The project “˜How Genocides End’ included the 2004 “˜Back from the Brink’ seminar at the Holocaust Memorial Museum, the SSRC Webforum, and two seminars at Harvard, co-convened by Jens Meierhenrich, during 2008. What are the preliminary conclusions? This posting is a personal reflection on the outcomes thus far.

Our first finding–emphasized by Dirk Moses–is that there is notably little research into the cluster of issues about how genocides and episodes of comparable mass violence end, and how near-genocides are averted. The genocide literature focuses overwhelmingly on a few cases (Armenia, Nazi Germany, Cambodia, Former Yugoslavia and Rwanda), excluding dozens of other candidates, ranging from Bangladesh, Burundi and Biafra through Guatemala, Indonesia and Iraq, to the USSR. The literature is rich on the background, origins and nature of genocidal violence, but when it comes to the ending of that violence, it becomes normative and prescriptive, concerned with how genocides should end””by military intervention and prosecution of the perpetrators.

Moreover, the normative ending of genocide reduces the options to just two””an intervention that terminates the episode, or in the absence of such an intervention, completed extermination of the target group by the perpetrators. This is a much simplified narrative that fails to capture the diversity of historical, empirical genocidal endings.

Analyzing the comparative historical record, we see a number of actual endings.

Invasion has ended some genocidal or near-genocidal regimes (Nazi Germany, Bangladesh in 1971, Cambodia’s Khmer Rouge, Idi Amin’s Uganda, Saddam Hussein’s Iraq). Note that in none of these cases was ending genocide either the main motivation or the immediate objective of the military action, which was undertaken when the regimes in question attacked or destabilized their neighbors or threatened the interests of a great power.

Armed resistance by groups identified with the target group has ended genocide in some cases, including the Yoweri Museveni’s National Resistance Army in Uganda and the RPF in Rwanda. Ben Kiernan has also pointed to significant resistance in Cambodia.

More common, it appears, is that a genocidal campaign ends when the perpetrators consider that they have done enough to achieve their political or military goals. This was the case for episodes of settler-colonial genocide in north America, Australia and Namibia. It was also the case in Stalin’s Soviet Union, Suharto’s Indonesia, the Burundi genocide of 1972, Mengistu Haile Mariam’s “˜Red Terror’ in Ethiopia in 1977-78, the Guatemalan counter-insurgency of the 1980s, and Saddam Hussein’s campaigns against the Kurds and the Marsh Arabs. The USSR is perhaps the definitive example: Stalin was able to start and stop his terror campaigns and his deportations overnight. In the case of Biafra, the genocide threatened by the advancing Federal army did not happen when that army won the war””as soon as it achieved victory, the killing stopped.

There are other cases of genocidal campaigns being halted by elite dissension or exhaustion. The Nuba Mountains example is one such. Siad Barre’s campaign in north-west Somalia beginning in 1988 is another case.

An important set of endings occurs through flight and asylum””when the target population gets itself out of harm’s way, or is evacuated. The case of the Somali Bantus, detailed by Catherine Besteman, is a prime example.

We are also compelled to ask, what constitutes an “˜ending’? In cases of mere respite, such as the settler-colonial genocides, Stalin’s Soviet Union, and Saddam’s Iraq, the apparatus and intent remain intact, and a genocidal campaign can be re-started at any time, against the same or different target groups. In other cases, the ending of one genocidal campaign prefigures another configuration of mass violence, perhaps by members of the former target group against their erstwhile oppressors. This was the case in Uganda and has continued to be a pattern throughout the Great Lakes.

Cases of invasion or humanitarian intervention have not stopped violence either. Kosovo is a problematic example of intervention, both because of the suspicions that the KLA incited the Serbs with the aim of spurring an intervention, and because of the ethnic cleansing of the Serb minority that took place under NATO occupation. Independent Bangladesh was unstable and violent, though far from genocidal, after 1971. War continued in Cambodia after the Vietnamese takeover and worsened in Uganda after 1979. And the invasion of Iraq, while it stopped Saddam’s genocides, did not lead to peace.

These observations also oblige us to disaggregate “˜genocide’, an exercise led by Jens Meierhenrich. His starting point is that the term “˜genocide’ began life to designate a crime, for which an individual could be prosecuted, and morphed into a sociological term in a rather unsatisfactory way. Meierhenrich proposes distinguishing between genocidal acts, genocidal campaigns, and genocidal regimes””suggesting that all three are needed if we are to determine “˜genocide’ with confidence. This is a useful distinction insofar as it allows us to identify whether an individual genocidal act (such as an ethnic massacre) is an isolated incident or part of a cluster of such incidents which are form a campaign. In turn, we can also determine whether a campaign of genocidal character is conducted in a limited fashion (perhaps in pursuit of a political or military objective) or is an intrinsic part of the political project of a genocidal regime. Our focus in the project is on genocidal campaigns and genocidal regimes””dealing with genocidal acts would bring us to an unmanageable agenda.

Another important insight is the distinction between essentialist violence, and instrumental violence. The most apt comparison is between Hitler’s Germany and Stalin’s Soviet Union. For Hitler, violence was an essential part of his project and there was no higher purpose than genocide. Something similar is evident in the ideology of Hutu Power in Rwanda in 1994. For Stalin, violence and his campaigns of terror, deportation and famine were all instrumental in pursuit of political ends. This is one reason why he could stop them so abruptly. Instrumental violence is far more common than essentialist violence.

These observations point us to an enduring problem with the criminology of genocide, namely what it means to determine the mens rea (state of mind) of the perpetrator. Unless intent to commit genocide is proven, a conviction for genocide cannot be obtained. Popular understanding and legal definition diverge.

In popular discourse, genocide is usually seen as the apex of crime, implying that genocidal intent trumps all other intents””it is not only consciously formulated but overriding and persistent. Successful prosecutions for genocide in Former Yugoslavia and Rwanda have lowered the bar, and inferred genocidal intent from circumstance. They have distinguished intent from motive, departing from the lay interpretation of genocide. There is no room in this emergent legal custom for the distinctions outlined in the previous paragraphs, namely whether the intent is persistent or momentary, overriding or secondary to another intent, essential or instrumental/circumstantial.

For the survivors of genocide, the impact of the crime on their lives will never go away, though a measure of personal coming-to-terms can be found through measures such as identifying the remains of their family and friends and giving them a dignified burial and memorialization, truth-telling, restitution of alienated land and property, and prosecution of perpetrators. These measures are hard, perhaps impossible, to undertake while the perpetrators remain in power. How a genocide ends has profound implications for what happens afterwards.

Most importantly from the policymaker’s perspective, this analysis puts the distinction between the ideal ending and the actual ending in our face””a point emphasized by Bridget Conley-Zilkic. The ideal ending (intervention) is an ending by us””the “˜international community,’ according to well-defined rules, which brings about peace, justice and democracy. Actual endings typically involve only the ending of genocidal campaigns, leaving ongoing armed conflict, displacement and exile, authoritarianism, impunity and other human rights abuses. These are all bad things that demand remedies””but they are not, we submit, genocide. The legacy and repercussions of a crime are different from the crime itself.

This analysis poses major moral and policy dilemmas. How are we to determine whether and when a “˜genocide’ has ended? Is it sufficient for killing and other acts that constitute genocide to stop, or be reduced, or does intent (however defined) have to disappear as well? Is it sufficient for a genocidal campaign to end, even if a genocidal regime remains in place with the potential to launch another genocidal campaign? Does genocide only “˜really’ end when justice has been done and restitution is complete? There are no obvious answers to these questions””but at least we have made a beginning by asking them.