SOCIAL DIFFERENTIATION BEFORE FIXED CASTE NARRATIVES

CASTE NARRATIVES 8.1 Scope: differentiation is not the same as a finished caste system Early historic society in the Middle Ganga valley was unequal, but inequality should not be reconstructed by placing every person into a later, fully elaborated caste order. Political rank, control of land and labour, ritual prestige, occupation, lineage, gender, wealth, freedom or dependency, and access to institutions could all shape social position. These axes overlapped, yet they did not always coincide. A wealthy merchant might possess substantial economic power without belonging to a ruling lineage; a ritual specialist could claim high prestige without controlling the largest productive resources; an artisan could be indispensable to urban life while occupying a lower social rank in some textual classifications. The task is therefore to reconstruct historically visible forms of differentiation before assuming the later fixity of endogamous jātis. This follows the broader methodological shift from treating caste as a timeless fourfold structure to examining how social formations changed with settlement, production, state formation and ideology (Thapar 1999; Chakravarti 1987; Sharma 1983). 8.2 The sources speak in different social vocabularies No single source provides a complete social census. Brahmanical normative texts organise society through varṇa and rules of ritual eligibility, marriage, inheritance and conduct. Early Buddhist texts frequently identify people instead through terms such as khattiya, brāhmaṇa, gahapati, seṭṭhi, vāṇijja, occupational labels, and categories of dependence such as dāsa and kammakāra. Archaeology reveals differences in house size, access to imported or prestige goods, craft production and consumption, but it rarely supplies the names by which people classified themselves. Later inscriptions make some occupational and donor categories more visible, yet they post-date much of the period considered here. These vocabularies must therefore be compared rather than collapsed. Apparent disagreement between texts may record different institutions, audiences and purposes rather than a simple choice between a “true” and a “false” social order (Fick 1920; Chakravarti 1987; Aktor 2017). 8.3 Social position worked through several overlapping axes The most useful model is multidimensional. Birth and lineage mattered in ruling groups and ritual claims. Property and control over productive resources mattered for substantial householders. Occupation mattered because people could be identified as traders, cattle-keepers, metalworkers, potters, barbers or other specialists. Political rank mattered because participation in a gana-sangha assembly was not open to every resident. Dependency mattered because servants, hired workers and enslaved persons did not command the same autonomy as household heads. Gender and age structured access to property, ritual, movement and authority. These axes could reinforce one another, but they could also diverge. Figure 28 therefore treats hierarchy as a field of intersecting relations rather than a single ladder. GAJENDRA THAKUR Figure 28 — Social differentiation before fixed caste narratives: several overlapping axes. No single axis explains all early historic hierarchy. 8.4 Varṇa was a powerful classificatory ideal, not a demographic spreadsheet The fourfold varṇa scheme—Brāhmaṇa, Kṣatriya, Vaiśya and Śūdra—became one of the most influential Brahmanical ways of imagining social order. Dharma literature attached consequences to these categories in ritual entitlement, marriage, work, punishment and social interaction. Yet modern scholarship stresses that the scheme was normative and ideological; other kinds of evidence do not simply reproduce it as a complete description of lived society (Aktor 2017). This distinction is especially important for north-eastern India in the age of the mahājanapadas. A textual rule assigning an occupation to a varṇa does not prove that every historical practitioner of that occupation belonged to one empirically identifiable group, nor does it establish that the same ranking operated uniformly in Vaishali, Champa, rural settlements and frontier zones. 8.5 Buddhist social vocabulary does not map neatly onto the four varṇas Uma Chakravarti’s analysis of early Buddhist literature, building critically on Richard Fick, emphasises that the categories most visible in Pāli social description do not line up neatly with the full Brahmanical fourfold scheme. Brāhmaṇas and khattiyas are conspicuous, but economic life is often described through categories such as gahapati, seṭṭhi, merchant, cattle-keeper and named crafts rather than through an empirical “Vaiśya” bloc. Likewise, labouring dependence is frequently expressed through dāsa and kammakāra rather than by making “Śūdra” the ordinary descriptive term for all service work (Fick 1920; Chakravarti 1987). This does not mean that varṇa ideas were absent; it means that social practice generated additional categories that a historian must take seriously on their own terms. 8.6 The gap between normative and empirical categories is historical evidence The mismatch between classificatory schemes is itself informative. It shows a society in which expanding agriculture, cities, merchant wealth, specialised crafts, political oligarchies and renunciant institutions created forms of status that could not be described adequately through one inherited ideological vocabulary. The historian should therefore resist translating gahapati as “Vaiśya,” khattiya as a simple equivalent of every later HISTORY OF MITHILA, VAJJI & ANGA — VOLUME II Kṣatriya caste, or dāsa-kammakāra as a synonym for Śūdra. Figure 29 visualises this non-equivalence. Categories could intersect, and later traditions could assimilate them more closely, but the early historic evidence must be allowed to retain its plurality. Figure 29 — Normative varṇa scheme and empirical social categories in early Pāli sources. The relationship is not one-to-one. 8.7 Brāhmaṇa status joined ritual authority to social competition Brāhmaṇas appear in Buddhist as well as Brahmanical sources, which indicates that the category had social visibility beyond texts written by Brahmanical specialists. Yet the sources also preserve arguments about what made a true brāhmaṇa: birth, learning, conduct, sacrifice, austerity or ethical achievement. Such debate implies that status was claimed and contested, not merely recorded. In the expanding eastern polities, ritual specialists interacted with rulers, householders and renunciants in an environment where new forms of wealth and political authority were developing. Their prestige could be substantial, but it was not identical to political sovereignty or mercantile wealth. The presence of contest over the meaning of brāhmaṇa status should therefore be treated as evidence of social negotiation rather than as proof that a single uncontested hierarchy governed every community. 8.8 Khattiya status was especially important in ruling lineages and gana-sanghas Pāli sources use khattiya prominently for ruling and aristocratic groups. In gana-sangha contexts such as the Licchavis and other oligarchic lineages of the north-east, khattiya identity was closely tied to political privilege and descent. This makes it dangerous to translate republican or assembly government into social democracy. A polity could distribute authority among multiple elite lineages while excluding most residents from the decisive political body. Thapar’s distinction between lineage systems and state formations is useful here: collective rule among chiefs or lineage heads is still compatible with sharp internal hierarchy (Thapar 1999). The political significance of khattiya status also helps explain why social rank cannot be reduced to ritual ranking alone. GAJENDRA THAKUR 8.9 Gahapati marked substantial household status in agrarian and urban society The gahapati is one of the most important categories that disappears when early society is forced into a four-varṇa template. In early Buddhist literature the term often denotes a substantial lay household head associated with property, agriculture, wealth, productive resources and local standing. Gahapatis could support renunciants, employ labour and participate in market and agrarian networks. Chakravarti therefore treats the category as central to the new social economy of the Middle Ganga valley (Chakravarti 1987). Not every gahapati was a merchant, and not every householder was a gahapati of comparable means. The term points to a social location produced by household control of resources; it should not be mechanically converted into a later caste name. 8.10 Seṭṭhi wealth created another axis of urban prestige The seṭṭhi or wealthy merchant-financier represents a form of prominence associated with commerce, capital, credit and urban institutions. Literary traditions about major donors illustrate how mercantile wealth could produce public influence without being identical to royal or ritual authority. Older scholarship sometimes attempted to fit such people directly into the Vaiśya category, but the texts often identify them by economic title instead. This distinction matters for Champa and other commercial centres discussed later in the volume. Urbanisation created social power through circulation and accumulation as well as through land and lineage. Wealth could support monasteries, public works and networks of dependants, giving economic elites a visibility that a purely varṇa-based reconstruction would understate (Fick 1920; Chakravarti 1987). 8.11 Occupation was a practical language of social identity Early Buddhist narratives identify people as traders, cattle-keepers, potters, metalworkers, barbers and other specialists. Such labels show that occupation could function as a meaningful social identity, especially for groups outside ruling and ritual elites. This does not mean that every occupation was already a closed hereditary caste. A profession could be learned within families, organised through local networks and associated with status without yet possessing all the later features of a jati. Fick’s early twentieth-century study already noted the complexity of north-eastern society, and Chakravarti’s later work emphasised the empirical importance of occupational categories in the Pāli record (Fick 1920; Chakravarti 1987). The safest historical language is therefore “occupational group” unless the evidence demonstrates heredity, endogamy and durable corporate boundaries. 8.12 Skilled crafts could be economically essential and socially unequal Metalworking, pottery, carpentry, textile production, ivory working, leather work and other crafts were indispensable to early urbanisation. Archaeological concentrations of specialised production show increasing division of labour, while texts supply names for some occupational specialists. Economic necessity, however, did not guarantee equal prestige. Work involving particular materials, bodily substances or service relations could carry different valuations in different textual traditions. At the same time, skilled production created expertise and bargaining power. The relationship between occupation and rank was therefore neither uniformly degrading nor uniformly empowering. Chapter 11 examines craft organisation in detail; here the key point is that occupational differentiation preceded and contributed to later social boundary-making without allowing us to identify every early craft group with a later caste bearing a similar occupation. HISTORY OF MITHILA, VAJJI & ANGA — VOLUME II 8.13 Dāsa and kammakāra reveal relations of dependence and coercion The presence of dāsa and kammakāra in early sources demonstrates that some households and institutions depended on labour performed by people with less autonomy than substantial household heads. The terms cover forms of servitude, slavery, hired labour and dependent work whose exact legal conditions varied by text and period. It would be misleading either to erase coercion by translating every dāsa as a generic “servant” or to assume that all such persons occupied one uniform form of chattel slavery. Sharma’s work on lower orders and Chakravarti’s social history both insist on relating status to labour and control over resources rather than discussing ritual rank in isolation (Sharma 1990; Chakravarti 2007). These categories make inequality visible from below. 8.14 Labour status could cut across later varṇa expectations A key methodological consequence follows from the prominence of dāsa-kammakāra and occupational workers: economic dependence did not need to be expressed through the same categories as Brahmanical ritual classification. A person could be socially subordinate because of labour relations, debt, household dependency or lack of property even when the text did not label that person a Śūdra. Conversely, the abstract assignment of service to Śūdras in normative literature does not identify the status of every worker in the archaeological or Buddhist record. This separation between economic relation and ritual category is crucial for writing a socio-economic history that does not substitute prescription for observation (Sharma 1983; Aktor 2017). 8.15 Gana-sangha government was oligarchic rather than socially egalitarian Vaishali’s association with the Vajjian confederacy can encourage anachronistic descriptions of a “republic” as if all inhabitants enjoyed equal citizenship. The surviving traditions instead point to assemblies of politically entitled lineages and chiefs. The famous Vajjian conditions in the Mahāparinibbāna Sutta concern collective meetings, established institutions, elders, women, shrines and arahants; they describe norms of political cohesion, not universal suffrage (Walshe 1995). Mishra’s reconstruction of the Vajjian constitution remains historiographically important but uses modern constitutional analogies that must be handled cautiously (Mishra 1962). Figure 30 therefore presents political participation as one axis of hierarchy: collective rule among elites could coexist with household wealth differences, occupational stratification and dependent labour. GAJENDRA THAKUR Figure 30 — Differentiated participation in a gana-sangha polity. Collective rule among entitled lineages did not imply universal social equality. 8.16 Lineage rank and economic wealth were related but not identical Ruling descent could confer access to office and assembly participation, while mercantile or agrarian wealth could create substantial influence outside the political lineage core. A wealthy gahapati or seṭṭhi might command more movable resources than a poorer member of an aristocratic lineage, yet the latter could possess privileges of birth and political participation unavailable to the former. This is exactly why a single hierarchy is insufficient. Early state formation brought several ranking systems into contact: descent, ritual status, control of labour, market wealth and office. Their interaction created opportunities for competition, patronage and alliance, but not necessarily a unified scale on which every person could be placed. 8.17 Gender structured every social category without producing one uniform female status Women were differentiated by family status, wealth, marital position, age, occupation, enslavement or dependency, and access to religious institutions. Elite women could control resources in exceptional cases; poorer women performed agricultural, domestic, craft and service labour that is much less visible in elite texts. Normative restrictions on marriage, inheritance and sexuality therefore need to be read alongside evidence for female donors, courtesans, nuns and household managers. Gender was not a separate hierarchy added after caste or class; it intersected with them. The same label of high family status could produce different opportunities for men and women, while renunciation could create new forms of female authority within continuing institutional constraints (Collett 2016; Chakravarti 2007). 8.18 Ambapālī shows why social categories must be read case by case Traditions about Ambapālī of Vesālī portray a woman of wealth, public visibility and control over property who interacts directly with the Buddha and the Licchavis. Her case cannot be made representative HISTORY OF MITHILA, VAJJI & ANGA — VOLUME II of all women, nor should later labels about courtesans be used to assign her a simple place in a caste ladder. Instead, the narrative demonstrates that urban societies could recognise forms of female economic agency that intersected with sexuality, patronage, civic prestige and elite competition. The example is valuable precisely because it resists a flat model: gender constraint coexisted with forms of wealth and public authority. Chapter 9 examines women in greater detail. 8.19 Renunciation challenged some status claims without abolishing social hierarchy Early Buddhist doctrine repeatedly questions the idea that spiritual worth is guaranteed by birth. Admission to the saṅgha could create a new institutional identity in which seniority by ordination and discipline mattered. Stories such as that of Upāli, traditionally remembered as a barber connected with the Sakyan princes, dramatise a reversal of worldly precedence when elite men are asked to honour someone ordained before them. Such narratives are powerful critiques of inherited status, but they should not be turned into evidence that caste or class disappeared in Buddhist communities. Monasteries depended on lay economies, retained gendered rules, and operated within societies structured by wealth, family and political power. The evidence therefore shows ideological challenge and institutional re-ranking, not an instant egalitarian social revolution (Chakravarti 1987; Horner 1952). 8.20 Ethical universalism and social practice could coexist in tension Buddhist texts can relativise birth status by making conduct and insight central to moral and spiritual evaluation. At the same time, the same textual world recognises khattiyas, brāhmaṇas, high and low families, wealthy householders, servants and occupational groups. The coexistence of these registers is historically important. Religious critique did not require authors to deny the social classifications around them; rather, it could redirect the criterion of ultimate worth. Historians should therefore avoid two opposite errors: treating Buddhism as simply reproducing Brahmanical hierarchy, or describing it as having abolished social distinction in practice. The tension between ethical universality and inherited social relations is itself part of the history of early north-eastern India. 8.21 Kula and jāti cannot simply be translated as modern caste Words translated as “family,” “birth,” “kind,” “lineage” or “caste” changed meaning across genres and centuries. Kula may refer to family or lineage status; jāti can indicate birth or a social category without necessarily denoting a fully formed modern-style endogamous caste. Thapar’s discussion of social terminology during the transition from lineage to state warns against assuming that later meanings were already fixed in the mid-first millennium BCE (Thapar 1999). The safest method is contextual: ask what a term does in a particular passage—does it rank families, regulate marriage, identify occupation, distinguish political descent, or make a doctrinal argument? Translation should follow that function rather than importing the whole later institution into a single ancient word. 8.22 Marginalised labels do not justify back-projecting later untouchability unchanged Early texts contain labels for people represented as low, impure, despised or socially distant, including caṇḍāla and pukkusa/pukkasa in some traditions. Such evidence is important for the history of exclusion, but it does not by itself prove that the full later institution of untouchability, with all its regional practices and hereditary boundaries, already existed in an identical form. Sharma’s longue-durée study of lower orders GAJENDRA THAKUR demonstrates that social degradation developed through changing ritual, occupational and economic relations over centuries (Sharma 1990). A historically controlled account should record early stigma and exclusion where the sources attest them while refusing to convert resemblance into timeless continuity. 8.23 Hunters, fishers and forest-linked workers complicate agrarian models Not everyone fitted neatly into an agrarian village or urban craft economy. Texts refer to hunters, fishers and other people whose livelihoods depended on rivers, wetlands, forests and animal resources. Archaeology from the region likewise demonstrates the continuing importance of fish, turtle, molluscs and wild resources discussed in Chapters 3 and 4. Such livelihoods could be economically essential while being represented ambivalently or negatively in some normative traditions. Their presence reminds us that social differentiation also followed ecological niches. Later caste histories sometimes attach particular communities permanently to such occupations, but early evidence should be described at the level it supports: livelihood, occupational naming and social valuation, not assumed unbroken community identity. 8.24 Urban neighbourhoods cannot be labelled by caste from artefacts alone Archaeological differences within a town—house size, craft debris, luxury goods, drains, wells, streets and waste—can reveal unequal access to resources and specialised activity zones. They cannot normally identify the caste of residents. A quarter containing metalworking debris is evidence for concentrated production, not automatically a “smith caste quarter.” A large residence suggests investment and resources, not a specific varṇa. This caution is essential at sites such as Champa and Vaishali where literary memory is rich enough to tempt retrospective identification. Material inequality and social classification must be connected only where independent evidence allows it (Allchin 1995; Chakrabarti 2009). 8.25 Consumption differences reveal inequality more securely than named status Variation in fine wares, ornaments, beads, imported materials, storage capacity and residential investment can indicate unequal consumption or access. Yet archaeological wealth is not a direct translation of legal or ritual rank. A prosperous merchant household could possess more expensive objects than a ritual specialist; goods could be gifts, heirlooms, trade stock or institutional property. The strongest archaeological claim is therefore comparative: some households or areas had greater access to resources than others. Linking that inequality to a named social category requires textual, inscriptional or spatial corroboration. This evidentiary discipline prevents artefacts from being used to “prove” a caste hierarchy they cannot name. 8.26 Social mobility existed, but it should not be romanticised Economic expansion could create new wealth through agriculture, trade and specialised production. Renunciation could confer new religious status. Service at court or participation in political institutions could alter a household’s position. Such possibilities show that social order was not completely frozen. Yet mobility was constrained by birth, lineage privilege, gender, dependency and inherited resources. Stories of individuals who cross status boundaries are therefore evidence that boundaries mattered, not proof that they were irrelevant. A historical account should recognise both movement and constraint, avoiding the twin myths of an entirely fluid society and an already immobile caste system. HISTORY OF MITHILA, VAJJI & ANGA — VOLUME II 8.27 Anga and Champa require local evidence rather than borrowed social templates Champa’s role as an urban and commercial centre makes it plausible that merchants, artisans, labourers, householders and political authorities interacted there in complex ways. Archaeology supports urban investment and craft activity, while Buddhist and Jain traditions place Anga within wider networks of the eastern Ganga valley. What is much harder to recover is the exact local ranking of named social groups. It would be methodologically weak to take a hierarchy reconstructed for Magadha or Kosala and simply relabel it “Anga.” Chapter 14 returns to Champa with a stronger economic focus. For the present chapter, Anga demonstrates the need to distinguish region-wide processes from locally proven institutions. 8.28 Nepal-side Mithila is especially vulnerable to retrospective classification For the early historic Tarai and Nepal-side Mithila, archaeological and textual evidence remains uneven. Later social formations, landholding structures and caste labels are much better documented, but they cannot be carried backward across a millennium or more. Shared religious traditions and cross-border mobility do not prove that every later community category existed with the same name and boundary in the early historic period. This volume therefore treats the Nepal Tarai as part of an interaction zone while keeping the evidentiary frontier visible. Where the data are insufficient, absence of reconstruction is preferable to invented continuity. 8.29 Later caste formations must be studied as historical outcomes More durable caste formations could emerge through the long interaction of occupational specialisation, hereditary transmission of work, marriage boundaries, ritual ranking, political incorporation, control of land and service obligations. But these processes operated unevenly across regions and centuries. Figure 31 therefore represents caste formation as a possible long-term stabilisation of several social boundaries rather than as a system that appeared fully formed at one date. This approach also prevents later Maithil genealogical, Panji, Śrotriya or other historically documented classifications from being projected into the age of the Vajji confederacy. The later evidence is important precisely because it has its own history. GAJENDRA THAKUR Figure 31 — From social differentiation to later caste formations: a long-term process rather than an instant or uniform system. 8.30 Conclusion: early historic hierarchy was plural, contested and changing The strongest conclusion is not that early Mithila, Vajji and Anga were “casteless,” nor that they already possessed the complete social architecture familiar from much later periods. The evidence shows stratification through lineage, ritual status, wealth, occupation, political privilege, gender and dependency. Brahmanical varṇa theory was historically influential, but Pāli texts reveal additional empirical categories created by agrarian expansion, urbanisation, trade and household production. Gana-sangha politics could be collective yet oligarchic; Buddhist ethics could challenge birth privilege without dissolving social inequality; occupational groups could be durable without being demonstrably identical to later jātis. Social differentiation should therefore be written as a process. The next chapter turns specifically to women, whose experiences expose particularly clearly how status, labour, family and institutional power intersected. Table 8.1 — Social categories, evidence and limits of inference Category / evidence What it can What it cannot Use in this chapter support establish by itself Varṇa terminology normative ranking, complete used as ideological ritual and legal ideals demographic classification, not structure of a census settlement Khattiya / ruling political descent and universal equality used to separate lineage elite participation in within gana-sanghas oligarchic some polities participation from citizenship Gahapati / seṭṭhi property, household automatic used to reconstruct standing, mercantile identification as economic status or urban wealth Vaiśya caste outside a fourfold template Occupational labels work, skill, practical fully hereditary or used for artisans, HISTORY OF MITHILA, VAJJI & ANGA — VOLUME II Category / evidence What it can What it cannot Use in this chapter support establish by itself social identification endogamous jātis traders, fishers and other specialists Dāsa / kammakāra dependence, labour one uniform legal used to recover subordination, form of slavery labour hierarchy possible coercion from below House size and unequal access to named caste, varṇa or used only as material wealth resources and genealogy archaeological consumption evidence of inequality Later later social unchanged early reserved for later caste/genealogical boundaries and historic identity chapters and never records historical memory retrojected