Thread by @shrikanth_krish

The Kashmir issue over the past days has made me ask a few questions

What causes a province in the very long run to be a part of the larger nation state?

What prompts it to break off?

Can this be theorized?

An attempt can be made using the statistical framework of ANOVA

ANOVA as many may be aware stands for Analysis of Variance - a way of analyzing variation in a given population

While I have only a passing acquaintance with it, what struck me is the idea of distinguishing b/w 2 types of variance

Within Group variance
Between-Group variance

In any large population, we have groups

Each group can be viewed as a cluster with some distinguishing characteristics that separates it from other groups

Yet within a group not everyone is alike
There are differences within a group. Just as there are differences across groups

So if you think of India as one LARGE population, it comprises of many groups of human populations

Karnataka
Andhra
Maharashtra
Bihar
TN
etc

Why are these groups different? Because we believe each to constitute a cluster and the mean of each cluster is very significantly different from the mean of another cluster

E.g. People in TN speak Tamil. Those in Karnataka speak Kannada. Those in Orissa speak Odia etc

But then the differences between groups is not HUGE. There are similarities too

Orissa is not as different from Karnataka as Orissa is from Ohio!

E.g. Orissa and Karnataka share a religion - Hinduism
They share the influence of Sanskrit
They share the tropical climate

But nevertheless the differences (especially lingual) are large enough for Orissa and Karnataka to constitute two separate clusters

But what makes Karnataka not break away from India and become a separate nation

Because we have "variance" within Karnataka. Not everyone in Karnataka = Mean value of Karnataka

Which is why though Karnataka is a separate cluster, the variation within it is critical

E.g. Karnataka has Brahmins. Gowdas. Lingayats, Shaivas, Vaishnavas, Smarthas, Muslims etc

And Orissa too has its Brahmins. Kayasthas, Kshatriyas, Shaivas, Vaishnavas, Christians among others

So the "within group" variance of Karnataka and Orissa somehow brings Karnataka and Orissa together

Though language divides Karnataka and Orissa

What happens if the "within group" variance reduces. While the "between" group variance increases?

The state develops secessionist tendencies

And it will eventually break off

So let's take J&K as an example -

It is clearly a distinct cluster. With a distinct geography. Deserves to be a distinct state. And not merely say an extension of Punjab or Himachal. No

But how does J&K fare in terms of "within group" variance?

It has considerable variance -

Muslims in the Valley
Buddhists in Ladakh
Pandits in the Valley (earlier, not now)
Dogras in Jammu
Punjabis in Jammu

But here's the thing - this variance is not uniform across the state. It is variance across different regions of the cluster

So in Bangalore, in a single street of Chamrajpet - you may meet 5 brahmins, 4 Gowdas, 2 Lingayats, 3 Muslims, and 2 Christians

While in Srinagar, you are likely to meet only Muslims.

Similarly you are likely to meet Buddhists only in Ladakh and seldom in Jammu or the Valley

So this creates sub-clusters within the cluster

And you get 3 sub clusters - Jammu, Ladakh, and the Valley

And while Jammu and Ladakh still have commonalities with other Indian clusters, the valley with its homogeneity and low within-group variance shows signs of breaking off

So what needs to happen for Kashmir valley to be a stable province and not break away?

You need to introduce "within group" variance

So that Srinagar, though Kashmiri, exhibits the kind of social heterogeneity that characterizes Chamrajpet in Bangalore or or RK Puram in Delhi

To my mind this is an insight that extends beyond provinces into larger debates of nation state formation

E.g. What makes Pakistan and India two different countries and not one?

Because circa 1947, the "within group variance" of Pakistan dropped because of religious nationalism, causing its differences w.r.t. rest of India to get accentuated, and its internal variations underplayed

25 years later, the "within group" variance surfaced in Pakistan

Causing East Pakistan to break away and form Bangladesh

West Pakistan remains an entity on account of religious nationalism

But religion is keeping the "within group" variance at bay uneasily. As religion weakens you will get the conflicts between Pashtuns, Sindhis, Punjabis, Balochis

Am not being cocky here by remarking on Pakistan

Because the same thing can be said about India too

Here Hinduism to some extent is helping unite the cluster-means in a single nation state

But there's another thing that works well for India

Our within-group variance does not cause the clusters to break away. As we have variance in each street in the form of caste and sect, which does not lend itself to sub-cluster formation

E.g. brahmins and Lingayats of Karnataka HAVE to live together. They can't form separate states unlike say Pathans of NWFP vs Punjabis of Multan/Lahore

So besides Hinduism, what works in India's favor is that while our within-group variance has the beneficial effect of bringing clusters together, it does not have the malevolent effect of engendering sub-clusters

As the within-group variance is vertical in nature
Not horizontal

So what should we do to make J&K work?

Increase its within-group variance, so that the "mean" difference between say J&K and any other Indian state (e.g. Punjab) becomes less critical!

That would mean - introduce heterogeneity in the valley

It can be religious. It can also be lingual. And ethnic

The greater within-group variance will counteract the now-somewhat remote cluster mean of J&K relative to other clusters