The Bengal Presidency Sati Returns

From 1815 the East India Company required its district officers in the Bengal Presidency to report every case of widow immolation to the Nizamat Adalat. The returns ran until the practice was prohibited in 1829, and they record 8,134 cases across those fourteen years.

That number is quoted constantly and is almost always quoted wrongly, as though it were the rate at which sati occurred in India. It is not. It is the number of cases that a particular administration, in one presidency, over fourteen years, managed to hear about and chose to write down.

The internal evidence that the series is a record of reporting rather than of practice is in the geography. More than half the recorded cases fall in the Calcutta division, in and around the seat of the administration doing the counting, and that is not where the practice is otherwise understood to have been concentrated. Two readings are possible, and both are held: that the division genuinely had an unusually high incidence, or that the returns are densest where the Company had the most officers and the strongest interest in the answer.

The returns also had a purpose. They were compiled while the question of prohibition was live, by an administration under pressure from evangelical opinion at home and from Indian reformers in Calcutta, and they became the evidence on which the case for prohibition was argued. A count assembled inside an argument is not thereby false, but it is not an innocent measurement either.

The question people actually want answered is how common sati was in India before anyone was counting. The returns cannot answer it, and neither can anything else that currently exists.

This is worth stating as a gap rather than as a controversy, because it is usually presented as one. There are loud positions in circulation. One holds that the practice was rare, largely confined to particular castes and regions, and inflated by colonial reformers who needed a barbarism to abolish. Another holds that it was widespread and that the colonial record understates it. Both positions are argued, and both are arguing about the interpretation of a fourteen-year series from one presidency, because there is nothing else to argue about. Neither has an independent measurement, and there is no series to compare the returns with.

What could change that is not a better reading of the returns. It is other evidence entirely. Memorial stones for immolated women survive in numbers across western and southern India and can in principle be dated and counted regionally; vernacular land, temple and family records exist for periods and places the Company never reached; and both bodies of material have been used locally without ever being assembled into a demographic picture.

Until somebody does that work, the honest statement about the frequency of sati before the nineteenth century is that it is not known, and that the tools to find out have not been picked up rather than tried and failed.