Violence data and analytical aggregation in S. Asia

These passages from (1) Jagath Senaratne’s book on political violence in 1977-1990 Sri Lanka (p. 146) and (2) Nandini Sundar’s new book on anti-Naxalite counterinsurgency in Bastar (p. 209) provide a lot of food for thought:

1) “evaluating conditions in the southern areas of the country in late 1989, Amnesty International reported that:

‘Violence is now so widespread that it is often difficult to establish with certainty who the agents of specific killings were – or even to identify the victims whose bodies are sometimes grossly mutilated, burned to ashes or transported long distances form the scene of arrests or abduction before being dumped” (Amnesty International (ASA 37/21/89) 1989:5).’

The confused, unstable, and dangerous situation led many to believe that the violence was random and meaningless. The imputations of randomness by some observers (mainly journalists) was a result of an inability to see the many different strands of violence: ‘The Violence’ was, in fact, a bundle composed of many separate strands of violence.”

2) this is Sundar interviewing a former policeman in Chhattisgarh:

“Me: Were all these Naxalites [people shot in Bijapur]?

Ex-policeman: Of course not. None of them were Naxalites. Sometimes an SPO would point out someone and tell us to shoot, sometimes we shot simply because the villager was running away and refused to stop when we called out. We call out in whichever language we knew – Telugu, Hindi, but the villagers didn’t understand.

Me: Did you record those deaths somewhere?

Ex-policeman: [Sounding shocked]. Our jobs would be in trouble if we did. We left the bodies in the jungles. We recorded it as an encounter only if someone was actually wearing a uniform or carrying a weapon. I personally never killed anyone, but if by chance my bullet hit anyone in an encounter, I hate to think of it.”

Two things are going on here. The first is the extraordinary difficulty of gathering clear, accurate data on patterns of violence.  I have found this in great detail on my own Sri Lanka research – even getting the Army and, now, Police fatalities data only captures a tiny amount of the overall violence, much of which didn’t involve security forces. In India, the only data I even vaguely trust is the self-reporting of their own deaths by the security forces; Drew Stommes and I are doing a lot on MHA and state police records.

Yet, as Anit Mukherjee has shown in the case of the Indian Army, even those are open to question. Press accounts are deeply problematic, since the press has limited or no reach into numerous physical and social spaces in conflict zones, self-censors or is intimidated, and/or is otherwise politicized. Talking to journalists in Kashmir and Nagaland has been revelatory on these questions.

So while there are hugely impressive micro-level events data projects ongoing that should be supported and encouraged, they are likely to be limited to a set of contexts that are well-studied and well-covered in the contemporary period or have unusually good primary documentation. That, I suspect, captures only a small, non-representative, subset of the universe of cases.

The second is Senaratne’s analytical claim that, despite these challenges, there are distinguishable logics of violence within this opaque political environment. The open question this then begs is how we can pursue these logics in the face of the first problem. I don’t have a great answer. At a minimum, I would suggest that we make sure to not equate the study of micro-level events data with the study of political violence. Aggregating up to meso- or macro-level politics may be the only tractable strategy in the cases where this is an insurmountable challenge.

Moreover, this move can offer distinctive theoretical leverage on the broader political incentives and contexts facing individuals and communities: the elite politics of national leaders, the bureaucratic politics of security institutions, the electoral politics of local political competition, etc. These tend to be taken for granted as natural, fixed features of politics in much recent literature, but they themselves are hugely variable. A turn to the micro has been enormously valuable, but I suspect we’re at a point in which disaggregation is hitting the flat of the curve in some domains. There is an opportunity to scale back up, and to bring bigger picture politics back in while retaining a sensitivity to the complexities and importance of micro-dynamics.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s