A recent article in the New York Times reminded me of recent work about why violence against civilians is so ineffective.
Referring to the shelling of several Lebanese towns near the Syrian border, Anne Barnard & colleagues document several responses by the Lebanese population:
“Residents said they believed that they were being targeted because Hezbollah, the pro-Syrian Lebanese Shiite militant group, is the political power in the village and bases some operations nearby. But Saad Hamedeh, the son of Hermel’s tribal sheik, said there were no military targets in the village. ‘They are trying to kill civilians,’ he said.”
This last line reminded me of Max Abrahms’ recent work, which focuses on how cognitive biases affect perceptions of violence against civilians. Armed actors generally target civilians for four main reasons: (1) to punish civilians for supporting their opponents (2) to deter them from doing so in the future; (3) to send a costly signal to a broader audience that the group is resolved and capable of inflicting maximum political damage; or (4) by accident or incompetence.
Regardless of whether the civilian targets are the victims of “costly signals,” punishment, or accident, Abrahms finds that the targets themselves often make their own, independent inferences about the armed groups’ motives. Importantly, when armed groups target civilians, the civilian population often understands these motives as simply murderous and senseless. Instead of understanding the violence to be politically instrumental, so to speak, they interpret the violence as “They want to kill us.” They see the violence as an end in itself.
To explain this phenomenon, Abrahms references a cognitive process called attribution error, where the object infers the subject’s motives based on the subject’s observed behavior rather than her stated and/or true intentions.
This is why violence against civilians, as opposed to violence against other armed combatants, has such a low success rate. It’s one thing to submit to rebels. It’s quite another thing to submit to people you see as “murderers.”
Interesting discussion. It seems like it would help to understand if the observing civilians (not directly targeted) are making similar (mistaken?) attributions. For example, if a civilian observes other community members being targeted, do they actually draw the inference that was intended by the armed actor that, for example, they should not give support to enemy combatants? Do they not simply throw-up their hands and draw the simple interpretation that “they want to kill us”? In other words, does a deterrence attribution exist among the people who are the intended targets of the deterrence?
This relates to Robert Jervis’s more general discussion (Monroe, ed., Political Psychology, 2002) of the difficulty of sending signals and the inability to control how the other side interprets those signals. Many years ago Frederick Sallager of RAND did a study of the British bombing campaign in the early years of World War II, which showed not only signals that were misinterpreted but also intended signals that were completely overlooked by the other side and events that were interpreted as signals even though they were never intended that way.