The Genesis of Caste in Ancient India: From Class Fluidity to Institutional Lock‑In

Abstract
Caste is often treated as an eternal feature of South Asian society, but a closer reading of archaeology, text, and economic logic shows a far more contingent story. This article traces that story from Mehrgarh’s first cereal plots to the smartphone age, weaving together narrative history and institutional theory. It argues that caste survived not because Indians loved hierarchy but because, at key junctures, it delivered coordination gains that outweighed the costs of inequality. Once those gains accrued, a path‑dependent feedback loop—formalised here through a replicator model—locked the system in. The essay concludes by identifying the policy levers that could finally break caste’s increasing‑returns logic.



1  Introduction: Why Another History of Caste?

Walk into an Indian newsroom, a Bangalore start‑up cafeteria, or a WhatsApp family group and you will hear two contradictory claims: that caste is obsolete and that caste, in fact, structures every marriage, job referral, and political rally. Both cannot be true—yet both are. Caste’s resilience defies easy explanations founded on culture or religion alone. This article therefore offers a hybrid lens. It braids together (a) a longue‑durée narrative beginning with Neolithic villages, (b) new‑institutionalist theory on increasing returns and path dependence, and (c) a game‑theoretic model that specifies when a society chooses hereditary endogamy over open status competition. The core argument is simple: once enough households found that marrying, hiring, and lending within caste reduced transaction costs, caste became the cheapest social technology available; reversing it would require an even cheaper alternative.

2  Institutions, Increasing Returns, and the Power of Early Advantage

Douglass North defines institutions as the “humanly devised constraints” that shape interaction. Some constraints bear constant or diminishing returns; others, like the QWERTY keyboard or caste, offer increasing returns—the more people adopt them, the more attractive they become. W. Brian Arthur’s 1989 model of “competing technologies” shows that when agents face switching costs and learn by doing, even a slightly better‑than‑random early advantage can snowball into near‑permanent lock‑in. Applied to caste, the implication is stark: once a quarter of households insist on endogamy and ritual purity, every additional household has an incentive to follow, lest it lose marriage partners, credit networks, and ritual services.

3  From Occupational Class to Ritual Rank: A Deep History

3.1  Mehrgarh to the Indus: Specialisation without Hereditary Status (7000–1500 BCE)

Excavations at Mehrgarh, dated as early as 7000 BCE, reveal dental drills, lapis‑lazuli beads, and granaries—clear signs of occupational specialisation. Yet burial goods show graded rather than categorical status differences: some tombs contain more ceramics, but none are segregated by exclusive iconography or inherited symbols of ritual rank. The Indus Valley cities (c. 3300–1300 BCE) pushed specialisation further—dockyards at Lothal, standardized weights, a city wall at Dholavira—but seals and figurines depict no priestly caste monopolising sacrifice. Specialists existed; castes did not.

3.2  Indo‑Aryan Migrations and Mythic Justification (c. 1500–1000 BCE)

Into this proto‑urban landscape arrived Indo‑Aryan clans whose economic base was mobile cattle. The Rig Veda, compiled over centuries, records skirmishes (dāsa‐ari battles), cattle raids, and sacrificial rituals. The hymn most frequently cited as caste’s seed is the Purusha Sukta (RV 10.90). Scholars debate its date, but few doubt its symbolic heft: society is imagined as a cosmic body whose mouth births Brahmins, arms Kshatriyas, thighs Vaishyas, and feet Shudras. Yet the same Rig Veda praises warrior‑priests and admits that poets of low rank could compose hymns—suggesting fuzzy rather than rigid boundaries.

3.3  Late Vedic Agrarianisation and Legal Codification (c. 1000–600 BCE)

As the Aryan frontier moved eastward toward the humid Gangetic plains, iron ploughs replaced cattle raiding. Wet‑rice agriculture demanded stable field labour and clear rules over land inheritance. Around this time the śrauta ritual manuals multiply, each assigning sacrificial fees by varṇa. The Dharmasūtras, redacted by 600 BCE, specify varṇa‑specific fines for crimes: a Shudra insulting a Brahmin incurs higher penalty than the reverse. What had been fluid status markers now intersected with penal codes.

3.4  Mauryan Buddhism: The First Major Disruption (268–185 BCE)

Emperor Aśoka’s edicts on stone pillars preach dhamma, frown upon animal sacrifice, and invite “all castes” to moral training. Buddhism’s sangha accepted lay converts regardless of birth, raising the expected benefit B of abandoning caste. Yet the Arthaśāstra, probably finalised under the Mauryas, still collects taxes by varṇa and punishes occupational mixing. The rural tax machine found it cheaper to tag peasants by caste than to audit individual incomes. Hence, Buddhism dented ritual supremacy but left agrarian caste intact.

3.5  Gupta Land Grants and the Path‑Dependence Trigger (4th–6th c. CE)

Gupta copper‑plate charters donated tax‑free villages (brahmadeya) to Brahmins, who, in turn, performed legitimising rites for the king. These grants made ritual status synonymous with fiscal status. Manusmṛti, often assigned a final redaction in this era, froze occupational duties (dharma) by birth and tied pollution rules to land. Every additional Brahmin village deepened the network of jajmani—hereditary service ties between potters, barbers, and landed patrons—thereby raising the coordination bonus k of playing caste.

3.6  Temple Polity and the Jajmani Order (7th–12th c.)

In the Kaveri delta, Chola inscriptions list washermen, oil‑pressers, and devadāsi dancers as tied to temples for generations. The jajmani system stabilised wages in grain, obviating courts and contracts. Endogamy guaranteed skill transfer and debt repayment within a single jāt cluster; defection threatened not only family honour but a multi‑cast network of reciprocal obligations, so the private cost c of marrying out soared.

3.7  Sultanate and Mughal Fiscal Bargains (13th–18th c.)

Contrary to nationalist tropes, Islamic rulers neither eradicated nor invented caste. They co‑opted it. The Delhi Sultans farmed land revenue through khots and muqaddams—often upper‑caste Hindus—who reconciled sharia tax demands with local ritual calendars. Akbar’s Āʾin‑i Akbarī lists occupational groups to calibrate revenue rates. Mughal farmāns exempted temple estates from jizya, exchanging tax compliance for political calm. Caste now served a cosmopolitan court: Persian chroniclers praised emperors for “upholding varṇa order,” proving that ideology travels well when it lowers administrative costs.

3.8  Colonial Codification and the Cementing Census (1790–1947)

The East India Company first tried to apply English common law to its new subjects but soon delegated “personal law” to Brahmin pandits. F. Max Müller’s philological labours and Herbert Risley’s 1901 census ranked 2,378 castes, each slotted under Brahmin‑Kshatriya‑Vaishya‑Shudra or “outcaste.” The census, once per decade, forced Indians to choose a single identity, magnifying the cost c of switching. Missionary schools offered alternative status through education, but government job quotas by caste re‑embedded the label in modern careers.

3.9  Republic, Mandal, and the Digital Age (1950–Present)

The Indian Constitution bans untouchability (Article 17) and promises affirmative action. Reservations altered the payoff matrix: for marginalised groups, the benefit B of caste identity rose because it unlocked government jobs and college seats. Upper‑caste groups responded by communal mobilisation—Religion based politics in the 1990s—thereby boosting the religious legitimacy multiplier R for hierarchy. Today, LinkedIn alumni lists, matrimonial apps, and AI résumé filters encode surnames almost as effectively as village gossip once did.

4  A Formal Model of Caste Lock‑In

Imagine a village where households choose between strategy C (observe caste) and O (open choice). Each year they must marry children and hire labour. Married‑within networks yield dowry reciprocity, debt insurance, and trust; married‑out couples pay social ostracism. Let base income be B. Playing C adds coordination bonus kx, where x is the share of households already playing C. Playing O subtracts boycott cost c. Pay‑offs:

P_C = B + kx,

P_O = B – c – R.

Under replicator dynamics,

dx/dt = x(1 − x)(kx + c+R).

Equilibria sit at x = 0 (open society) and x = 1 (universal caste). The tipping share is x* = c/k. If coordination benefits outstrip boycott cost (k > c), the tipping point is small. Empirical estimates place caste‑based credit premia near 40 % of annual income, while boycott costs in non‑violent contexts hover at 10 %. Thus x* ≈ 0.25: once a quarter of households enforce caste, chaos for dissenters exceeds reward, and conformity becomes rational.

5  When and Why the Equilibrium Shook

5.1  Buddhism raised B (alternate status) and lowered R (ritual stigma) yet failed to slash k (agrarian coordination). Monks depended on lay donors who themselves stayed endogamous.

5.2  Bhakti and Sufism weakened ritual barriers—Low‑caste saints like Kabir mingled with Muslims—but left land contracts intact, so k remained high.

5.3  Colonial Schools lifted B for English‑educated groups but Risley’s census inflated c; conversions and casteless nationalism never reached tipping mass.

5.4  Post‑Mandal Politics multiplied B via reservations yet bolstered k through caste vote‑banks—the equilibrium stretched but re‑stabilised.

6  Global Echoes: Social Darwinism and the Biology of Inequality

Caste’s Indian longevity is extraordinary, but the institutional logic is global. Victorian Britain deployed phrenology, and the American South invented Jim Crow, to naturalise labour control. Herbert Spencer’s “survival of the fittest” and Thomas Malthus’s arithmetic of hunger legitimised laissez‑faire misery. Each system offered elites coordination gains (cheap labour, predictable hierarchy) against the moral cost of oppression. Caste differs in longevity, not in kind.

7  Policy Levers: Lower k, Slash c, Raise B Wisely

Universal public goods—healthcare, crop insurance—substitute caste insurance, depressing k.

Fintech credit authenticated by biometric IDs can bypass caste‑based lenders, again cutting k.

A strong anti‑boycott law that fines caste councils for harassing inter‑caste couples directly lowers c.

Media campaigns that glamorise mixed marriages and inter‑caste startups can raise the reputational benefit of openness, effectively boosting B for strategy O.

A caste census could help if paired with redistribution; without reforms it risks raising c by freezing identity further.

8  Conclusion: From Cosmic Body to Economic Puzzle

Caste began as a poetic metaphor for social differentiation and hardened into a low‑transaction‑cost technology for managing labour, credit, and marriage. Each epoch—Gupta land grants, Mughal revenue farms, colonial censuses—added layers of coordination benefit or switching cost, until endogamy became a self‑reinforcing equilibrium. Dismantling caste therefore requires more than moral exhortation; it demands a competing social OS whose insurance, credit, and coordination services outclass caste’s 3,000‑year‑old package. The economic calculus that once fossilised hierarchy might yet be harnessed to dissolve it.

Select Bibliography

Abū‑l‑Fażl. Āʾin‑i Akbarī (c. 1590).

Ambedkar, B. R. Annihilation of Caste (1936).

Arthur, W. Brian. “Competing Technologies and Lock‑In by Historical Events” (1989).

Bayly, Susan. Caste, Society and Politics in India (1999).

David, Paul. “Clio and the Economics of QWERTY” (1985).

Dirks, Nicholas. Castes of Mind (2001).

Manusmṛti, critical ed. Patrick Olivelle (2005).

North, Douglass C. Institutions, Institutional Change and Economic Performance (1990).

Risley, Herbert H. The People of India (1908).

Truschke, Audrey. Culture of Encounters (2016).