Letter to a colleague: on LTTE

Back from the long siesta... Oh well, what goes up has to come down. Anyway. Here is a letter I wrote to a colleague on LTTE and terrorism.
-------

Here is a piece by Aatish Taseer in New York Times on the LTTE saga: (Linky). I thought I would comment on the article.

I am not sure if you know about Aatish. He is the son of an Indian journalist and commentator, Tavleen Singh, and a (now deceased) Pakistani politician Salman Taseer. Salman Taseer, as long as he lived, did not openly acknowledge Aatish as his son and Aatish primarily grew up in India with his mother. Towards the end, the father and son reconciled. But then, as fate would have it, Salman Taseer defended a Christian woman who was also accused of blasphemy and insulting Islam. So Salman got killed by a jehadi named Mumtaz al Qadri, who is probably in jail now but with a lot of sympathy from the general population of Pakistan which still sees this act as wajib-ul-qatl (crudely translated as due punishment for apostates, blasphemers and their supporters). Salman also had a visceral hatred for many things India, and it was an irony that just like most things in Pakistan, the primary thing that defined Pakistan (not being India with Islam being in danger in India and safer in Pakistan) goaded someone to kill another defender of the same idea. 

That partly explains Aatish's empathy for the Sri Lankan Tamils, at least in this article. I think the most common theme missed by many commentators on Sri Lankan matters is that things flow back and forth over time. The momentary hopelessness of the Tamil polity in forging their destinies is somehow seen as a bad thing or a good thing, if one gets emotionally invested in the situation. But if one discards that approach and sees things from an observationalist viewpoint, the triumphalism that is very common in the Singhalese population is pretty much a self-goal. Further, the triumphalism is displayed not only against the Tamils, but also the Muslims (who are also mostly Tamils, but never accorded that respect by either the Tamils or the Singhalese for their own complicated reasons). For their part, the Muslims have remained divorced from the violence of the different Tamil outfits, but there is often a last straw that breaks a camel's back. 

I will probably give the Tamils 10 to 15 years of cooling-off time before they start getting violent again. The one thing that is keeping the powder dry is the skewed demographics as of now, due to the emigration of Tamils to Europe, Canada and Australia and also the losses in the various wars. The Singhalese believe that posting ex-Army people in Tamil territories in the North and the East amidst the Tamil population would give them early warning signals of trouble that could allow them to ship soldiers to flashpoints, which is why they are speeding up infrastructure projects in collaboration with China. There are often limits to every such contingency measure. When things go unpredictable and belly-up, it is often a new issue that noone had imagined would happen which could then be post-facto rationalized -- the classical black swan argument or the argument on how complex systems fail. It is often not a single reason that causes things to go belly-up, its often a collection of small things just like in the Titanic or Pearl Harbor. 

I am not sure if there is a theory of self contradictory outcomes somewhere in the sociology literature. But one such candidate theory would be: any action that is well thought out to make a certain outcome less probable would often produce a certain other outcome which in turn could make the original outcome more probable. I am pretty sure that a variant of Murphy's laws can be twisted to this form. In any case, in the signal processing literature, we have a variant of this tradeoff called the bias variance tradeoff (Linky). To put it in simplistic terms, if the goal of an algorithm is to game/predict a certain unknown outcome based on indirect observations that are random, one could hope that the policies prescribed on average produce the intended outcome (no bias). But then with a certain realization of the observations that is actually seen, the variation from the intended outcome may be too huge so as to defeat and render useless the average property of this class. 

I think the Singhalese are trying too hard to game this scenario into the far future in case things go wrong. But there are simpler means to achieve that than by planting ex-Army men and land grabbing in what are seen as Tamil territories. I think reconciliation and trust is a better approach with some carrots and sticks. But then, we would nt have had the problem of the LTTE in the first place if reconciliation and trust was there to start with. The LTTE in itself came up only after the meltdown of the Bandaranaike-Chelvanayakam pact (the Banda-Chelva pact as it is called), the loss of trust by the Tamil people on a mostly-democratic outfit TULF, leading to violent outfits from which emerged the LTTE as the victor. 

I believe some form of reconciliation has been achieved in how the Government of India (GoI) deals with some terrorist outfits. Of course, some of it is purely accidental and somewhat deep-down philosophical in terms of approach and life, and not because of great planning. For example, when the Mizo National Front (MNF) abandoned violence in ~1986 in one of the northeastern states (with a mix of tribal groups of which the Mizos were the dominant demographically), the Chief Minister of the state resigned to pave the way for a MNF government followed by a general elections in which the MNF won the popular vote. This regime only lasted for a short while as the leader of MNF died of natural causes, but Mizoram has more or less remained peaceful relative to the otherwise violent Northeast of India. But to coax the MNF to the table, the Indian Air Force was used to shell civilian locations, one of probably very few (if not perhaps the only one) uses of aerial bombardment against civilians in India. 

The current model of dealing with what the GoI deems as the pro-talks faction of one of the major outfits in the Northeast, the ULFA, is also precisely this. The pro-talks leaders were all safely put up in Bangladesh, hosted by the various Islamist regimes under Zia-ur-Rehman, Hossein Mohd. Ershad and then Begum Khaleda Zia from the late 70s and early 80s till now. When a change of regime happened in 2009, GoI coaxed the Bangladeshis to hand over these people in return for a land swap agreement with net loss of territory for India (the final fruition of the Indira-Mujib accord of 1974) and water sharing on the Teesta river which is absolutely critical for Bangladesh. Neither of these has fruitioned now for Bangladesh, but that is a different story. The GoI has treated these ULFA leaders with a good amount of respect, even though electorally if they contest today, it will be a walkover for the current regime in Assam. I think the GoI will be happy to see them form a political party and burn their credentials over a period of time, as every terrorist outfit always does if it takes the democratic approach.

Historically too, that approach has been followed by the GoI in different forms. The Communist party formed the first democratically elected government anywhere in the world in the southern state of Kerala in 1956. Nehru, despite being a Fabian socialist and a good friend of Stalin, Brezhnev and Khruschev, was deeply suspicious of dictatorship of any kind, even the Proletariat one. So it did take a fair amount of moving away from his position via rhetorical gymnastics to allow the CPI to take over the Kerala government. The CPI takeover of educational institutions from the Catholic Church and the consequent violence led to the Constitutional use of an approach (now sparingly used, thankfully) to disband the state government. But I think the CPI has been coaxed into democratic politics with all the barrel of a gun speeches primarily restricted to rhetorical fanfare. This is a remarkable transformation which the Maoists accuse the Communists of. In fact, the Maoists would treat the Communists as their biggest enemies before attacking any other party in India today.

The same can be said about many other political parties in India in different states: DMK in Tamil Nadu, Shiv Sena in Maharashtra, Akali Dal in Punjab (which then splintered into many outfits with different consequences), Asom Gana Parishad in Assam, and so on. Even the definition of a successful terrorist outfit is not well defined. I believe the most successful ones are some of these political parties, DMK especially. Despite all these attempts over the years, India still has approx. 50-60 terrorist outfits, a good ~30 of them would be based out of the Northeast, ~10 Islamist outfits, mostly from J&K, ~10 Leftist outfits, and a few Hindu and Sikh ones. 

Its amazing that these attempts at neutralizing outfits, Indian style, has not been studied that much. There are a large number of lessons to be learned, good, bad and ugly, and one could try to discern conscious policies from historical accidents. The limits of predictability in policy making and the consequent agnosticism that brings into the picture have not been well understood, either in the Indian context or elsewhere, which is why we see folks keep getting too protective of their own ideas and turfs. Again, signal processing literature rescues us by allowing an inevitable Cramer-Rao lower bound to any estimator/predictor.

Further still, the metrics used for defining success for the outfit as well as the government is also needful of reassessment. A terrorist outfit would do well to transform to a political party that can pilfer the state coffers at will. The government would want this eventuality to happen as it would enhance the State's claim to suzerainty in a theoretical sense and would also be de facto assured in a practical sense with certain caveats. Yet, despite a commonality of eventual goals, we do not see all terrorist outfits abandoning themselves overnight. I think this is because terrorism is a cat and mouse game where from a game-theoretic viewpoint, the Nash equilibrium is clear. But both parties are greedy to seek a solution in the Pareto boundary that favors only itself. 

To connect it back, this is precisely what Prabhakaran was. Even amongst the greedy terrorist outfit chieftains, he was greedy beyond explanation. Which is one reason why the people who were coaxed into condoning him and his actions, either of their own free will or unwillingly, will have to bear the cross for his sins. But then this is just another short-term momentariness in a long-term back and forth, which could be easily predicted to a certain degree under the caveats mentioned above.