Sati: a practice with almost no scriptural ground, condemned from inside long before it was banned

The single Rigvedic verse ever offered in defence of widow-burning tells the widow to get up and rejoin the living. Rigveda 10.18.8, in Ralph Griffith's translation, reads: "Rise, come unto the world of life, O woman: come, he is lifeless by whose side thou liest." Nothing in the hymn asks her to burn.

That is close to the whole scriptural foundation, and it points the wrong way for anyone defending the practice. Sati has no Vedic injunction behind it. It enters Sanskrit legal literature late, as one option among two, and by the seventh century Hindu writers were arguing against it in Sanskrit with no outside pressure of any kind behind them. It also happened, in real numbers, heavily concentrated in one region. It was made illegal on 4 December 1829 by a colonial administration acting after a campaign run by a Bengali Hindu. Independent India had to legislate against it again in 1987. Every one of those sentences is true at the same time, and the article that drops any of them is lying.

The funeral hymn tells her to stand up

Rigveda 10.18 is a burial hymn. Its middle verses stage a ritual: the wife lies down beside her dead husband, and is then raised and led back. Verse 7 describes women who are not widows, wearing ornament and unguent, going up first to the place where the dead man lies. Griffith renders that line "first let the dames go up to where he lieth". Verse 8 then addresses the woman on the ground directly.

"Rise, come unto the world of life, O woman: come, he is lifeless by whose side thou liest. Wifehood with this thy husband was thy portion, who took thy hand and wooed thee as a lover."

Rigveda 10.18.8, translated by Ralph T. H. Griffith

Verse 9 continues in the same direction, with the bow taken from the dead man's hand. The sequence is a symbolic descent and return. She goes down towards death, she is called back, life continues.

The whole textual argument turns on one word in verse 7, yonim, which carries the sense of seat, place or womb, and which Griffith gives as "where he lieth". It is not the word for fire, and the 2014 Oxford translation by Stephanie Jamison and Joel Brereton, now the scholarly standard, does not render it as fire either.

No other Rigvedic verse has ever been produced in support of the practice, and the Brahmana literature does not prescribe it. This matters, because the Rigveda is the layer of the tradition that carries the most authority. A practice with no footing there has to be argued for from somewhere else.

The agneh story, and why the accurate version is stronger

You will meet a claim, often repeated, that the verse was tampered with. The story goes that the word agre, meaning "first" or "forward", was altered to agneh, meaning "of the fire", and that this forgery manufactured a Vedic licence for widow-burning.

Handle this carefully, because as usually told it is wrong, and one refutable sentence undoes a page of true ones.

The Rigvedic text at 10.18.7 reads agre. There is no Rigvedic manuscript that reads agneh. This is not a disputed reading, because there is nothing to dispute. What exists is a modern argument, usually cited to the historian Anand A. Yang, that the word was mistranslated or misapplied at a much later date, perhaps as late as the sixteenth century, in the way the verse came to be quoted by others. That is a claim about later citation practice. It is not a demonstrated corruption of the Vedic text, and nobody has produced the manuscript that would prove one.

State it that way and it holds. State it as proven tampering and any reader with a search engine can flatten it in a minute, which contaminates everything accurate next to it.

The correction costs nothing, because the honest version is the better argument. The Rigveda does not need to have been forged for the pro-sati case to fail. Read exactly as transmitted, the verse calls the widow back to the world of the living.

The law books argue with each other

The Dharmashastra material is where the practice finally shows up, and it does not speak with one voice. Manu's fifth chapter deals at length with the widow and prescribes an ascetic life. In Georg Bühler's translation the key verse is 5.157: she may live on pure flowers, roots and fruit, and must never mention another man's name after her husband's death. The next verse tells her to remain patient and self-controlled until death.

A note on numbering, since it trips people up. That same verse is 5.155 in Ganganatha Jha's edition of Medhatithi's commentary. Editions disagree by a verse or two through this stretch, so any citation between 5.155 and 5.168 needs checking against the edition being used. The substance is not in doubt. Manu describes a long widowhood. Manu does not tell her to die.

The Vishnu Smriti, one of the latest works in the tradition, does something different at 25.14. In Julius Jolly's translation the widow is given two courses: to preserve her chastity, or to ascend the pile after her husband. Later digests and commentaries went further in both directions, some approving the practice and some rejecting it outright.

Notice the shape of that. Even in the text most often cited in favour, sati appears as a permission and an alternative, never a duty. A woman who chose the other path broke no rule. The Dharmashastra corpus does not appear to prescribe any penance for the widow who goes on living, a strange gap in a literature that specifies penances for almost everything else.

The contradiction is real and should not be smoothed over. The tradition holds a text that ignores the practice and a text that permits it. That is what an authentically old, unedited, internally argumentative body of literature looks like.

Bana, seventh century: a course that does the dead no good

Bana wrote at the court of Harsha in the seventh century, and he handled the subject twice, in two different registers.

In the Harshacharita he narrates it. Queen Yasovati resolves to enter the fire while her husband, the emperor Prabhakaravardhana, is still alive and dying. Her son pleads with her not to. She goes anyway. Bana is recording a court event rather than endorsing it, and the son's plea is left standing in the text. E. B. Cowell and F. W. Thomas translated the episode in 1897.

The argument proper is in his prose romance, the Kadambari, and the two books are often conflated. The passage comes when Mahasveta, having lost Pundarika, wants to follow him in death, and Candrapida talks her out of it. Two technical terms matter here: sahagamana, going with him, meaning death on the same pyre, and anumarana, dying after him at a later date. It is anumarana the passage takes apart, describing it as fruitless, as a course followed by the ignorant, as infatuation and shortsightedness rather than merit.

The reasoning is not sentimental, which is what makes it durable. Dying with the dead does the dead no good, while the survivor who lives can perform the rites that help. In C. M. Ridding's 1896 translation, the one who kills herself at another's death makes that person a sharer in the guilt. P. V. Kane, the historian of Dharmashastra, singled it out as one of the strongest condemnations of anumarana in Sanskrit.

Hold that date in mind. A Sanskrit author in a seventh-century Indian court, writing for a Hindu audience under a Hindu emperor, put a sustained case against widow-burning into the mouth of a sympathetic character and let it win the argument. That is roughly twelve hundred years before Bentinck.

Medhatithi: permitted and still forbidden

Two centuries or so after Bana, Medhatithi wrote his commentary on Manu, probably in ninth-century Kashmir. He is among the earliest surviving commentators on the text and the most influential of them, and he argues against sati directly.

His reasoning is technical and worth following, because it is the kind of argument that only comes from inside a tradition. Sati, he says, is suicide, and killing oneself after one's husband is forbidden to a woman. He grounds this in a Vedic prohibition against departing before one's natural span, and holds that a shruti prohibition outranks any later permission.

Then he makes the sharper move. He compares the case to the shyena sacrifice, the rite performed to destroy an enemy. The shyena exists in the ritual literature. It has a procedure, and the procedure works. It is still transgressive, and performing it is still a sin. Scripture describing how a thing is done is not scripture telling you to do it. On Medhatithi's reading, a widow who burns is not acting according to the shastra even where a text has described the act.

He also notices the material situation. He reads the provision as bearing on widows who are childless and without property, women with nothing to live on and every pressure on them, and answers that such a woman should support herself by spinning or similar work, not burn.

It happened anyway, and the map is the interesting part

None of the above stopped it. The earliest surviving reports are foreign and literary. Aristobulus of Cassandreia, who travelled with Alexander's expedition around 326 BCE and survives in fragments quoted by Strabo, describes wives burning with their husbands among the Kathaioi of the Punjab. Diodorus records a case after the battle of Gabiene in 316 BCE: the two widows of an Indian officer in Eumenes' army competed for the right to burn, and the Greek commanders settled it in favour of the younger, the elder being pregnant.

The earliest evidence from inside India is not literary but inscribed: the Eran pillar inscription of 510 CE, from the reign of Bhanugupta, which records that the warrior Goparaja fell in a great battle and that his wife, clinging to him, entered the mass of fire. The inscription does not name the enemy, and the common association with the Hunas is a scholarly inference rather than something the stone says. From the medieval period there are memorial stones across Rajasthan, Karnataka and elsewhere.

Then the colonial administration started counting, and the counting is where the argument about scale gets settled. Between 1815 and 1828, official returns in the Bengal Presidency recorded 8,134 cases. In the ten years from 1815 to 1824, of roughly 6,632 recorded across the Bengal, Bombay and Madras presidencies, about 5,997 were in Bengal, a little over ninety per cent.

Within the Bengal Presidency the distribution was equally lopsided. For 1815 to 1828 the Calcutta Division accounted for around 5,099 of the 8,134 cases, with Benares at about 1,165, Patna about 709, Dacca about 610, Murshidabad about 260 and Bareilly about 193. A practice supposedly commanded by universal Hindu scripture was, in practice, close to a regional phenomenon with a local epicentre.

These are administrative figures collected by officials who had their own reasons for collecting them. They almost certainly miss cases in remote districts, and the Calcutta Division returns were compiled by a denser, better-staffed administration than anything in Bombay or Madras, over a much larger population. Treat the numbers as an order of magnitude, not a census. Even so, the concentration is too extreme to be an artefact of who was counting.

One structural fact belongs next to the map, without overreading it. Bengal followed the Dayabhaga school of inheritance, under which a widow could hold her husband's share, where the Mitakshara regions covering most of the rest of India gave her no such claim. The suggestion that this created a material interest in her death is old and plausible and has never been established. Label it a hypothesis every time it is repeated. What the numbers establish on their own is narrower and still decisive: the practice was not evenly distributed, and a scriptural command would have been.

4 December 1829

Rammohan Roy published his first tract against the practice in 1818 and a second in 1820, both cast as dialogues between a defender and an opponent. His method was to fight on the ground his opponents had chosen. He argued from the Sanskrit texts themselves that the practice had no compulsory basis, that the shastras favoured the ascetic widow, and that a permission granted for the sake of a promised reward is inferior to a duty performed without one. In the second tract he widened it, arguing that the supposed weakness of women was a product of how they were raised and educated rather than of their nature.

On 4 December 1829, Lord William Bentinck, Governor-General in Council, promulgated Regulation XVII, which declared the practice illegal and punishable by the criminal courts. Those who abetted it were made liable to fine or imprisonment, and in cases involving force or compulsion, liable as for culpable homicide. The Madras and Bombay presidencies passed their own versions in 1830.

Here is the part that cannot be dressed up.

Roy campaigned. Bentinck signed. The law that ended sati in British India was an act of a colonial government, not an internal reform of Hindu society, and no honest account of 1829 can make it one. What the record supports is narrower and still worth saying: the intellectual case against the practice was made in Sanskrit centuries earlier, and the campaign that produced the statute was led by a Hindu arguing from Hindu texts. The pen was Indian. The signature was not.

The opposition was also Hindu. Radhakanta Deb founded the Dharma Sabha in Calcutta in 1830 specifically to resist the regulation, and it petitioned the Privy Council in London, arguing that the Company had broken its undertaking not to interfere in religion. The appeal was rejected in 1832 and the ban held.

So the picture is not colonisers against Hindus. It is Hindus against Hindus, with a foreign government holding the pen and deciding the outcome by statute. That is a worse story for everybody involved, which is a reasonable sign that it is the true one.

Deorala, 1987

On 4 September 1987, in the village of Deorala in Sikar district, Rajasthan, an eighteen-year-old woman named Roop Kanwar died on her husband's funeral pyre. Whether she went willingly was contested from the first day and has never been resolved to anyone's satisfaction.

What followed mattered as much as the death. A shrine went up, crowds came, and a chunari ceremony was organised at the site. The Rajasthan government issued an ordinance on 1 October 1987, and Parliament passed the Commission of Sati (Prevention) Act, 1987, extending to the whole country. Prosecutions ran on for decades, through separate sets of accused and separate courts, and ended in acquittals, the last of them concluding in 2024, thirty-seven years after the death.

The most revealing feature of the 1987 Act is not the section criminalising the act itself, which was already homicide under the general law. It is the section against glorification: against the temples, the processions, the collection of offerings, the honouring of a woman who has died this way.

That is a legislature identifying where the practice actually lived. Not in a verse, since the verse said the opposite. Not in Manu, who prescribed a long widowhood. It lived in what a village did the following week, in the shrine, the donation box and the crowd. Every Deorala case brought under that section and carried to a verdict ended in acquittal.