Ambedkar - Caste as Closure, and the Argument with Gandhi

Bhimrao Ramji Ambedkar (1891–1956) produced the best theory of caste anyone has written, drafted the constitution of the world's largest democracy, and is almost entirely absent from sociology's theoretical canon.

His 1916 paper, written at twenty-five, is a piece of structural sociology of the first rank. This course has already used it twice (4.6.3, 6.2.1). This lesson gives it the treatment it should have had in Part 4.

A boy, a bullock cart, and a station.

Goregaon, 1901. Ambedkar is nine. He and his brothers are travelling to meet their father, a retired army subedar. They arrive at the station well dressed, in new shirts, with silk-bordered caps.

The station master is friendly until he asks their caste. They are Mahars — untouchable.

His manner changes. No cart driver will take them; the drivers would have to purify the cart afterwards. Eventually one agrees, at double the fare, on the condition that the boys drive it themselves and he walks alongside, so he need not touch them or their belongings.

They travel through the night. They have no water — nobody will give them any, because they may not draw from a well.

Ambedkar wrote this down decades later, in an unfinished autobiographical fragment, Waiting for a Visa. He wrote several such episodes: being made to sit apart on a piece of sacking in a classroom, which he had to carry himself because the school's servant would not touch it; being unable to learn Sanskrit because a Brahmin teacher would not teach him; being driven out of lodgings in Baroda in 1917 after taking a post in the state service — at which point he held a doctorate from Columbia and had studied at the London School of Economics.

Hold those two facts together, because they are the whole of his intellectual position: a man with a Columbia PhD could not obtain a room, and no amount of individual achievement altered it.

Which is why his analysis is structural. He was not theorising about a hierarchy he had read about. He was explaining a mechanism that had operated on him at every stage of a life spent, by any measure, surpassing everyone applying it.

Ambedkar's question, posed in 1916 to a Columbia anthropology seminar, is a genuinely theoretical one, and it is worth appreciating how good it is.

The available explanations of caste were three, and he demolishes all three.

The racial theory — Risley's, and the standard scholarly view (6.1.1) — held that caste reflects the mixture of Aryan and pre-Aryan populations. Ambedkar's objection: the anthropometric evidence does not support it; populations within a region resemble each other far more than caste-fellows in different regions do; and India is "a composite of races which have long intermingled."

The occupational theory — caste is a division of labour that hardened. Objection: division of labour exists everywhere without producing caste, and caste persists when occupations change. The occupation is not doing the work.

The religious theory — caste follows from Hindu doctrine. Objection: this explains the justification and not the mechanism, and it cannot explain why caste exists among Indian Muslims, Christians and Sikhs, whose doctrines forbid it.

So what is the mechanism?

Ambedkar's answer, and it is a single move: caste is not a division of labour but a division of labourers, and what divides them is marriage.

"Caste is an enclosed class."

Everything else follows.

The mechanism: endogamy and its enforcement

Ambedkar's argument in "Castes in India: Their Mechanism, Genesis and Development" (1916):

Indian society was originally a plurality of groups with permeable boundaries — and, crucially, the same cultural stock. He is emphatic that India is culturally homogeneous, which is what makes the divisions require explanation.

A group closes itself by prohibiting marriage outside. That is the whole operation. Endogamy is not one feature of caste among several; it is caste. "Exogamy is a creed and none dares to violate it... superimposed on exogamy is endogamy, and it is this superimposition that creates caste."

And here is the analytical insight that makes the paper important: a closed group faces a structural problem that endogamy itself creates.

If a group must marry only within itself, it must maintain parity of the sexes. And death disrupts it. A husband who dies leaves a surplus woman; a wife who dies leaves a surplus man. Each is a threat to the boundary, because each is available to marry outside it.

Ambedkar's argument is that the three practices most associated with caste oppression are solutions to this problem:

For the surplus widow: sati (burning her) or enforced widowhood (making her socially dead — head shaved, ornaments removed, excluded from festivity, unmarriageable).

For the surplus widower: child marriage — marrying him to a girl not yet of an age to be a competitor for other women.

And the asymmetry is the point. Burning is available for women and not for men, because "man is the master" — the group will not destroy its own men. So the woman is killed or made socially dead; the man is remarried to a child.

Assess this argument honestly, because it is bold and its status matters.

What is powerful: it is a functional argument with a specified mechanism (5.1.3's requirement), it explains why three apparently unrelated practices cluster, and it treats caste as something continuously done rather than something that is. It also, remarkably, makes the position of women structurally central to the analysis of caste — in 1916.

What is contested: historians have questioned the account's chronology and the claim that these practices originated as solutions to the surplus problem rather than acquiring that function later. Sati's actual distribution was geographically limited and its history is entangled with colonial legal debate (4.2.4, Lata Mani). And the "imitation" mechanism by which Ambedkar says the practice spread from Brahmins downwards is thinner than the rest.

What is not in dispute: that endogamy is the reproductive mechanism of caste, that it is enforced, and that the enforcement falls overwhelmingly on women. Contemporary evidence supports this precisely: inter-caste marriage rates remain low; violence against inter-caste couples is documented and continuing; and the policing of women's marriage choices is where caste boundaries are defended most fiercely.

And the theoretical priority is real. This is social closure (4.3.6) with a specified mechanism, published in 1916, sixty-three years before Parkin. Weber's account of caste — from texts, without fieldwork, in a library in Heidelberg — was the standard sociological reference throughout that period.

Annihilation, not reform

Annihilation of Caste (1936) is his most important political text, and it has an extraordinary history.

It was written as a presidential address to the Jat-Pat Todak Mandal, a Lahore organisation of caste-Hindu reformers dedicated to breaking caste.

They read the text and cancelled the conference. They objected to his argument that the Vedas and Shastras must be repudiated, and to his announcement that he intended to leave Hinduism. He published it himself, at his own expense, along with the correspondence.

The argument, in outline:

Caste cannot be reformed by economic or political means alone. Ambedkar addresses socialists directly: a revolution in property relations will not end caste, because caste divides the workers who would have to make it. "Can it be said that the proletariat of India, poor as it is, recognises no distinctions except that of the rich and the poor?" — a direct challenge to the Marxist position (4.1.8).

Caste is not a division of labour but a division of labourers — and, in his phrase, "a hierarchy in which the divisions of labourers are graded one above the other."

And it is not merely a division; it produces a specific moral condition. Ambedkar's sharpest formulation: "Caste has killed public spirit. Caste has destroyed the sense of public charity. Caste has made public opinion impossible. A Hindu's public is his caste. His responsibility is only to his caste."

Caste cannot be abolished by inter-dining or inter-marriage campaigns alone, because these treat symptoms. The root is the religious sanction, and while scripture authorises the hierarchy it will regenerate.

Therefore: "You must destroy the religion of the Shrutis and the Smritis. Nothing else will avail."

The most consequential disagreement in modern Indian thought, and it is usually flattened into a clash of personalities.

Gandhi's position, as expressed in the 1930s: untouchability is an evil excrescence on Hinduism and must be abolished. He called untouchables Harijans — "children of God" — and campaigned for temple entry and against untouchability throughout.

But he defended varnashrama dharma — the fourfold order — as a sound principle of social organisation, arguing that it was a division of duties rather than a hierarchy, that hereditary occupation was not degrading, and that the evil lay in untouchability rather than in caste itself.

Ambedkar's response was that this is incoherent. Untouchability is not an excrescence; it is what the graded hierarchy produces at its bottom. Retain the principle of graded status by birth and the bottom will exist. "The out-caste is a bye-product of the caste system."

He also rejected "Harijan" as patronising — a name given by others, casting Dalits as objects of religious charity rather than as claimants of rights. The self-chosen term is Dalit — "broken," "crushed" — which names a condition inflicted rather than a status bestowed.

And the Poona Pact of 1932 is where the disagreement became concrete.

The Communal Award granted separate electorates to the Depressed Classes: reserved seats elected by Dalit voters only. Ambedkar had argued for this at the Round Table Conferences, on the ground that reserved seats within a general electorate would produce representatives chosen by caste Hindus — dependent on their votes and therefore not answerable to Dalits.

Gandhi began a fast unto death against it, arguing that separate electorates would divide Hindu society permanently.

Ambedkar was placed in an impossible position. If Gandhi died, the consequences for Dalits — in a country where Gandhi was revered — would have been catastrophic. He conceded, and the Poona Pact substituted a larger number of reserved seats within joint electorates.

His own account of it is bitter: he described being compelled to save Gandhi's life at the cost of the political mechanism he judged essential.

And the analytical question remains live. Which arrangement produces better representation for a dispersed, subordinated minority: separate electorates, or reserved seats in a general electorate? The Indian evidence is genuinely mixed, and the question recurs wherever minority representation is designed.

Note what this example is doing here. It is not a story about two great men. It is a case in which a theoretical disagreement about the mechanism of caste — excrescence or product — produced a concrete institutional dispute with consequences that persist, and in which the sociologically better analysis lost.

Go deeper

The Constitution, and what he built

Ambedkar chaired the Drafting Committee of the Indian Constitution and served as independent India's first Law Minister.

What is in it, substantially through him: the abolition of untouchability (Article 17); fundamental rights including equality before law and non-discrimination; reservations in legislatures, public employment and education for Scheduled Castes and Tribes; and a strong central state, which he argued was necessary because local self-government would mean local caste domination — "the village is a sink of localism, a den of ignorance, narrow-mindedness and communalism."

That last position put him directly against Gandhi, whose vision was of village self-rule, and it is a sociological judgement: Ambedkar thought the village was where caste was most complete and escape least possible, and that a modern state and urban anonymity offered more.

And his closing speech to the Constituent Assembly (25 November 1949) contains the most quoted warning in Indian political thought:

"On the 26th of January 1950, we are going to enter into a life of contradictions. In politics we will have equality and in social and economic life we will have inequality... How long shall we continue to live this life of contradictions? How long shall we continue to deny equality in our social and economic life? If we continue to deny it for long, we will do so only by putting our political democracy in peril."

He also resigned from the cabinet in 1951, over the failure of the Hindu Code Bill — a reform of Hindu personal law on marriage, divorce, inheritance and adoption that would have substantially improved women's legal position. The connection is not incidental: his analysis made the control of women's marriage central to caste (above), so reforming family law was not a side issue.

And in 1956, weeks before his death, he converted to Buddhism at a mass ceremony in Nagpur with several hundred thousand followers — having stated in 1935 that though he had been born a Hindu he would not die one. The Navayana Buddhist movement he founded continues, and the census records millions of Buddhists in Maharashtra, almost all former Dalits.

Why he is not in the canon

The mechanisms are the ones this course has documented four times (3.3.3, 4.5.5, 4.6.1, 4.6.3), operating together.

Institutional position: he held no academic appointment. His work was published by movements, presses and the state, not by university publishers with international distribution.

Reclassification: he is filed as a political leader, a constitutional figure and a Dalit icon — not as a sociologist. The 1916 paper appears in Indian sociology syllabi and almost never in general theory courses.

Language and circulation: his major works were published in India, by small presses, and the collected writings were not assembled until the 1980s and 90s by the Government of Maharashtra.

And the standpoint problem in its sharpest form. Indian sociology was built substantially by upper-caste scholars (6.2.1). A theory of caste written by an untouchable was not read as theory by a discipline whose practitioners were the beneficiaries of what it analysed — which is 5.6.1's standpoint argument, and 4.5.5's erasure, occurring simultaneously.

And the Du Bois connection

Ambedkar wrote to W. E. B. Du Bois in 1946, having learned of the NAACP's petition to the United Nations on the condition of Black Americans, and asking for copies — noting that "there is so much similarity between the position of the Untouchables in India and of the position of the Negroes in America that the study of the latter is not only natural but necessary."

Du Bois replied that he had "every sympathy with the Untouchables of India" and sent the material.

Two of the most important social theorists of the twentieth century, both excluded from the canon of the discipline they were transforming, corresponded about the structural similarity of the systems they were analysing — and neither correspondence nor either body of work entered the mainstream for decades.

That fact is itself the argument of Part 6, and it does not require elaboration.

"Ambedkar was a political leader who also wrote." He held doctorates from Columbia and the LSE, produced a structural theory of caste at twenty-five, and wrote extensively on economics, law and religion. The politics followed the analysis.

"Caste is a division of labour." Ambedkar's central correction: it is a division of labourers, "graded one above the other," reproduced by endogamy.

"Sati and child marriage are religious customs." His argument is that they are mechanisms for maintaining endogamy by managing the surplus produced by death. Contested in its historical detail; the underlying point about the enforcement of endogamy on women is well supported.

"Gandhi opposed caste." He opposed untouchability and defended varnashrama dharma as a division of duties. Ambedkar's reply was that untouchability is the graded hierarchy's product, not an excrescence on it.

"The Poona Pact was a compromise both accepted." Ambedkar conceded under a fast unto death that would have had catastrophic consequences for Dalits if it had succeeded, and described the concession bitterly.

"He is now recognised." In India, iconically. In sociology's international theory canon, still barely — which is the point.

Why it matters

It is a theory of closure with a specified mechanism, produced sixty-three years before Parkin and in a case Weber got wrong from a library. Ambedkar belongs in Part 4, and this course put him in Part 6 for the same reasons the discipline did — which is worth stating rather than concealing.

It makes gender structurally central to caste. The argument that endogamy creates a surplus problem solved by controlling women's marriage, widowhood and death is, whatever its historical difficulties, an analysis in which the position of women is not an additional topic but the mechanism itself. In 1916.

It supplies the sharpest available critique of two positions at once: that class analysis alone can explain a system that divides the class (against Marxism, 4.1.8), and that a hierarchy can be reformed at its bottom while its principle is retained (against Gandhi).

And it is the discipline's clearest single case of what exclusion costs. The best theory of caste, written by the person the system was most completely designed to exclude, unread by a field that continued to cite an account produced without fieldwork by someone who had never been there. Every mechanism in 4.5.5 and 6.1.1 is present, and the result was a century of teaching the worse theory.

Ambedkar's 1916 paper demolishes the racial, occupational and religious explanations of caste and replaces them with a mechanism: caste is "an enclosed class", and the enclosure is endogamy. India is culturally homogeneous, which is what makes the divisions require explanation.

And endogamy creates a structural problem it must then solve: maintaining parity of the sexes against death. Sati and enforced widowhood dispose of the surplus woman; child marriage disposes of the surplus man. The asymmetry — killing available for women, remarriage for men — is the point. Contested in its historical detail; sound in its core claim that endogamy is enforced and that the enforcement falls on women.

Annihilation of Caste (1936), written for a conference that cancelled it: caste is "a division of labourers... graded one above the other"; economic revolution will not end it because caste divides the class that would make the revolution; "a Hindu's public is his caste"; and it cannot be reformed while scripture sanctions it.

Against Gandhi: untouchability is not an excrescence on Hinduism but the product of graded hierarchy — "the out-caste is a bye-product of the caste system." Harijan rejected as a name given by others; Dalit — "broken" — chosen instead. And the Poona Pact (1932), in which he conceded separate electorates under a fast unto death.

The Constitution: Article 17, reservations, a strong centre against village self-rule ("a sink of localism"), and the closing warning about "a life of contradictions" — political equality alongside social and economic inequality.

Why he is not in the canon: no academic appointment, reclassification as a political figure, publication outside international circuits, and a discipline of upper-caste scholars not reading a theory of caste by an untouchable.

And in 1946 he wrote to Du Bois about the structural similarity of the two systems. Two of the century's most important social theorists, both excluded, corresponding about it.

Caste as enclosed class — Ambedkar's definition: a class that has closed itself through endogamy.

Division of labourers — as against division of labour: caste divides workers and grades them.

Endogamy as mechanism — marriage restriction as the reproductive engine of caste.

Surplus man / surplus woman — the parity problem endogamy creates, and which sati, enforced widowhood and child marriage address.

Graded inequality — hierarchy in which each level is above one and below another, producing investment in the system by all but the lowest.

Annihilation versus reform — the argument that caste cannot survive the removal of its religious sanction.

Harijan / Dalit — the name given by others, and the name chosen.

Separate electorates / reserved seats — the institutional dispute of the Poona Pact.

Navayana Buddhism — the movement founded by Ambedkar's 1956 conversion.

One — read the 1916 paper. "Castes in India" is about twenty pages, freely available, and was written by a twenty-five-year-old. Read it as theory, which is what it is.

Two — test the closure mechanism elsewhere. Take any hereditary group system and ask what happens to marriage across its boundary. The answer will tell you how the boundary is maintained.

Three — apply graded inequality. Ambedkar's insight is that a hierarchy with many rungs is more stable than one with two, because almost everyone has someone below them. Find a system with this structure and ask who has an interest in maintaining it.

Four — check the marriage data. Look up inter-caste marriage rates in India, or intermarriage rates across any major social boundary in your own country. The rates are the measure of closure.

Five — ask why you had not read him. If you have studied social theory and not encountered Ambedkar, ask through which of 4.5.5's six mechanisms he was excluded. All of them apply.