Figures 6-8) and which point to possible funerary feasting in the early centuries of the second millennium BC.
The radiocarbon record at MARP-79 resumes between the fifteenth and twelfth centuries BC with assays from Burials 1, 6 and 19, all of which are combusted organic coffin burials (Tables 2 & 3; Figures 4, 5, 6 & 9). These burials demonstrate continuity in the mortuary practices from those first observed in Burial 11 several centuries earlier (i.e. excarnation prior to burial, burned organic coffins, slipped and polished serving vessels and globular slipped jars), with modifications including stone slab elements, iron and copper grave goods, and, in the case of Burial 19, the inclusion of animal remains (e.g. young sheep/goat elements) and greater numbers and types of serving vessels. The ceramic vessels and animal remains from Burial 19 suggest that funerary feasting formed part of the associated mortuary practices. Burial 19 is similar to graves excavated by Thapar (1957) at Maski (MARP-97), which were found between Neolithic and Iron Age habitation deposits (see the OSM), collectively suggesting an expansion of funerary feasting during the Neolithic-Iron Age transition. Terracotta and simple pit burial assemblages continue into the later occupations at MARP-79. Burial 2, for example, contained a terracotta sarcophagus covered by stone slabs (Table 3; Figure 5). The inclusion of an iron blade tool in Burial 2 suggests that the practice of terracotta sarcophagi interments extended into the latter part of the second millennium BC; regionally, iron objects and iron production are documented by at least the fourteenth century BC (Johansen 2014). Terracotta sarcophagi, often with legs and other stylistic elaborations, are documented from burials elsewhere in the Deccan (e.g. Kudatini) Social distinctions during the south Indian Neolithic Combusted ashy organic coffin burial with tabular caping stones. Heavily damaged and exposed in partial section.
Fragmentary remains Slipped and polished ware serving vessels noted.
2 None Terracotta sarcophagus burial dug into saprolite. Sarcophagus covered by stone slabs. Heavily damaged and exposed in partial section.

Skeleton within sarcophagus
Slipped and polished ware serving vessels and an iron blade noted.
3 None Pit burial sealed by a layer of granite boulders. Heavily damaged and exposed in partial section.

Extended skeleton, heavily decomposed
Slipped and polished red ware bowl noted. 6 N = 1 Combusted ashy organic coffin burial. Heavily damaged and exposed in partial section.
Fragmentary remains Slipped and polished ware serving vessels noted; copper bangle; four carnelian beads. 7 N = 1 Simple pit exposed in partial section and damaged.
Not observed. Slipped and unpolished ceramics noted. 8 N = 1 Terracotta sarcophagus burial placed in a pit. Damaged and exposed in partial section.

Skeleton within sarcophagus
Black-and-Red Ware slipped and polished serving vessels noted; dolerite ground stone axe. 11 N = 4 Combusted ashy organic coffin burial with plaster coating. Coffin was interred in an east-west oriented pit dug into saprolite. Burial exposed in excavation.

Excarnated and rearticulated: single adult
Five short-necked, globular, red slipped ware jars; seven slipped and polished ware (Black-and-Red Ware) serving vessels.

N = 3
Longitudinal section exposure of an extended pit inhumation. Pit fill included a layer of ashy matrix capped by broken ceramics.

Fragmentary remains; heavily decomposed
Plain micaceous dark grey ware sherds noted.

N = 1
Combusted ashy organic coffin burial with wood and plaster. Coffin was interred in an east-west oriented pit and capped with tabular stone slabs. Generally well preserved and largely exposed through excavation.

Excarnated and rearticulated; single adult
Four short-necked, globular, red slipped ware jars; 16 slipped and polished Black-and-Red Ware vessels, including dish-on-stand serving vessels; fragment of plain buff ware vessel; medium-sized mammal remains and iron blade tool. 20 None Stone circle cairn burial. Grave placed in a circular arrangement of naturally occurring boulders. Body placed at the base of a pit excavated into the saprolite capped by two granite slabs supporting a cairn of boulders and cobbles. Burial was heavily damaged and exposed through excavation.  ( Mushrif et al. 2003) and date to as late as the Early Historic period (300 BC-AD 500) in other regions of south India (Haricharan et al. 2013). The stone slabs, like those observed in later combusted organic coffin burials, suggest a developing preference for the inclusion of large stone elements, prefacing the large stone constructions of Iron Age megalithic graves. Other burials at MARP-79 featured stone elements in their architecture. Burial 3, an undated primary burial exposed in section, contained a layer of granite boulders sealing a shallow burial pit ( Figure 5). Burial 20 (Figure 10) was another partially preserved grave exposed in section, reminiscent of the more formally constructed stone circle megaliths of the Iron Age and Early Historic periods. It is similar to Thapar's excavated burial Types iv and v (see the OSM), and to a stone circle cairn burial at Terdal that contained diagnostic Neolithic grey ware ceramics and a copper bracelet in an otherwise Iron Age megalithic cemetery (Sundara 1969(Sundara -1970. This suggests a gradual development of burial forms, from the use of stone slabs to cap burial pits and containers towards more elaborate megalithic constructions featuring stone circles, cairns and later dolmens.

Discussion
The developments in mortuary rituals that begin at MARP-79 by at least the Neolithic IIB (Table 1) document a diversification in the ways that regional communities commemorated and memorialised their dead. Here, we approach the social history of mortuary practices at MARP-79, and the South Deccan generally, by considering mortuary rituals and their instantiation in the creation of enduring places (e.g. cemeteries, grave and memorial features), as assemblages of materials, knowledge, practices and people-both living and dead (Johansen & Bauer 2018; Bauer & Johansen 2020). As assemblages, mortuary rituals constitute histories of relationships and interactions between humans and things, in which materials play important roles in social relations organised around death and the disposal of the dead. Indeed, for more than a century, anthropologists and sociologists (e.g. Hertz 1960;Metcalf & Huntington 1991;Fahlander 2020) have acknowledged that mortuary ritual centres around a problem constituted by the material properties of the human corpse: that of decay and decomposition. Yet, it is the confluence of these material and biological processes with culturally inflected practices, values and knowledge systems (e.g. metaphysical logics, eschatological beliefs) that provides the historically situated context for socially meaningful perceptions and experiences of death (cf. Bauer & Kosiba 2016). Hence, while mortuary ritual addresses the physical problem of the human corpse, it also encompasses a historically situated, culturally inflected social and metaphysical problem of death. Through the mediation of cultural practices and social relationships, this provides a venue for a wider politics through which social affiliations and differences may emerge (see Bauer 2015;Johansen & Bauer 2011, 2018. By the twentieth century BC, there is archaeological evidence at MARP-79 for a diversifying set of mortuary rituals that assembled new knowledge, practices and materials, including the earliest known use of slipped and polished fine ware serving vessels and terracotta sarcophagi and, later, copper and iron artefacts ( Figure 6). Taken together, these new assemblages expanded the range of social and symbolic resources involved in attending to the problems Figure 10. Burial 20: partially exposed in section prior to excavation (above), bone pendants in situ (left below), exposed skeletal remains (below centre) and excavated copper bracelet (below right) (figure by the authors).
of death and the corpse, providing a range of material and procedural possibilities for Neolithic communities to commemorate and memorialise their dead at any given time. The inclusion of recently developed ceramic wares and vessel forms, terracotta burial sarcophagi, and constructed and combusted organic mortuary coffins likely involved a growing number of participants, and possibly wider networks of exchange or future obligations with others (e.g. kin, acquaintances, neighbours). Likewise, the raising and slaughter of animals and the provisioning and preparation of other types of foods and beverages for funerary feasts would have provided further opportunities and challenges in negotiations over access to social and symbolic resources. Variation in the residues of these ritual mortuary assemblages suggests that some inequalities of access to productive resources may have developed from the Neolithic IIB onwards.
Yet the archaeological remains of these diversifying mortuary assemblages do not simply mirror or reflect the rank or status of the deceased, as per the epistemological logic of an earlier cultural-evolutionary archaeology (e.g. Binford 1971;Saxe 1971). These new materials and practices appear to have been mobilised in strategic ways into commemorative assemblages that included people, living and dead, with the potential to galvanise and instantiate old and new social relations emanating from the problem presented by the recent dead. It was through these ritual mortuary assemblages that historically contingent social relations of affiliation and difference were negotiated and constituted by and for the living (e.g. Bauer 2015; Bauer & Johansen 2020).
Empirically, these observations also challenge earlier culture-history arguments that suggested Iron Age 'peoples' were migrants or invaders from regions north of peninsular India, who introduced a developed set of megalithic mortuary practices c. 1200 BC (cf. Leshnik 1974;Allchin & Allchin 1982). Our radiocarbon assays of individual burials have allowed us to identify variation in mortuary practices and explore how their localised diversification over time incrementally developed into the more punctuated and often resource-and labourintensive megalithic burial practices of the subsequent Iron Age, when a diversity of megalithic commemorative features are distributed across most of south India (Darsana 2010;Haricharan et al. 2013;Mohanty & Thakuria 2014;Rajan 2015). We have also recorded the earliest dated use of slipped and polished ware serving vessels, centuries earlier than previously documented (Figures 4 & 6-8), further challenging the idea that these wares first accompanied megalith-building migrants from the north. These wares and forms would be used in commensal mortuary ritual for hundreds of years at MARP-79 through the ensuing Neolithic period and into the Iron Age, when their use would expand to quotidian consumption practices in settlement contexts across south India (see Sinopoli 2016).
The early use at MARP-79 of terracotta sarcophagi, burning, excarnation and, later, stone slab capping and cairn packing, are all further elements of mortuary variability observed in later Iron Age and Early Historic period megalithic burial practices, the origins of which we can now conclusively trace to Neolithic ritual mortuary assemblages of the second millennium BC in south India. Less well documented, undated burials from Ramapuram and Tekkalakota (Nagaraja Rao & Malhotra 1965;Indian Archaeology: a Review 1984) suggest that the gradual in situ development of Neolithic mortuary ritual observed at MARP-79 was not an isolated phenomenon. These results together demonstrate that there was not a static and monolithic Neolithic 'culture' that can be easily characterised and conveniently assigned to arbitrarily defined periods-an observation pertinent to archaeologists working within the parameters of regional culture-histories both in South Asia and in other regions of the world.

Conclusion
The cemetery at MARP-79 documents important patterns of change and continuity in mortuary ritual spanning nearly 1500 years, from the Neolithic IB (2500-2200 BC) through to the beginning of the south Indian Iron Age, providing the first radiocarbon assays of burials from a Neolithic cemetery. By the turn of the second millennium BC, we observe, alongside the continuity of simple pit burials with diagnostic Neolithic ceramics, the appearance of mortuary rituals that consisted of at least two additional burial assemblages, including excarnated secondary burials placed in combusted organic coffins, and skeletal remains interred in terracotta sarcophagi. This diversification documents a novel set of social and symbolic resources for Neolithic communities through which a politics of social affiliation and difference could be constructed and negotiated through commemorative memorial practices. Our radiocarbon assays of these burials have enabled us to begin to assess how these assemblages developed over the course of the remaining Neolithic period and into the south Indian Iron Age. Without the radiometric analyses of individual mortuary contexts, the interpretation of the MARP-79 cemetery risked analytic collapse into the temporally flattened 'black box' of culture-historical categories, and a misunderstanding of the gradual diversification of mortuary practices as a punctuated transition between arbitrarily defined periods, misidentified by allegedly temporally diagnostic ceramic types (i.e. slipped and polished wares). Rather than an abrupt transition of south Indian mortuary practices at the start of the Iron Age, MARP-79 demonstrates that change and continuity in funerary rituals were localised and long-standing, unfolding over the course of more than a millennium.