Christopher Sullivan had an excellent guest blog post over at Political Violence @ a Glance yesterday. It’s a fascinating, rigorously academic look at an important question: does the use of torture actually help a state accomplish its security goals (i.e. reducing violence by terrorists or insurgent movements)? As a side note, it’s interesting that debates on torture inevitably fall back onto the empirical “does it work” side, which says something about the persuasive condition of moral discourse in our society – but that’s a point for another day.
Sullivan tackles the “does torture work” question by examining the conflict in Guatemala in close analytic detail, looking for micro-level correlations between the use of torture and various conflict outcomes (increased or decreased violence). Unsurprisingly (even to him), he found that the use of torture by the government did not have any effect on the level of violence by the insurgents, whom the torture was presumably supposed to help stop. This is an unremarkable finding, although it is good to have it verified in a careful and objective fashion.
What Sullivan did find surprising was that the use of torture was “robustly associated” with an increase in killings by the counter-insurgent forces. That is, the same people (or, at least, people on the same side) who were committing the torture also became more likely to engage in killing in other conflict contexts. He offers no explanation for this finding, at least in his blog piece, but does seem somewhat surprised by it.