The 2017 revision removed four pages on the Bengal famine from the class eight textbook. The stated reason, in the committee minutes, is that the chapter ran long and the material was covered in class ten.
I have read the class ten chapter. It devotes two paragraphs to the famine, both of which describe it as a consequence of wartime disruption, and neither of which mentions the export of rice from Bengal during it, or the decision-making in London, or the term "denial policy". So the material was not covered in class ten. It was removed from class eight and a sentence was written down explaining that it had moved somewhere it did not move to.
This needs saying carefully, because there is a genre of writing about Indian textbook revision that I find unserious — the genre in which every edit is an ideological outrage and the writer's own preferred emphasis is simply history. Textbooks are revised constantly, by every government, in every country, and most revisions are exactly what they claim to be: reorganisation, updating, cutting for room. A committee that shortens a chapter is not by that fact engaged in a conspiracy, and treating it as such makes it impossible to identify the times when something else is happening.
So the question I care about is methodological: how do you tell an ordinary edit from a substantive one, when the stated reason is always ordinary?
The answer I have arrived at, after fifteen years of reading these documents, is that you cannot tell from the edit. You tell from the pattern of what survives.
Consider the same 2017 revision. Alongside the famine cut, the chapter on the Mughal administrative system lost about a third of its length. The chapter on the Vijayanagara empire gained. Regional revolts against colonial rule gained substantially. A section on Partition violence was rewritten in a way that reduced the number of named perpetrators on both sides from six to two.
Any one of those, alone, is defensible. Chapters do run long. Vijayanagara was underweighted for decades and a correction was overdue — I have argued for that correction myself, in print, before this committee existed. The Partition rewrite could plausibly be about age-appropriateness, which is a real consideration for thirteen-year-olds.
But look at the set. What was cut is material in which the suffering was caused by an authority. What was added is material in which the suffering was inflicted by an outside force and resisted. That is a shape, and a shape is visible where an individual edit is not.
This is the same evidentiary problem as in any archive: a single silence proves nothing, and a pattern of silences across independent decisions is evidence about the decision rule. The difference is that here the decisions were made by named people in recorded meetings, which ought to make it easier and in practice does not, because the minutes record the stated reason and the stated reason is always length.
Now the complication, which is the part I think is genuinely hard and which most writing on this subject ducks.
The previous revision, in 2006, under a different government, did the same kind of thing in the opposite direction. Chapters on caste in pre-colonial society expanded considerably. Material on the composite culture of the Mughal court gained. The treatment of the independence movement was rebalanced towards figures the incoming government's tradition claimed and away from others. The committee minutes of 2006, which I have also read, cite pedagogical modernisation and the removal of dated material.
I supported most of the 2006 changes. I think several of them made the textbooks better history. And I have to be honest that if I apply my own pattern test to that revision, it produces the same finding: a shape, a set of edits pointing in one direction, and a stated reason that does not explain the pattern.
That is uncomfortable and I do not think the discomfort should be resolved. The temptation is to say that 2006 was correction and 2017 is distortion, that one moved towards the evidence and the other away. Sometimes I do think that. But the test I have proposed is a test of shape, not of truth, and it flags them both — and a test I only apply when I dislike the result is not a test.
What I have settled on is narrower and more useful than a verdict. A textbook committee should be required to state, for every removal, where the material now appears — with a page reference. Not a justification, not a philosophy. A page number, checkable by anybody in an afternoon.
That single requirement would have caught the famine cut immediately, because there is no page to point at. It would have been satisfied honestly by most ordinary edits, because most ordinary edits really do move material. It does not require anyone to agree about history, or about emphasis, or about what thirteen-year-olds can handle. It requires the stated reason to be true.
I have proposed this in two consultations. It has not been adopted, and the objection raised both times was that it would be administratively burdensome — which is, I note, an argument about length.
