My thesis topic, nonviolent responses to mass atrocities and genocide, is not the most straight forward. Few scholars have written about it in depth, and, if I do say so myself, it’s very conceptually difficult. The idea does pop up frequently in related literature, but it’s almost always dismissed within a paragraph. Luckily for me, these claims don’t hold much water.
When authors do address nonviolence in response to genocide (as opposed to mass atrocities or civil war violence, which gets more nuanced attention), the standard line is that nonviolence is powerless against an enemy committed to killing a certain group. In the face of this type of single-minded hate, violence is the only defense. At first glance, these conclusions make sense, but authors often contradict themselves later in the works where they make these claims. Two examples here are political scientist Oliver Kaplan’s dissertation on civilian autonomy in Colombia and Chirot and McCauley’s Why Not Kill Them All? I pick on these two works not because they’re problematic; both works, and Kaplan’s in particular, are great works of scholarship, but they both make the mistake of dismissing nonviolence as a response to genocide.
Kaplan cites Valentino on why civilian resistance to genocide is futile. He contradicts this, however, by proposing that creating community processes for conflict resolution reduces the chances residents will use armed actors to settle local grudges. This happened, for example, during the Armenian genocide, when Kurdish tribes allied themselves with the Ottoman Empire to eliminate Armenians. The same thing happened with Banyamulenge, the Rwandan army, and Congolese Hutu. Community conflict resolutions processes aren’t a silver bullet for stopping genocide, but they could potentially contribute to a decrease in violence by eliminating community divisions that can be exploited by armed groups with genocidal intentions. Perhaps an even better example is in Why Not Kill Them All?, where Chirot and McCauley make a similar argument as Valentino. They examine how “contact programs” and a strong civil society can provide a bulwark against the rise of genocidal ideologies and a fear of the “other” that lead to mass killing. Both works outline nonviolent strategies that can prevent or mitigate genocide despite their claims to the contrary.
There are two central causes for this inconsistency in works on genocide. The first is an overly simplistic conception of what stopping genocide entails. Doing that is a long process that doesn’t commence in the middle of atrocities. Basically, the authors have forgotten about genocide prevention, most of which is done nonviolently. Secondly, these dismissals are based on a mistaken interpretation of nonviolence. Nonviolence in response to genocide is so much more than unarmed civilians physically confronting their would be murderers, because we all agree that wouldn’t be very effective. Conflict resolution programs, anti-hate education, finding employment for young men, the dissemination of truthful news, and humanizing portrayals of a potential victim group can all be used to prevent genocide (and it’s important to remember that all these strategies can be used after violence has started, because genocide develops gradually, meaning there is no point before which it’s “prevention” and after which it’s “response”).
So yes, it’s very much possible to prevent and respond nonviolently to civil war violence, mass atrocities, and even genocide. Genocide is not a unique phenomenon, as compared to other types of violence, that only responds to force and not to “reason”. It’s time modern scholarship accepted that.
The Widening Lens 
True, but you’re overlooking some more recent developments in conflict processes scholarship. What about recent quantitative work that shows that naming & shaming by NGOs, IOs, etc. reduce the risk of atrocities occurring (Jackie DeMeritt’s recent work), and reduce the severity of them once they begin (my recent work, as well as Jackie DeMeritt’s)?
No, you’re right on that. My categorization of the relevant scholarship was probably too generalized. I’ve seen the claim a lot recently which influenced this post.