The Census Can Collect 46 Lakh Caste Names. Who Decides What They Mean?

The Count Is Not the Problem. The Classification Always Has Been.

India has tried to count caste before and each attempt ran into trouble at the same point: after the counting, when someone had to decide what the answers meant.

A Pattern That Runs From 1901 to 2015 In 1901, Census Commissioner **Herbert Hope

Risley** did not simply record castes. He tried to rank them, sorting communities by "social precedence" into a varna-style ladder. The result was a flood of petitions from communities disputing where they had been placed.

The act of classification turned the census into a site of social struggle. By 1931, J.H. Hutton had stepped back from Risley's ranking. That census recorded 4,147 castes and it remains the last complete caste count India has.

In 1941, caste data was collected under wartime conditions and never published. In 1951, independent India stopped enumerating caste except for Scheduled Castes and Scheduled Tribes. Then came the 2011 Socio-Economic Caste Census.

It used an open field and the returns came back as over 46 lakh distinct caste names. Variant spellings, clan names, surnames and regional labels all entered as separate "castes". In July 2015, the raw data was handed to the states to consolidate.

An expert group under Arvind Panagariya was announced for the same task. It never met and the data was withheld.

Why 2027 Repeats the Risk

The open column in Census 2027 is not the flaw its critics claim. Letting people name their own identity is more honest than making them pick from a list drawn up in advance. The Bihar and Telangana surveys used lists and every list decides in advance which identities exist.

But an open column moves the real power from the enumerator to the classifier. Someone must decide whether dozens of spellings are one caste or many and whether a sub-group is counted separately or merged. Those decisions shape OBC sub-categorisation, the debate over the reservation ceiling and welfare targeting.

Officials now say discussions on this step have not started. The file will go up to the Union Home Secretary and the decision will be taken by the Ministry or the Cabinet. That is exactly the stage at which SECC 2011 stalled.

Where the Parallel Breaks

Risley was a colonial official who imposed categories from above. Today's classifier is a democratic state, answerable to Parliament and to the courts. The respondent also has more say: in 2027 people name themselves and can even choose "no caste".

So the risk now is not an imposed hierarchy. It is the opposite failure, raw data that is never made usable or classified out of sight. The census already has a working model for this. It routinely consolidates thousands of raw mother-tongue returns into a standard classification of languages.

Caste can be treated the same way, if the coding rules are published before fieldwork ends and experts, as scholar G.N. Devy has urged, are involved from the start.

For the exam: caste enumeration in India has repeatedly failed at classification, not at collection. An answer that sees this pattern, from 1901 through 1941 to 2011, will go further than one that just lists the arguments for and against a caste census.