Mahabharata as Source: Kinship and Caste

A statement question on this chapter typically tests whether a candidate can name Draupadi's marriage as the exam-set example of polyandry among ruling Kshatriya elites, and whether jati is treated as just another word for varna or as a genuinely distinct, unbounded category that kept multiplying while the four varnas stayed fixed. Both traps sit on real, textbook-verified ground: the epic itself offers two separate justifications for one polyandrous marriage, and the NCERT text is explicit that the number of jatis had no ceiling while the number of varnas never moved from four.

This note assumes two things covered elsewhere on this site: the Vedic-age shift from tribal Sabha and Samiti assemblies to a varna-stratified kingdom system, covered in Ancient India Essentials, and the Vedic textual corpus itself, the Samhita-Brahmana-Aranyaka-Upanishad structure of the four Vedas, covered in Vedic Literature, Archaeology and Historiography. This note covers different, genuinely distinct ground: the Mahabharata specifically as a historical source, and the kinship and marriage rules of the period c. 600 BCE to 600 CE.

The Mahabharata as a historical source

The Mahabharata is described in early Sanskrit tradition as an itihasa, literally "thus it was", conventionally translated as "history". In its present form it runs to over 100,000 verses, built around the story of a feud between two sets of warring cousins, the Kauravas and the Pandavas, who belonged to a single ruling family, the Kurus, a lineage dominating one of the janapadas of the Kuru-Panchala region.

Historians read the Mahabharata's own composition history as evidence in itself, not just its content. The text was composed over about 1,000 years, from roughly 500 BCE onwards, through identifiable stages. The original story was probably composed by charioteer-bards known as sutas, who accompanied Kshatriya warriors to the battlefield and composed poems celebrating their victories, compositions that circulated orally. From the fifth century BCE, Brahmanas took over the story and began committing it to writing, a transition NCERT links to the Kurus and Panchalas themselves becoming settled kingdoms that wanted their itihasa preserved more systematically. A second phase, c. 200 BCE-200 CE, coincides with the growing worship of Vishnu and Krishna's identification with him. A third phase, c. 200-400 CE, added large didactic sections resembling the Manusmriti; with these additions, a text that had initially perhaps run to fewer than 10,000 verses grew to about 100,000, traditionally attributed to the sage Vyasa. Historians classify the resulting content under two broad heads, narrative sections and didactic sections (most famously the Bhagavad Gita), while noting the division is not watertight, since didactic sections often carry stories and narrative sections often carry a social message.

Whether the epic preserves the memory of a real historical war is genuinely unsettled: some historians think an actual conflict amongst kinfolk was preserved in the narrative, others point out there is no other corroborative evidence for the battle. One attempt at corroboration comes from archaeology: the archaeologist B.B. Lal excavated Hastinapura, in Meerut, Uttar Pradesh, in 1951-52. Whether this was the epic's Hastinapura cannot be proven from the name alone, though the site's location in the upper Ganga doab, where the Kuru kingdom is placed, makes the identification plausible. Lal found five occupational levels; the second phase (c. twelfth to seventh centuries BCE) showed mud and mud-brick walls with reed-marked plaster, and the third phase (c. sixth to third centuries BCE) showed mud-brick and burnt-brick houses with soakage jars and brick drains. NCERT poses this as an open methodological question rather than settled proof: was the epic's description of the city added once urban centres had actually flourished in the region after the sixth century BCE, or is it poetic description that cannot be verified against other evidence either way.

A separate scholarly project addressed the text's own internal variation. One of the most ambitious projects of twentieth-century Indology, the Critical Edition of the Mahabharata, began in 1919 under the Sanskritist V.S. Sukthankar, whose team collected Sanskrit manuscripts from across the subcontinent, compared verses, and published the ones common to most versions, running to over 13,000 pages across 47 years. The exercise revealed both a common core, evident in manuscripts from Kashmir and Nepal to Kerala and Tamil Nadu, and enormous regional variation in transmission, documented in footnotes and appendices that account for more than half those 13,000 pages. Historians read this variation itself as evidence of the dialogues, consensual and conflictual, between dominant Brahmanical traditions and resilient local practices that shaped early Indian social history. An English translation of the Critical Edition, begun by J.A.B. van Buitenen in 1973, remained incomplete at his death in 1978.

Kinship and marriage: patriliny, gotra and the Draupadi question

At one level, the Mahabharata is itself a story about a change in kinship norms. The Kaurava-Pandava conflict ended in battle, the Pandavas emerged victorious, and patrilineal succession, tracing descent from father to son and grandson, was proclaimed afterwards. While patriliny had existed before the epic's composition, its central story reinforced the idea that it was valuable: sons could claim a father's resources, including the throne, once the father died. Most ruling dynasties from around the sixth century BCE claimed to follow this system, though NCERT notes real variation: sometimes there were no sons, sometimes brothers succeeded one another or other kinsmen claimed the throne, and in very exceptional circumstances women, such as the Vakataka queen Prabhavati Gupta, exercised power directly. Concern with patriliny is also evident in Rigvedic ritual mantras, including one probably inserted around 1000 BCE, chanted by the priest at a wedding, asking that the bride "have fine sons and be fortunate in her husband's love".

Within this framework, marrying daughters into families outside one's own kin group was considered desirable, a system called exogamy (literally, marrying outside), giving rise to the belief that kanyadana, the gift of a daughter in marriage, was a father's important religious duty. From around 500 BCE, Brahmanas laid down detailed codes of social behaviour in Sanskrit texts called the Dharmasutras and Dharmashastras; the most important, the Manusmriti, was compiled between c. 200 BCE and 200 CE. These texts recognised eight forms of marriage, of which only the first four were considered "good"; the remaining four, condemned by the texts, were probably practised by those who did not accept Brahmanical norms.

A second, distinct Brahmanical practice, evident from around 1000 BCE, was to classify people, especially Brahmanas, by gotra, each named after a Vedic seer whose descendants everyone in that gotra was regarded as being. Two rules about gotra mattered: women were expected to give up their father's gotra and adopt their husband's on marriage, and members of the same gotra could not marry each other, gotra being itself a unit of exogamy. Whether this was actually followed can be checked against real names: the Satavahanas, a ruling lineage of western India and the Deccan (c. second century BCE-second century CE), are known from inscriptions to have used names derived from gotras, such as "Gotami-puta" ("son of Gotami"). Some Satavahana rulers were polygynous, and their queens' names show many retaining names derived from their own father's gotra rather than their husband's, contrary to the rule; some of these women even shared their husband's gotra, meaning marriage within the kin group, endogamy, rather than the exogamy the Brahmanical texts recommended, a practice prevalent among several south Indian communities precisely because marrying kinfolk such as cousins kept a community close-knit.

The single most debated kinship episode in the period is Draupadi's marriage to the five Pandava brothers, an instance of polyandry central to the Mahabharata's narrative. Drupada, king of Panchala, organised a competition to string a bow and hit a target; Arjuna won and was garlanded by Draupadi. When the Pandavas brought her home, their mother Kunti, without having seen her, told her sons to share whatever they had won; realising her mistake only once she saw Draupadi, Kunti's command was nonetheless treated as one that could not be violated, and Yudhishthira decided Draupadi would be their common wife. When Drupada protested, the sage Vyasa offered two separate justifications: first, that the Pandavas were in reality incarnations of Indra, whose wife had been reborn as Draupadi and was destined for all of them together; second, that in an earlier life a woman had prayed to Shiva for a husband five times over by mistake, and had been reborn as Draupadi with Shiva fulfilling all five prayers at once. Convinced, Drupada consented.

NCERT's own reading of this episode resists a single tidy conclusion. Present-day historians suggest that the very fact the author or authors described a polyandrous union at all indicates polyandry may have been prevalent among ruling elites. At the same time, so many different explanations stacked onto a single episode suggest polyandry was gradually falling into disfavour by the time these sections were compiled. Some historians connect the practice to the Himalayan region, where it was, and remains, prevalent; others suggest a shortage of women during warfare led to polyandry as a response to crisis rather than a stable norm; some early sources indicate polyandry was never the only, or even the most prevalent, form of marriage in the period. NCERT's caution against over-reading the episode is explicit: creative literature has its own narrative requirements and does not always literally reflect social realities, so an unusual practice in a story is evidence to weigh, not proof of how most people actually married.

A related episode, Draupadi's question, tests the same gendered property logic differently. During a dice game, Yudhishthira staked and lost Draupadi after already staking and losing himself; Draupadi asked whether he had the right to stake her at all, having already lost his own freedom. Two contrary opinions are recorded: that his wife remained under his control regardless, so he could still stake her, or that an unfree man could not stake another person at all. The matter is left explicitly unresolved; ultimately Dhritarashtra restored both the Pandavas' and Draupadi's freedom. Alongside this sits stridhana ("a woman's wealth"): women retained gifts received at marriage as separate property, inheritable by their children, over which a husband had no claim, even as cumulative evidence suggests land, cattle and money more broadly remained largely under male control.

From varna to jati

The Dharmasutras and Dharmashastras laid down rules of occupation for the four varnas. Brahmanas were primarily meant to study and teach the Vedas, perform sacrifices, and give and receive dana (gifts); Kshatriyas to engage in warfare, protect people, administer justice, and also study the Vedas and make dana; Vaishyas to engage in agriculture, pastoralism and trade, alongside similar textual and ritual duties; and Shudras were expected to assist the other three varnas. Other traditions pushed back against treating this as a fixed, birth-determined hierarchy: the Bhagavad Gita itself, alongside Buddhist texts, suggests the four-fold order was based on guna (qualities) and karma (actions) rather than birth alone.

Real political history did not track the varna scheme closely. According to the Shastras only Kshatriyas could be kings, yet the social background of several major ruling lineages is genuinely contested. The Mauryas' own origins were hotly debated even in ancient sources: later Buddhist texts described them as Kshatriyas, Brahmanical texts as of "low" origin; the Shungas and Kanvas, the Mauryas' immediate successors, were Brahmanas by tradition. In practice, NCERT concludes, political power was effectively open to anyone who could muster support and resources, rarely depending on birth as a Kshatriya. The Satavahana ruler Gotami-puta Siri-Satakani illustrates the gap between varna as ideology and varna as practice within a single inscription: he claimed to be simultaneously a unique Brahmana (eka bamhana) and a destroyer of the pride of Kshatriyas, and claimed to have ensured there was no intermarriage among the four varnas, while entering into a marriage alliance with the kin of Rudradaman, the best-known ruler of the Shakas, a Central Asian dynasty the Brahmanas themselves classed as mlechchhas, outsiders to the varna system altogether.

Set against this fixed, contested four-fold scheme is jati, a separate category Sanskrit texts also used for social groups. While the number of varnas stayed fixed at four, there was no restriction on the number of jatis, which kept increasing as society encountered new social groups and occupations. Jatis sharing a common occupation were sometimes organised into shrenis, or guilds. Documentary evidence for these groups' own histories is rare, but a Mandasor inscription (Madhya Pradesh, c. fifth century CE) is an exception: it records a guild of silk weavers who had originally lived in Lata (Gujarat) and migrated to Dashapura, present-day Mandasor, with their children and kinfolk, after hearing of the local king's greatness. The inscription shows the guild's members sharing more than a common craft, collectively funding a temple to the sun god, and, tellingly, some members had by then taken up entirely different occupations, including music, storytelling, religious discourse, Vedic astronomy and even warfare, showing jati and shreni membership was no rigid, permanent occupational cage. Occupation and varna did not always line up either: the play Mrichchhakatika (Shudraka, c. fourth century CE) describes its hero Charudatta as both a Brahmana and a sarthavaha (merchant), and a fifth-century inscription describes two brothers who funded a temple as kshatriya-vaniks, Kshatriya merchants, a hybrid label the four-fold varna scheme has no room for.

Beyond the varna framework altogether sat populations whose practices were not shaped by Brahmanical ideas: forest-dwelling hunting communities such as the nishadas (the community Ekalavya is said to have belonged to), nomadic pastoralists who did not fit a settled-agriculturalist framework, and speakers of non-Sanskritic languages, often labelled mlechchhas and looked down upon. A sharper division classed certain groups, designated chandalas, as "untouchable", on the ground that handling corpses and dead animals was maximally "polluting". The Manusmriti laid down that chandalas had to live outside the village, use only discarded utensils, wear the clothes of the dead, and could not walk about villages or cities at night; they also disposed of unclaimed bodies and served as executioners. Foreign travellers corroborate this from outside the Brahmanical tradition itself: the Chinese Buddhist pilgrim Fa Xian (c. fifth century CE) recorded that "untouchables" had to sound a clapper in the streets so others could avoid seeing them, and Xuan Zang (c. seventh century) observed executioners and scavengers forced to live outside the city.

Quick revision points

• The Mahabharata is an itihasa ("thus it was"), composed over about 1,000 years from c. 500 BCE onwards, growing from perhaps under 10,000 verses to about 100,000 verses; traditionally attributed to the sage Vyasa.
• Composition stages: charioteer-sutas composed and orally transmitted the original story; Brahmanas began committing it to writing from the fifth century BCE; a Vishnu-Krishna phase followed (c. 200 BCE-200 CE); large Manusmriti-like didactic sections were added last (c. 200-400 CE).
• Whether the epic preserves a real historical war is explicitly unsettled in the NCERT text; B.B. Lal's 1951-52 Hastinapura excavation is suggestive of the site, not proof of the epic's events.
• The Critical Edition of the Mahabharata: begun 1919 under V.S. Sukthankar, took 47 years, ran to over 13,000 pages, documenting both a common textual core and enormous regional variation.
• Patriliny (father to son descent) versus matriliny (through the mother); the epic's own plot reinforces patrilineal succession after the Kaurava-Pandava war.
• Gotra: named after a Vedic seer, from c. 1000 BCE; two rules, gotra exogamy (no marriage within gotra) and women adopting the husband's gotra, both contradicted by real Satavahana queens who retained their father's gotra and sometimes married within it (endogamy).
• Draupadi's marriage to the five Pandavas: the exam-set example of polyandry, justified twice over by Vyasa (Indra-incarnation story, Shiva-boon story); historians read the layered justification as a sign polyandry was already falling out of favour, not proof it was typical.
• Varna was fixed at four categories, ideologically birth-based per the Dharmashastras but contested even for kings (the Mauryas' Kshatriya status was itself disputed).
• Jati had no numerical ceiling and multiplied with new occupations and migrant groups (the Mandasor silk-weavers' guild inscription); jati and shreni membership did not lock a person into one occupation or align cleanly with varna.
• Chandalas sat outside the varna order altogether as "untouchable", their Manusmriti-prescribed exclusions corroborated by the foreign travellers Fa Xian and Xuan Zang.

Hold the sequence of the Mahabharata's composition, the specific gotra and exogamy rules, and the varna-jati distinction firmly, and the statement-based traps this chapter is built around stop working.