Social Classification and Its Historical Formation

Formation Social hierarchy in early medieval Mithila, Vajji and Aṅga was real, but its categories were historical rather than timeless. Land, office, ritual authority, occupation, migration, household standing and stigma generated overlapping distinctions. This chapter reconstructs how those distinctions became socially legible and, in some settings, increasingly hereditary. It refuses two opposite shortcuts: treating normative varṇa as a complete census, and projecting later caste lists backward into centuries for which they are not documented. Contemporary charters from Munger and Bhagalpur provide the principal regional anchors; wider eastern Indian evidence supplies comparison; later Maithil genealogy is used only with explicit chronological controls. 29.1 Social classification as a historical problem Every society classifies people, but the categories used by rulers, ritual specialists, households and later ethnographers are not identical. Early medieval Mithila, Vajji and Aṅga were marked by inequalities of landholding, service, ritual prestige, occupation, dependency and political power. Some of these differences were expressed through Sanskrit terms that later became associated with caste. Others were recorded as offices, household ranks, ethnonyms or local status labels. The task is therefore not to ask when a complete modern-style caste system suddenly appeared. It is to reconstruct how social distinctions became more explicit, more hereditary in some settings, and more closely connected with marriage, land, ritual authority and occupational reproduction. This chapter treats classification as a process. It uses contemporary charters as its strongest regional evidence, normative texts as evidence for ideals, and later genealogical or caste records only after applying a strict chronological filter. 29.2 Varṇa and jāti must be kept analytically distinct Varṇa supplied a broad normative language of ranked social order. Jāti, in historical usage, could refer to birth, kind, community or more bounded social groups, and its meanings changed across genres and periods. Treating the four-varṇa scheme as a literal census obscures the much denser social vocabulary of inscriptions. Conversely, translating every occupational or regional label as a fully formed caste imports later closure into earlier evidence. The early medieval record contains Brahmans, officials, soldiers, cultivators, rural notables, service personnel and groups placed low in Brahmanical rankings, but these categories operate at different scales. Varṇa is most useful for understanding normative ranking; jāti formation is better approached through evidence of repeated group names, marriage boundaries, inherited occupation, collective rights and durable social reproduction. Those forms do not appear at the same speed everywhere, which is why regional chronology matters. 29.3 Four evidence streams and their limits Figure 112 separates the evidence into charters, normative texts, material history and later records. Charters are dated transactions and therefore provide the strongest control for categories made administratively visible. Normative texts tell us how learned traditions imagined social order, but prescription does not equal practice. Archaeology reveals unequal settlement, access to goods, specialised production and institutional demand, yet it rarely names the social identity of the people who made or used objects. Later genealogies, caste handbooks and colonial descriptions can document identities that had become important by their own time, but they are dangerous when projected backwards. The chapter therefore asks four 313313 GAJENDRA THAKUR questions of every label: when is it attested, who used it, for what purpose, and what social relation does it actually demonstrate? Only after these questions are answered is comparison across periods legitimate. 29.4 The Munger plate: classification in an administrative act Devapāla’s Munger plate of year 33, issued from Mudgagiri, is unusually informative because its notification moves through a long chain of office-holders and then addresses residents through an explicitly broad social range. The charter names royal and provincial officers concerned with administration, coercion, tolls, forts, transport and animals. It then turns to local society, including Brahmans, mahattaras and kuṭumbins, and extends the formula to Medas, Andhrakas and Caṇḍālas. This is not a neutral ethnographic list. Its purpose is to announce a grant and command compliance. Yet precisely because the grant had to be socially effective, the formula shows categories that the issuing administration considered intelligible. It also shows that political authority imagined the village social field as differentiated rather than homogeneous. What remains uncertain is the population, wealth, occupation and marriage practice of each category named. 29.5 The Bhagalpur plate: hierarchy, service and local society Nārāyaṇapāla’s Bhagalpur plate of year 17, also issued from Mudgagiri and concerning a village in Tīrabhukti, repeats the broad logic while adding a rich roster of officials and servants. It addresses royal personnel, provincial and village officers, transport-related staff, cāṭas, bhaṭas and sevakas, then turns to residents extending from Brahmans through locally prominent categories to Meda, Andhra and Caṇḍāla. The same charter separately instructs resident cultivators to render the assigned revenues to the beneficiary. This produces a layered documentary picture: political office, service, local standing, cultivation and ranked social categories overlap without becoming identical. The plate is particularly valuable for Mithila because the granted village lay in Tīrabhukti. It demonstrates a classificatory language operating within the region, while leaving open how rigidly its categories structured everyday marriage and occupation. 29.6 From Brahman to Caṇḍāla: what the formula does and does not prove The phraseology stretching from Brahmans toward groups placed at the lower end of the social scale communicates comprehensiveness and hierarchy. It tells the reader that the order binds the whole inhabited community, not only elites. It also reflects a Brahmanical vocabulary in which some groups carried stigma. But a rhetorical range is not a population table. The formula does not tell us what proportion of a village belonged to any category, whether all named groups lived in every granted settlement, or how local people described themselves. Nor does it establish a single occupational definition for Meda, Andhra or Caṇḍāla. The secure inference is narrower and more important: royal documentary practice recognised a graded social field and expected that field to be administratively intelligible. The historian must resist turning that recognition into a finished caste diagram. 29.7 Mahattara and kuṭumbin: status without automatic caste identity Mahattara and kuṭumbin are among the most useful labels for reconstructing rural differentiation because they point toward social standing without fitting neatly into a varṇa chart. Mahattaras appear as locally important or senior persons; kuṭumbins are substantial householders in many eastern Indian inscriptions. Their authority could derive from land, household resources, participation in boundary knowledge, or local mediation. These positions could become socially reproducible, yet the terms are not equivalent to modern caste names. Their presence reminds us that hierarchy was generated not only by ritual classification but also by agrarian resources and local influence. A family controlling land, cattle, labour and HISTORY OF MITHILA, VAJJI & ANGA — VOLUME II information could exercise power whether or not a surviving text gives it a fixed jāti label. Social formation therefore involved the interaction of economic position and classificatory language, not the simple application of a pre-existing fourfold schema. 29.8 Cultivator is an economic relation before it is a caste label The Bhagalpur plate’s resident kṣetrakaras matter because they were responsible for producing the revenues transferred by the grant. The word identifies cultivators more securely than it identifies a closed hereditary community. Cultivation rights could coexist with differences in tenancy, household scale, labour dependence and status. Some cultivators may have been substantial householders; others may have worked land under more dependent conditions. Over time, repeated agrarian roles could contribute to inherited social identities, but the charter itself does not prove that all cultivators belonged to one bounded caste. This distinction is essential in a region where later agrarian communities developed strong caste identities. Historical method requires starting from the relation documented in the ninth-century charter and tracing later consolidation forward, rather than using a modern agricultural caste list to interpret an earlier kṣetrakara automatically. 29.9 Offices, professions and social heredity Administrative titles could become associated with families because literacy, court access and technical knowledge were often transmitted through households and patronage networks. Yet an office is not automatically a caste. The Pāla rosters distinguish ministers, military commanders, record-keepers, revenue and police staff, frontier officials, messengers and transport supervisors. Some of these roles may have drawn repeatedly from particular families or social networks. Others could have been appointments dependent on royal favour. The evidence is strongest when a title recurs across generations in identifiable kin groups; it is weaker when only the occupational designation survives. This is especially important for later identities connected with scribal service. The route from professional literacy to a hereditary corporate community is historically plausible, but its timing must be demonstrated. Early medieval charters reveal a specialised service world; they do not by themselves yield a complete genealogy of later service castes. 29.10 Figure 112: why a caste history needs multiple archives Figure 112 summarises the chapter’s method. A charter can show that a category mattered in an administrative transaction; a normative text can show that learned authors ranked or theorised it; archaeology can show the material inequalities within which people lived; a later genealogy can show that a corporate identity became important at a later date. The strongest historical argument arises where these streams overlap without being collapsed. If a later genealogy supplies a lineage name that never appears in earlier inscriptions, the absence matters. If a charter names a group but later records reorganise it under another label, the change matters. Caste history is therefore not the recovery of an ancient master list. It is the history of how different institutions repeatedly named, ranked, reproduced and contested social boundaries. 315315 GAJENDRA THAKUR Figure 112 — Social classification: distinct evidence streams 29.11 Agrarian expansion created new contexts for ranking Chapter 26 showed that land grants altered rights over revenue, resources and settlement. These changes also affected social classification. When new or reorganised agrarian zones were granted to religious beneficiaries, the state had to identify local authorities, cultivators and people liable to obey the transfer. Brahman settlement networks could introduce or strengthen Sanskritic status vocabularies, while local household hierarchies continued to shape actual production. This interaction could generate new forms of ranking. A community that controlled land or mediation might gain recognition; another linked to stigmatised labour could be placed lower in normative schemes. Yet agrarian expansion was not a uniform programme of caste creation. The balance among migrant beneficiaries, established residents, local elites and dependent workers differed by place. Social classification should therefore be read as one dimension of agrarian reorganisation rather than as a separate religious superstructure imposed unchanged on every village. 29.12 Brahman settlement networks and the construction of authority Ryosuke Furui’s work on early medieval Bengal emphasises that Brahman authority itself had a history. Brahmans established networks through migration, land grants, ritual expertise and ties to kingship; their clearer corporate identity was produced through these processes rather than simply present outside history. The same caution is useful for Mithila and Aṅga. Grants to Brahman or religious beneficiaries could create durable centres of learning, ritual and property management. Such centres helped circulate genealogical, gotra and normative vocabularies. But the regional record must decide when particular corporate forms became important. A ninth-century charter identifying Brahmans does not automatically prove the full institutional structure of later Maithil Brahman communities. The historical question is how authority became local, reproducible and connected with land, marriage and political patronage over successive centuries. 29.13 Political service could generate rank as well as income Service to court and administration offered more than wages or perquisites. It created proximity to rulers, access to records, mobility and opportunities for landholding. Titles such as rājanaka, amātya, mahāsāmanta HISTORY OF MITHILA, VAJJI & ANGA — VOLUME II or record-office positions expressed different levels of political incorporation. Families able to reproduce such service might convert office into social prestige. At the same time, service categories such as cāṭa, bhaṭa and sevaka could denote much more dependent relationships. The state therefore produced its own hierarchy, intersecting with ritual ranking but not reducible to it. A person of high political office could derive status from command and patronage; a Brahman beneficiary could derive status from ritual and textual authority; a substantial householder could derive it from land and local influence. Early medieval social order was formed where these sources of rank met, reinforced one another or came into tension. 29.14 Religious institutions were also social institutions Buddhist monasteries, Śaiva establishments, temples and Brahmanical centres organised patronage, labour and access to resources. Their inscriptions classified donors, teachers, beneficiaries and service relationships in institution-specific ways. The Bhagalpur charter, for example, supports a Śaiva establishment while being issued by a Pāla ruler within a wider Buddhist political tradition. This plural institutional world complicates any notion that one religious ideology mechanically produced all social categories. Religious institutions could reinforce hierarchy through rules of eligibility and purity, but they could also create alternative networks based on monastic rank, initiation, donation or learned authority. Social classification was therefore plural. A person could simultaneously occupy a household position, occupational role, ritual category and institutional identity. Later caste boundaries sometimes gathered these dimensions together, but the earlier record often preserves them separately. 29.15 Ethnonyms and origin labels in the charter world Pāla charters include names such as Gauḍa, Mālava, Khaśa, Hūṇa, Karṇāṭa and Lāṭa among military or service personnel. These labels remind us that royal establishments were mobile and multi-regional. They should not be forced into a caste taxonomy. Some refer most naturally to geographic or ethnic origin; some may have become occupationally stereotyped in particular contexts; some could acquire new social meanings after migration. The key point is that early medieval classification included origin and political service alongside ritual status. A person’s social identity could be shaped by where a group was thought to come from as well as by what it did locally. This mobility challenges models in which every community is imagined as immobile within a fixed village hierarchy. Migration was itself one of the forces through which new identities could be formed and later naturalised. 29.16 Meda, Andhra and Caṇḍāla: naming the social margins Groups placed at the lower end of charter address formulas deserve attention because elite historiography has often rendered them invisible. Their inclusion shows that subordinated communities were part of the productive and governed landscape. Yet the labels are difficult to interpret. Andhra may function as an origin term in some contexts; Meda has different associations across texts; Caṇḍāla carries strong stigma in Brahmanical normative traditions. The charter does not provide occupational biographies. It tells us instead how the issuing authority represented the social range of its subjects. The historian should therefore recover two facts at once: real inequality and uncertain local semantics. Stigma in elite language is evidence of hierarchy, but it cannot substitute for the voices, livelihoods or self-identifications of the people named. Classification from above must be read as power, not simply description. 317317 GAJENDRA THAKUR 29.17 Untouchability requires evidence, not retrospective assumption Brahmanical texts developed increasingly elaborate rules concerning purity, contact and groups regarded as impure. These traditions are indispensable for the history of exclusion, but a textual rule is not proof that every locality enforced the same boundary in the same way. Early medieval inscriptions can demonstrate subordinated social placement; they rarely describe the daily mechanics of avoidance, residential segregation, access to water or commensality in enough detail to reconstruct a uniform regime. Archaeology also seldom identifies caste directly. The safest conclusion is that graded notions of purity and stigma were historically powerful and that some named groups were placed at severe social disadvantage, while the exact practices of exclusion must be documented where possible. This avoids the opposite errors of denying hierarchy because evidence is incomplete and projecting a fully documented nineteenth-century untouchability regime unchanged into the ninth century. 29.18 Household reproduction made inequality durable Classifications become socially powerful when they are reproduced through households. Land, skills, ritual knowledge, debt, patronage and social reputation can all pass between generations. Marriage links households into wider networks and can either reinforce or cross social boundaries. Early medieval charters rarely record these processes directly for ordinary residents, but the logic of household property is visible in agrarian society, while genealogical claims become increasingly important in later periods. The central point is not that every category was already endogamous. It is that household reproduction supplied a mechanism through which unequal positions could become durable. A locally powerful kuṭumbin family could transmit resources; a craft household could transmit skill; a Brahman lineage could transmit textual and ritual capital; a dependent household could inherit vulnerability. Jāti formation is best understood against this broader history of social reproduction. 29.19 Gender is part of classification even when women are unnamed Social ranking was reproduced through marriage, inheritance, labour and ritual eligibility, all of which were gendered. Yet royal charters overwhelmingly name male office-holders and beneficiaries, making women less visible than their social importance warrants. Normative texts discuss women extensively but often from the standpoint of male household authority. The historian must therefore distinguish documented rules from actual practice. Marriage restrictions are especially important because later caste systems made endogamy central, but the date at which specific regional marriage circles hardened must be demonstrated. Women’s property rights, widowhood, dowry-like transfers and household labour also varied by legal and social context. Chapter 32 examines women, marriage and property directly. Here the methodological point is narrower: any history of caste formation that ignores the gendered reproduction of households misses one of the principal mechanisms through which social boundaries became durable. 29.20 Figure 113: the charter field is not a pyramid Figure 113 deliberately avoids drawing a caste pyramid. The Munger and Bhagalpur formulas bring together royal service, local standing and a ranked social range in the same administrative act. These domains overlap, but the charter does not tell us that every official occupied a fixed ritual rank above every cultivator, or that each named group formed a closed endogamous unit. A pyramid would therefore create more certainty than the source permits. The diagram instead represents multiple fields of classification whose members were all required to recognise the grant. This is historically significant because it shows that HISTORY OF MITHILA, VAJJI & ANGA — VOLUME II hierarchy had several institutional bases. A person’s position in political service, rural property, ritual status and local community could interact without being reducible to one universal scale. Figure 113 — A Pāla charter addresses a social field, not a census 29.21 Markets and towns complicated village classification Chapter 27 showed that villages, markets, ferries, towns and institutions formed connected networks. Mobility through those networks could loosen as well as reinforce social boundaries. Merchants dealt across settlements; transport workers crossed jurisdictions; artisans served patrons beyond one village; officials moved with courts and camps; students and monks travelled among religious centres. Such mobility created opportunities for new associations based on profession, patronage and commercial interest. It also created situations in which outsiders had to be classified. Market society therefore did not abolish hierarchy; it produced additional contexts in which social reputation and group identity mattered. The relation between caste and commerce was dynamic. Some occupational communities could become more corporate through repeated exchange networks, while individual mobility could challenge a purely local hierarchy. Early medieval social formation must accommodate both tendencies. 29.22 Mudgagiri and Aṅga: a politically cosmopolitan social setting Mudgagiri’s repeated role as a Pāla jayaskandhāvāra and the large administrative rosters issued from it make the Aṅga–Munger zone especially important for the study of classification. Court, army, transport and documentation brought together people of different geographic origins and service positions. The Munger plate’s list of Gauḍa, Mālava, Khaśa, Hūṇa, Kulika, Karṇāṭa and Lāṭa personnel is evidence for this cosmopolitan political environment, regardless of the precise status of each label. Such a setting could generate new dependencies and reputations. It also demonstrates why later caste geography cannot be imagined as a collection of isolated villages. State centres connected local society to wider circuits of recruitment and migration. Aṅga’s social history therefore includes both rooted agrarian communities and mobile political-service populations, whose classifications operated on different principles. 319319 GAJENDRA THAKUR 29.23 Mithila and the problem of uneven evidence Tīrabhukti appears in the Bhagalpur charter, giving a direct regional anchor for the ninth century, but the surviving social documentation for Mithila is not as continuous as historians might wish. Later medieval records are much richer for Brahman genealogies, court culture and marriage regulation. The temptation is to bridge the documentary gap by treating later institutions as survivals from remote antiquity. This chapter rejects that method. Where a ninth-century charter names a category, it may be used for the ninth century. Where a fourteenth-century or nineteenth-century source describes a ranked community, it must first be used for its own period. Continuity can then be argued only if intermediate evidence supports it. This chronological discipline is especially important in Mithila, where powerful later traditions of genealogy can easily create the illusion that every later social distinction was already fully formed under the Pālas. 29.24 Regional identity and caste identity are different things Labels such as Maithila, Gauḍa or Karṇāṭa can refer to regions, cultural affiliations, scholarly identities or political origins, depending on context. A regional adjective may later become attached to a caste community, but the two categories should not be conflated from the start. Migration can convert regional origin into a durable social marker, especially when newcomers reproduce marriage networks and institutions. Conversely, people of different social ranks can share the same regional identity. The emergence of a specifically Maithil Brahman identity in later medieval evidence is therefore a historical development involving territory, genealogy and authority, not proof that every Brahman in ancient Videha belonged to the later corporate community. The same rule applies elsewhere: geography can contribute to social identity, but regional belonging does not by itself establish caste boundaries. 29.25 Brahmanisation and Sanskritisation are processes, not a one-way map Historians often use terms such as Brahmanisation or Sanskritisation to describe the spread of Brahmanical institutions, practices and status vocabularies. These concepts are useful only if they preserve negotiation. Local rulers sought ritual legitimation; Brahmans depended on patronage and agrarian support; communities could adopt, reinterpret or resist ranked categories; religious traditions competed and overlapped. Classification therefore emerged through interaction rather than one-sided cultural replacement. A local group might seek higher status by changing ritual practice or genealogy, while elites could redefine local identities within textual schemes. Such processes also produced exclusions. The result was neither total fluidity nor a single unchanged code. It was a historically changing field in which social actors used available languages of rank to secure resources, honour and marriage alliances. 29.26 Mobility, claim-making and the negotiation of rank Hereditary classification does not eliminate social mobility; it changes the forms through which mobility is pursued. Families can seek new titles, land, patronage, ritual recognition or genealogical narratives. Political upheaval can elevate military or administrative groups; migration can detach a community from an older rank; wealth can support new claims; defeat or dependency can lower status. Early medieval inscriptions preserve some of these processes indirectly through the appearance of new local magnates, subordinate rulers and beneficiaries. Later genealogies often naturalise the achieved position by projecting it backward. A historical approach therefore reads status claims critically. The question is not whether a community’s tradition is “true” or “false” in a simple sense, but when a claim appears, what institutions recognised it, and what social interests it served. Rank was real precisely because it had to be repeatedly reproduced and acknowledged. HISTORY OF MITHILA, VAJJI & ANGA — VOLUME II Figure 114 — From social role to more bounded identity: possible pathways 29.27 Dharmashāstra ideal and inscriptional practice Dharmashāstra literature provides an elaborate language of varṇa, mixed descent, occupation, inheritance and purity. It shaped elite ideas and legal reasoning, but inscriptions show a social world with categories that do not fit neatly into four boxes. The Pāla charters are crowded with offices, household ranks, service groups and regional names. This discrepancy is not evidence that normative texts were irrelevant; rather, it shows that norms worked on a complex social field. Learned authors could reinterpret occupational or regional groups within varṇa-based schemes, while political documents continued to use categories required by administration. The history of caste formation lies partly in the growing effort to connect these two worlds: to systematise diverse social groups through Brahmanical classification while those groups retained distinct economic and regional histories. 29.28 Why later caste lists cannot be projected backward Late precolonial and colonial observers often present social groups as if each possessed an ancient, bounded and universally recognised place. Such lists are valuable evidence for the period in which they were compiled, but they are not keys that decode every earlier inscription. A modern caste name may descend from an older occupational term, absorb several earlier groups, split from a broader community, or acquire a new ranked meaning. Figure 115 therefore applies a time-depth filter. Contemporary inscriptions receive the greatest weight for their period; near-contemporary normative texts require genre analysis; later genealogies are evidence for later corporate formation; colonial surveys are late snapshots. Backward projection erases change, which is precisely the phenomenon the historian is trying to explain. 321321 GAJENDRA THAKUR Figure 115 — Time-depth filter for using social labels 29.29 A specific Mithila caution: do not medievalise later Brahman sub-castes Mithila’s later Brahman genealogical system is unusually rich, but its richness makes chronological caution more necessary, not less. Later ranked labels within Maithil Brahman society must not be inserted into the Pāla or early medieval record unless contemporary evidence supports them. In particular, this volume does not treat Śrotriya as a fixed Maithil Brahman sub-caste in the medieval period. The project’s Panji-based chronology places the emergence of Śrotriya as a distinct sub-caste around 1800 CE; earlier uses of śrotriya in Sanskrit can denote Vedic learning or qualification and are not the same corporate category. Chapter 30 examines Brahmanical institutions without retrospective caste projection, and Chapter 31 treats the Panji tradition directly. The present chapter therefore establishes the governing rule: later internal rankings belong first to the period in which they are actually evidenced. 29.30 Conclusion: classification was made, reproduced and revised The strongest evidence from Mithila, Vajji and Aṅga does not support either a timeless caste order or a society without hierarchy. Pāla charters reveal graded social language, powerful Brahmans, local notables, cultivators, specialised service populations, mobile origin groups and communities placed at the social margins. Agrarian expansion, political office, institutional patronage, household reproduction and normative classification could make some identities more hereditary and more bounded over time. Yet the process was uneven, and later genealogies reorganised earlier social memories. Social classification must therefore be written as history: categories were made visible by institutions, reproduced through resources and marriage, contested by mobility and claim-making, and repeatedly reinterpreted. Chapter 30 now turns specifically to Brahmanical institutions, applying this same chronological discipline so that later Maithil caste structures are not projected backward into periods that did not yet possess them. Table 29.1 — Evidence for social classification and limits of inference Evidence Category / content Historical use Limit of inference HISTORY OF MITHILA, VAJJI & ANGA — VOLUME II Evidence Category / content Historical use Limit of inference Munger plate of Officials; Brahmans; Administrative Not a census or Devapāla mahattaras; recognition of complete map of kuṭumbins; Meda, differentiated and marriage and Andhraka and Caṇḍāla ranked society occupation range Bhagalpur plate of Royal personnel, Direct Mithila-region Formulaic address does Nārāyaṇapāla service groups, evidence for layered not prove fixed residents and classification hereditary closure cultivators in Tīrabhukti Mahattara / Local seniority and Rural hierarchy Should not be kuṭumbin substantial household outside a simple four- translated as modern standing varṇa chart caste names Kṣetrakara Resident cultivator Agrarian role and Occupation does not liable to deliver fiscal relationship automatically equal assigned revenue caste Origin / military Gauḍa, Mālava, Khaśa, Mobility and multi- Geographic or service labels Hūṇa, Karṇāṭa, Lāṭa regional recruitment labels are not and others inherently jātis Dharmashāstra and Varṇa, purity and Normative language Prescription and Purāṇic classification mixed-category models influencing social polemic are not ranking demographic description Early medieval Increasing organisation Model for gradual Must be tested against eastern Indian of landholding and systematisation Mithila–Aṅga comparison occupational groups toward jāti evidence rather than imported wholesale Later Maithil Corporate identities, Essential evidence for Cannot be projected genealogy and caste marriage regulation later formation backward without accounts and ranked internal intermediate evidence categories 323323