The caste system: what Hinduism says

One English word for two Indian things

Almost every confused argument about “caste in Hinduism” begins with a single translation error. The English word caste comes from the Portuguese casta, “breed” or “lineage,” a word the Portuguese used in the sixteenth century to describe a social order they did not understand. It has been flattening two very different Indian ideas into one ever since.

The first idea is varna: a fourfold theological model of society, four broad classes, described in scripture. The second is jati: the thousands of actual birth-groups, tied to occupation and region, into which Indians are in fact born, marry, and live. When a news report or a textbook says “the caste system,” it usually means jati but quotes scripture about varna, and the two do not line up. Getting the distinction clear is not a technicality. It is the whole key to the question, because the tradition’s texts, its history, and its critics all say different things depending on which one you mean.

So this is a piece about both: what the texts actually say, how a flexible ideal became a rigid hierarchy, and how the tradition has argued against itself on this question for as long as the question has existed.

Varna: the fourfold ideal

The word varna means “colour,” and by extension “class” or “type.” The classical model names four. Brahmins are the priests and teachers, keepers of learning and ritual. Kshatriyas are the rulers and warriors, charged with protection and justice. Vaishyas are the producers, farmers, herders, and traders. Shudras are those who serve and labour. Four functions, imagined as the working parts of a single social body.

That bodily image is not a metaphor added later. It is there in the founding text. The Rigveda, in the late hymn called the Purusha Sukta (10.90), describes the cosmos as born from the sacrifice of a primordial being, and Griffith’s public-domain translation renders the crucial verse:

“The Brahman was his mouth, of both his arms was the Rajanya made. His thighs became the Vaishya, from his feet the Shudra was produced.”

It is worth noticing two things at once. The image is organic: mouth, arms, thighs, feet are parts of one living body, each needed, none complete alone, which is a picture of interdependence. And the hymn is late; Mandala 10 is among the youngest layers of the Rigveda, so even within the oldest scripture the fourfold scheme arrives near the end, not at the beginning. The idea of varna is ancient, but it was not the Vedic religion’s first word.

Was varna ever meant to be by birth?

Here the tradition’s own texts pull in more than one direction, and honesty requires holding the tension rather than resolving it too quickly.

The single most quoted verse on the question is Bhagavad Gita 4.13, where Krishna says: chaturvarnyam maya srishtam guna-karma-vibhagashah, “the fourfold order was created by me according to guna and karma,” that is, according to quality and action. The verse does not mention birth. Read on its own, it makes varna a classification of temperament and vocation, the way one might sort people by aptitude, and many reformers have leaned on exactly this line to argue that true varna has nothing to do with heredity. The Gita reinforces this elsewhere by describing the four classes through their innate qualities rather than their parentage.

But the same text also assumes the settled society of its day. Krishna urges Arjuna to do his own sva-dharma, the duty proper to him, and Arjuna’s own anguish at the start of the poem includes a fear of varna-sankara, “the confusion of varnas,” the social chaos he thinks war will bring. Readers who want the Gita to endorse the hereditary order point here. So the fair summary is that the Gita grounds varna in quality and action, while taking for granted the inherited social world around it. Both threads are in the text, and interpreters have pulled on one or the other for two thousand years. The larger idea of duty, dharma, that frames all this is treated in Dharma, Karma and Rebirth.

Jati: caste as actually lived

Whatever the theology says about four varnas, no one has ever actually lived in a world of only four groups. The lived reality is jati.

Jati, from a root meaning “birth,” refers to the thousands of distinct communities that make up Indian society on the ground. A jati is typically hereditary, linked to a traditional occupation, tied to a region and language, and, most consequentially, endogamous: people marry within it. There are jatis of weavers, potters, scribes, cultivators, merchants, priests, and countless others, and their local ranking and relations vary enormously from region to region. This, and not the tidy fourfold varna, is what governs the marriages, the community networks, and much of the discrimination that the word “caste” points to in real life.

The relationship between the two is loose and much argued over. Traditionally, jatis were mapped onto the varna scheme, each local community assigned a place as broadly Brahmin, Kshatriya, Vaishya, or Shudra, but the fit was always rough, contested, and regionally variable. Varna is the theological map. Jati is the territory, and the territory is vastly more complicated than the map.

The ones left outside

The cruelest feature of the system is not inside the four varnas at all. It is the exclusion of groups placed entirely outside them: the communities historically stigmatised as “untouchable,” assigned the work a purity-obsessed society deemed polluting, and subjected to systematic humiliation, segregation, and violence. Today many of these communities identify as Dalit, a word meaning “broken” or “oppressed,” chosen as a name of dignity and resistance.

It must be said plainly that this exclusion has no warrant in the earliest scripture and represents one of the gravest injustices carried under the name of religion. It was denounced from within the tradition by saints and reformers, and it was made illegal at the birth of the modern Indian republic: Article 17 of the Constitution of India, adopted in 1949 and in force from 26 January 1950, abolishes untouchability and forbids its practice in any form. Law, of course, is not the same as lived reality, and caste discrimination persists. But the direction of the tradition’s own moral argument has long run against the exclusion.

How the hierarchy hardened

If the Gita can read varna as quality and action, where did the rigid, birth-locked hierarchy come from? Part of the answer is the Dharmashastra literature, the law books, and one text above all.

The Manusmriti, the “Laws of Manu,” composed sometime between roughly the second century BCE and the third century CE, sets out an elaborate, birth-based ranking of the varnas, with detailed and often harsh rules subordinating Shudras and women. It is the text most often quoted, by admirers and critics alike, as “the Hindu law of caste.” But its status needs care. It is one smriti, a remembered human composition, among many such law books that disagree with each other; it was never a single scripture binding on all Hindus everywhere; and its authority has always been contested. Its rules describe an ideal that particular Brahmin jurists wished to impose, which is not the same as a lived constitution that all of India obeyed.

Then comes a more modern and, to many, more surprising part of the story. A large body of scholarship argues that caste as a single, rigid, nationwide system is substantially a product of the colonial encounter. In Castes of Mind, the historian Nicholas Dirks argues that under British rule caste “became a single term capable of naming and … subsuming India’s diverse forms of social identity,” and that the decennial census from 1872 onward, by forcing every Indian to record one fixed caste, helped freeze fluid, local, negotiable identities into hard bureaucratic categories. Susan Bayly’s careful history traces how caste’s meaning shifted across the eighteenth and nineteenth centuries. Dirks is explicit that the British did not invent caste out of nothing; rather, colonial administration, law, and ethnography hardened and systematised what had been far more varied and flexible. The strong version of this thesis is itself contested: historians such as Sumit Guha and Norbert Peabody argue that pre-colonial Indian states already counted and enforced caste, and Bayly’s own account treats colonialism as reshaping rather than inventing it. Louis Dumont’s classic Homo Hierarchicus influentially read the whole system through a single logic of ritual purity and pollution, a reading that later scholars such as Declan Quigley sharply criticised for imposing one theory on a messy reality. The scholarly picture, in short, is that caste is neither timeless nor simple, and that its most rigid modern form has recent as well as ancient roots.

The tradition against itself

The most important thing missing from most outside accounts is that Hinduism has always contained its own fierce opponents of caste. The critique is not a modern import. It is internal, old, and continuous.

The bhakti movements, the traditions of loving devotion that swept across India from the medieval centuries, again and again declared caste irrelevant before God. Their poet-saints included men and women from across the social order, and some of the most revered were from the very communities the system despised. Ravidas, a saint from a leather-working community, sang of a casteless city of God. Basava, in twelfth-century Karnataka, founded a community that rejected caste and ritual hierarchy outright. Kabir mocked the pretensions of both Brahmin and mullah. The devotional logic was simple and radical: if the one reality receives all who love it, then birth cannot rank souls. That current is surveyed in The Bhakti Saints.

In the modern era the critique became a political and constitutional force, above all through B.R. Ambedkar, himself born into an untouchable community, who became independent India’s foremost jurist and the architect of its Constitution. On 25 December 1927, during the Mahad Satyagraha, a campaign for the simple right of Dalits to draw water from a public tank, Ambedkar publicly burned a copy of the Manusmriti as a rejection of the religious sanction for untouchability. In his classic text Annihilation of Caste he argued that caste had to be abolished rather than reformed. Whether one reads him as a critic from within Hinduism or from outside it, his argument is now part of the tradition’s own long quarrel with caste.

What the honest answer sounds like

So, does Hinduism have a caste system? The truthful answer refuses both the apologist’s denial and the caricature.

Hinduism supplies the concept of varna, a fourfold model of society that its own highest texts tie to quality and action rather than birth, and that its founding hymn imagines as the interdependent parts of one body. Layered over this, the lived institution of jati, with its thousands of hereditary communities and its cruel exclusion of those placed outside, grew into the hierarchy the world calls caste, hardened by particular law books and, many scholars argue, by colonial systematisation. And running through the whole history is the tradition’s own dissent, from the bhakti saints to Ambedkar, insisting that birth cannot measure the worth of a soul.

To understand caste honestly is to hold all of that at once: an ideal, an injustice, and an argument the tradition has never stopped having with itself. The dignity of the human being, and the claim that the same reality dwells in every one of them, is not a modern gloss on the tradition. It is one of the oldest things the tradition says, and it is the ground on which its own reformers have always stood. The wider question of what “Hinduism” even is, and why it resists single answers like this one, is taken up in What is Sanatana Dharma?.