Independent India conducted its first census in 1951. However this was not novel given that census exercises had been conducted every decade starting from 1870s. 1931 was the last "proper" census conducted in British India, as the one in 1941 was badly hampered by war

1931 census is also very significant and worthy of study, as it was the last time a census exercise examined India through the prism of caste and community - dimensions which have been taboo in census reports post 1951.

So when we read the 1951 report - we are told the Indian literacy rate was ~18%. But we do not know how to interpret that number. How did that number vary by caste, and religion, geography and gender, or even a combination of all these factors.

To get insights into how Indian literacy and education levels varied by these factors the best we can do is to dig deep into the 1931 All India census - a remarkable document with a wealth of information that has been denied to us post 1951 for better or worse.

Now let's look at the 1931 document. While the document is rich in many ways, we will focus specifically on data pertaining to education and literacy - and how it varies across three important dimensions - A) Region B) Caste / Community C) Gender

Before we start looking at literacy rates, we must first understand how the report in 1931 defined "literacy" -

As per this report a literate person was someone who can write a letter, and also read an answer to it, in ANY language

So this naturally excluded people who could just sign their name and not much else. It also excluded for eg : some Muslims who were literate enough to read Quran for religious purposes, but were illiterate for all other practical purposes

Also the literacy rates were published by using the population aged 5 and above as the denominator. So children below the age of 5 would not count among illiterates.

So based on these qualifiers here's the literacy rate in 1931!

British India that year had 28.1 million literates out of a population of 296 million aged 5 or more. That translates to a "literacy" rate of 9.5%.

How did this look by gender? The literacy rate was 15.5% among men and 3.0% among women.

Now this was a huge gap that existed across the board, regardless of how we slice the data going forward by region or caste.

And how do these figures compare with the 1951 census?

In 1951, the literacy rate was said to be ~18%. So it would seem the rate almost doubled from 9.5% to 18% between '31 and '51. Alternatively it is possible the 1951 "definition" of literacy was laxer (worth investigating)

Next we turn our attention to region. How did literacy rates vary by region? Short answer: it varied a LOT. Eg : The most literate part of India was Burma at 36.8% followed by Cochin at 33.7% while J&K & Sikkim languished at 4% and 3.5% respectively.

Here's a quick look at the literacy rates in some major regions

Travancore: 28.9%
Baroda: 20.9%
Delhi: 16.3%
Bengal Presidency: 11.1%
Bombay Presidency:10.8%
Madras Presidency: 10.8%
Punjab: 6.3%
United Provinces: 5.5%
Bihar and Orissa: 5.3%
Hyderabad : 5%
J&K : 4%

So obviously region mattered. But the point to note is that the differential is less than what we would come to expect

Eg : A state like Madras was less than 2 times more literate than United Provinces (modern UP). And 9 in 10 people in Madras presidency were illiterate!

Another point to note is the relatively high literacy (25%+) in Travancore, and Cochin states, regions that were governed by princes (and hence not technically part of British Raj) - and which roughly correspond to modern Kerala

For several decades now we have remarked on high Kerala literacy . But this proves it has NOTHING to do with Kerala Communist / Congress govts post independence, but that Kerala had a head start even as early as 1931.

If anything the credit must go to the princely states!

Now let's turn to caste. This is the MOST interesting part of this census. As it was the last census that tracked literacy by caste.

Today we have a very elaborate caste-based reservation system in India. Yet educational outcomes are not tracked by caste. And that's unfortunate.

So the best we can do is to go back to 1931, and see how communities stood in comparison to each other

Did the "Brahmins" enjoy a huge edge in education, as often supposed, on account of their assumed "cultural capital"? And how did this vary by region? Questions worth asking.

Before we turn to caste, let's first be done with "religion". Here are the literacy rates by religion. It is clear that the two major religions were equally illiterate.

Parsis: 79.1%
Jain: 35.3%
Christian: 27.9%
Sikh: 9.1%
Hindu: 8.4%
Muslim: 8.4%

The religion grouping is not very useful, particularly when it comes to Hindus - given that different castes were at different levels of attainment and this varied by region. So looking at the numbers by caste becomes very important.

The 1931 census publishes the caste specific literacy rates by province, but it does not aggregate the castes across province. So there is no single all India brahmin literacy rate available.

But by weighting a bunch of numbers at my end I estimate the Brahmin literacy rate in 1931 to be - 29%

So the brahmins as an all India grouping, supposedly the highest varna in the Indian social system were atleast 70% illiterate in 1931.

Sure, 29% is about 3 times the 9.5% country average. But still it is much lower than the state with the lowest literacy rate even as late as 1991.

Also was the brahmin literacy rate uniform across regions? No.
Here's an examination of brahmin literacy by province in '31 -

Mysore: 56%
Madras: 54%
Baroda: 53%
Bombay: 50%
Bengal: 43%
Bihar/Orissa: 19%
United Provinces: 16%
Punjab: 15%
Rajputana: 13%
Gwalior: 12%

So the interesting thing here is that in the North Indian plain, which was home to over 50% of all brahmins back then (as now), the literacy rate was barely 15%! Not even 2 times the national average literacy rate of 9.5%!!

So while it is all well to talk of the high "social" status of brahmins, the reality was that this was a group with no great edge.

While the brahmin excellence in Madras and Mysore was very real, in most other provinces there were several communities that were running it close.

An interesting thing that you may have noted by now is that the 16% literacy of Brahmins in United Provinces (modern UP) was actually way lower than the ~30% literacy of the entire provinces of Travancore and Cochin (with NO caste filter applied!)

To understand the interplay of caste and geography better, let's look at caste specific rates within each geography.

Let's first pick Madras and Mysore provinces, areas where the brahmin dominance was the greatest and least challenged.

Madras province :

Brahmins : 54%
Nair : 44%
Chettiar : 28%
Kallar : 12%
Vanniyar : 8%
Mala : 1%

Mysore :
Brahmins: 56%
Lingayat: 16%
Vokkaliga: 7%
Kuruba: 3.5%

Clearly there was an edge that brahmins had in these provinces. Not true in others. Now let's look at Bengal.

Now here's Bengal Presidency for a few select communities:

Brahmins : 43%
Baidya : 62%!!
Kayastha : 39%
Baishnab : 16%
Jogi: 14%
Santal : 1%

So the community topping the list for Bengal is not brahmin but Baidya! And Kayasths are also pretty close!

Now let's look at UP

In UP, the story is totally different -

Brahmins : 16%
Kayastha : 45%
Sayid Muslims : 24%
Rajput : 8.5%
Jat: 4.5%
Kurmi : 2.7%
Chamar : 0.3%

So here brahmin literacy is less than 2 times the national avg. The ritually "low" Kayasths, were much more literate. Almost 3 times!

The story is very similar in North West India. Here are the numbers for Punjab -

Khatri: 27%
Baniya: 26%
Arora: 21%
Brahmins : 15%
Rajput : 5%
Jat : 3%
Ahir: 1.6%

So in NW, the merchant castes were way more literate. Brahmins were not even a close second!

The story that is emerging from these numbers is that brahmins were a dominant group in terms of education in the South, and to some extent the West.

But in the East, they were excelled by the Baidyas/Kayasths, and in the North by Merchant castes (Khatris / Baniyas) and Kayasths

So far we have focused on general literacy. Now let's turn to English language literacy. Here again, the brahmin dominance is limited to the South and lesser extent the East. Not the other geographies.

We discuss English literacy next -

In the English language, the all India literacy rate was around 1.1% back in 1931 (as compared to the overall literacy rate of 9.6%.

But this varied a lot by geography and caste.

Here are the English literacy rates for a few upper castes in '31 -

Bengali Baidya: 33.5%!!
Mysore Brahmin: 18.7%
Madras Brahmin: 17.3%
Bengali Brahmin: 16.1%
Bengali Kayastha: 15.5%
Punjab Khatri: 7%
Punjab Baniya: 2.5%
BIhar Brahmin: 1.9%
UP Brahmin: 1.3%!!!!!

So it is clear that when it came to English literacy, the dominant groups were very small and in certain geographies - mainly Bengali Baidyas/brahmins/kayasths and Southern brahmins.

The UP brahmin English literacy rate was only marginally higher than the all India rate of 1.1%!

Clearly the traditional varna system was a poor guide to figure who was educated and who was not, back in 1931.

Maybe varna was a strong indicator 2000 years ago. But sorry, it failed as a strong predictor variable in 1931.

At the same time, it would be dishonest to claim that caste was a non-factor. Clearly it was a factor. But it has to be studied in relation to geography. Not in isolation.

A stunning fact: Lingayat literacy in Mysore in 31 was about the same as UP Brahmin literacy the same year!

Now many may be wondering - why dig up these 80 year old numbers in this day and age. Why?

The reason is that caste remains relevant in Indian public discourse and caste based angsts and perceived sense of injustices drive our politics. (Contd..)

So it becomes important to understand if these so-called injustices or "enormous advantages" were real?

Modern censuses don't help us answer this qn as they ignore caste altogether. Hence the need to look into the '31 census - we don't have a choice but to go back by 8 decades.

The story that emerged from this study is nuanced.

Sure. Certain groups did do pretty well in the British Raj relative to others. But they don't neatly correspond to our understanding of "hierarchies". Also even the communities that excelled were faring poorly in absolute terms

The 1931 census remains fascinating, and requires a larger audience.
It can help calm down frayed nerves in Indian society, as it reminds us that both the senses of superiority and inferiority nursed by groups are largely misplaced and not supported by history as recent as 1931!

The numbers for this thread are based on the actual 1931 Indian census report available in public domain. Also a big thanks to @entropied for many discussions on the same, that got me interested in this topic!