- playing down the number of tutsi victims
- accusing the RPF of shooting down the plane
- playing down the crimes of the Hutu militia
As we make no claims about the shooting down of the plane, we will ignore this comment. The other two, however, merit a response as it is addressed by our work – at least, partially in the film. Before beginning, however, it is useful to note that the letter of complaint identifies us as working for the Defense at the ICTR. Actually, the exact wording in their letter is: “two US academics who worked for a team of lawyers defending the génocidaires at the ICTR.” Several things merit attention about this comment.
First, we were already working on doing a detailed study of the violence undertaken in Rwanda during 1994 paid for by the National Science Foundation. This involved conducting interviews with genocidaires as well as survivors, conducting focus groups among those who lived in Rwanda during 1994 (Tutsi as well as Hutu), conducting a survey in Butare of experiences during 1994 as well as compiling diverse databases from different Rwandan ministries and human rights organization (both inside and outside Rwanda) as well as compiling information from all media coverage of Rwanda between 1985 and 1999. The ICTR interaction was a small part of the larger research project.
Second, the letter of complaint incorrectly represents the nature of our relationship with the ICTR. For example, our services were requested by the Prosecution of the ICTR and this was our first contact with the organization. Specifically, they wanted:
- Complete coding of witness statements which they had in their possession but they did not have the resources to enter and code (a “master chronology” from the ICTR Prosecution);
- Conduct a comparative analysis of the chronology generated using their witness statements vs. other sources in our possession, noting zones of correspondence and disagreement;
- Create an animation of killings on a day-by-day basis for country (lowest level will be determined by census data they can provide);
- Create detailed maps of killings through out country, as well as maps of weapons types used and changing nature of violence over time;
- Provide a GIS regression analysis to establish if killing patterns were non-random or systematic.
We were told that they wanted this so that they could “contextualize” their cases and have a broader understanding of what had happened during 1994. To meet these needs we requested the following:
- Individual level records for 1991 census data (in order to calculate %Tutsi population) – source: Rwandan Ministry of Census;
- 2001 cell level census data – source: Rwandan Ministry of Local Affairs;
- Physical locations of party HQ or locations of individuals claimed to be associated with execution of genocide plan;
- 1:50,000 topographic maps of Rwanda from Canadian defense forces (we were willing to provide the funds for this ourselves; they simply needed to make request on our behalf);
- Complete set of interview records as well as existing database based on witness interviews so that we could code this information and conduct the comparisons noted above (the coding would be paid for by our NSF grant); and,
- Unrestricted access to prisoners already convicted by the ICTR to ask them questions about what they did, what they saw and who was where in the country and when.
We asked for no consulting fee or anything besides what was requested above as well as the right to publicly distribute all information after the project was completed. This was agreed to but we were told that everything would take a little time.
Before we began this work, however, we were told that the Prosecution was no longer interested in a broader understanding of what happened and they were going to concentrate on the specific cases that they had on their docket.
Somewhat after this point, we were contacted by the Defense and asked if we would be willing to provide exactly what the Prosecution has earlier requested. After we were informed that we would not be explicitly associated with the Defense (which was a concern for us) but rather that we would be providing information for the court (the Judges, Prosecution, Defense and official archive to be compiled as well as later made available to the public), we then repeated our list above about what was needed. After much negotiation, approximately 7000 readable witness statements were provided, a detailed topographic map of Rwanda on April 6th (this is before the revelations of available US satellite imagery that might be day to day) and access to some former members of the RPF and FAR to corroborate geographic locales.
The characterization of our relationship with the Defense and the nature of our research endeavor in the letter of complaint was thus incorrect. But this is just the beginning/background. Now, we move to the specific points themselves.
1) It is suggested that we minimize the crimes of the Hutu militia. This is simply false. If anyone had seen the documentary and/or seen/read our discussion of what we found, then they would clearly see that we attribute the majority of the killing in 1994 to territorial zones under the jurisdiction of the FAR (the Rwandan government at the time) and those associated with them (including the various organizations referred to uniformly as the “Hutu militia” but which represented a conglomeration of distinct institutions). We are not precise on the identification of the perpetrators of these crimes (e.g., presidential guard vs. different militias vs. gendarme vs ordinary citizens or combinations) because the evidence on this is unclear. In some locations, individuals have been able to identify some of those responsible for killing. In other locations, either no one was left alive to identify those responsible for killing or the survivors were not able to identify who was responsible – something that is quite understandable under the circumstances as well as something that is quite common in instances of mass killing. Lacking this type of precise information, which no one in the public domain has with any degree of accuracy, we opted for an evaluation of killing designated by territorial control. This is something that is more reasonable to estimate with available information. Now, the differences between what specific perpetrators were involved in specific activities at a particular time (hour/day) and place (e.g., village) do not matter, if one is simply interested in whether or not someone died. If one is trying to affix blame, hold individuals accountable or understand what motivated individuals to engage in the different killing, however, then such information is crucial.
Was some of the violence that took place within the area of control of the FAR what would be legally classified as genocide? Yes, without a doubt. We have said this repeatedly. Some of violence was clear as there were representatives of the state present, there were utterances that spoke to the issue of their intent in a specific time and place and there was violence. Unfortunately, however, we cannot be clear on how much of the violence (i.e., what proportion of all acts considered were legally classified as genocide) and that is because making such a judgment requires extremely detailed information about the victims, the perpetrators and the latter’s motives – the last characteristic being the hardest to understand as they reside within an individuals mind. Our approach has been to adopt an approach based on detailed evaluations of political violence and assessing responsibility through a consideration of territorial control throughout Rwanda during 1994. Where possible we corroborate such information with ICTR eyewitness testimony and comments made in books written about Rwanda. Only through the comparison of the different sources can we begin to approach a reasonable answer. Again, all the information used for these calculations are provided on the GenoDynamics webpage so that individuals can make their own investigations/evaluations.
Part of what is at issue is a belief that large numbers of casualties equals genocide, but this is not the case as it is a very specific legal phrase. Large numbers of casualties are horrific, crimes against humanity and human rights violations, which have different criteria for identification, but they are not by definition genocide. The ICTR here is instructive:
ICTR Statute 2010
Article 2: Genocide
1. The International Tribunal for Rwanda shall have the power to prosecute persons committing genocide as defined in paragraph 2 of this Article or of committing any of the other acts enumerated in paragraph 3 of this Article.
2. Genocide means any of the following acts committed with intent to destroy, in whole or in part, a national, ethnical, racial or religious group, as such:
- Killing members of the group;
- Causing serious bodily or mental harm to members of the group;
- Deliberately inflicting on the group conditions of life calculated to bring about its physical destruction in whole or in part;
- Imposing measures intended to prevent births within the group;
- Forcibly transferring children of the group to another group.
3. The following acts shall be punishable:
- Conspiracy to commit genocide;
- Direct and public incitement to commit genocide;
- Attempt to commit genocide;
- Complicity in genocide.
B. Crimes against Humanity
Article 3: Crimes against Humanity
The International Tribunal for Rwanda shall have the power to prosecute persons responsible for the following crimes when committed as part of a widespread or systematic attack against any civilian population on national, political, ethnic, racial or religious grounds:
- Persecutions on political, racial and religious grounds;
- Other inhumane acts.
C. Violations of Article 3 Common to the Geneva Conventions and of Additional Protocol II
Article 4: Violations of Article 3 Common to the Geneva Conventions and of Additional Protocol II
The International Tribunal for Rwanda shall have the power to prosecute persons committing or ordering to be committed serious violations of Article 3 common to the Geneva Conventions of 12 August 1949 for the Protection of War Victims, and of Additional Protocol II thereto of 8 June 1977. These violations shall include, but shall not be limited to:
- Violence to life, health and physical or mental well-being of persons, in particular murder as well as cruel treatment such as torture, mutilation or any form of corporal punishment;
- Collective punishments;
- Taking of hostages;
- Acts of terrorism;
- Outrages upon personal dignity, in particular humiliating and degrading treatment, rape, enforced prostitution and any form of indecent assault;
- The passing of sentences and the carrying out of executions without previous judgment pronounced by a regularly constituted court, affording all the judicial guarantees which are recognized as indispensable by civilised peoples;
- Threats to commit any of the foregoing acts.
D. Individual Criminal Responsibility
Article 6: Individual Criminal Responsibility
- A person who planned, instigated, ordered, committed or otherwise aided and abetted in the planning, preparation or execution of a crime referred to in Articles 2 to 4 of the present Statute, shall be individually responsible for the crime.
- The official position of any accused person, whether as Head of state or government or as a responsible government official, shall not relieve such person of criminal responsibility nor mitigate punishment.
- The fact that any of the acts referred to in Articles 2 to 4 of the present Statute was committed by a subordinate does not relieve his or her superior of criminal responsibility if he or she knew or had reason to know that the subordinate was about to commit such acts or had done so and the superior failed to take the necessary and reasonable measures to prevent such acts or to punish the perpetrators thereof.
- The fact that an accused person acted pursuant to an order of a government or of a superior shall not relieve him or her of criminal responsibility, but may be considered in mitigation of punishment if the International Tribunal for Rwanda determines that justice so requires.
HRW ICTR Case Law Digest
“ICTR jurisprudence correctly recognizes the mental element of genocide in Article 2(2) above as its distinguishing feature, namely the requirement of a specific intent (dolus specialis) to destroy, in whole or in part, a national, ethnical, racial or religious group, as such. This mental element applies to all material acts of genocide enumerated under Article 2(a)-(e) of the Statute. Since the underlying acts - such as killing or causing serious bodily or mental harm - are not international crimes as such, '[i]t is this specific intent that distinguishes the crime of genocide from the ordinary crime of murder.'' Thus, in addition to defining genocide, the requisite mental element also delineates the normative sphere of international criminal law from that of domestic law” (Akhavam 2005: 992).
There is also a conflict/overlap between Art. 2(3)(e) “complicity in genocide” and Art. 6(1) “or otherwise aided and abetted planning, preparation or execution of a crime referred to in Articles 2 to 4”. There are, according to Akhavam, different interpretations of the ICTR regarding the dolus specialis/dolus generalis requirement for liability.
A side note on Data and our sources
Related to the criticism noted above, it is suggested in the letter that we provide "false figures". This is problematic because all of "our figures" (i.e., statistical estimates) come from the existing Rwandan government (i.e., the Ministry of Education; the Ministry of Youth, Culture and Sport and the Ministry of Local Affairs) as well as three reputable human rights organizations (e.g., Human Rights Watch, African Rights and The Tutsi Survivor organization Ibuka). All we have done is take the information from all of these different sources (in addition to ICTR eyewitness records, a survey in Butare, about a dozen focus groups and dozens of interviews), in order to generate estimates of the number killed as well as utilizing census records and demographic projections regarding the probable number of each ethnic group killed. The project has been extremely detailed about every step of this process. Indeed, unlike those that have criticized us, we not only make the original documents available for each source utilized (so that one can see what was done in each case) but we have provided our coding of that material (noting the precise page numbers from which it is taken), detailed descriptions of how we generated our estimates as well as the raw data that was generated. We note very clearly that our estimates are improved with the inclusion of additional data and this is why we have continued to collect as much information as possible (to this day). The imprecise nature of casualty estimates demands both the pursuit of additional sources and transparency. If our figures are believed to be false, therefore, then that is a claim delivered against the current Rwandan government, one of the most prominent human rights organizations in the world and an organization that has been identified as the leading institution representing the Tutsi survivors.
2) It is suggested that we play down/trivialize the number of Tutsi victims. To begin, there are no deaths that are trivial. All human suffering is unacceptable and we (as a community of human beings) should work toward no one suffering from political violence – from any actor. As for our estimation, again, if anyone had seen/read our discussion of what we found, then they would clearly see that we believe that there are a range of estimates on the number of total casualties as well as the number of casualties by the different ethnic groups involved. There is no single number. In fact, there is no way to discuss casualties figures without such ranges given the highly imprecise manner in which such figures are identified in conflict situations.
Some objections that could be raised concern our use of the census taken in 1991 by the then Rwandan government. For example, some wish to argue that the census of 1991 (a common starting point for an inquiry into how many Tutsi and Hutu died) is correct and that this is how the Rwandan government identified those that it wanted to kill. Some wish to argue that the census of 1991 was not correct, however, and that it underestimated the number of Tutsi on purpose given the situation prompted by the interstate war but that the government still knew the actual number and used this to enact it’s killing. With this one could subtract the number believed killed from this number to derive at some understanding of how many of each group died (i.e., how many tutsi and how many hutu had been killed). If one did not trust the 1991 census, however, then they could go back to earlier censi (1978, the one undertaken in the 1950s or colonial records) – as we and others have done – and project forward with a specific population/ethnic group growth rate.
The problem with the position that the census was off but the government still knew who was who is that no one has produced the document that was used by the government agents. There is no “kill list” of 500,000 or 1,000,000 individuals that were located at different locations throughout the country that has been found – at least nothing that has been publicly distributed. Even partial lists (e.g., the names that were read over the radio) were small in comparison to the number of individuals commonly believed to be killed. The population is largely illiterate and thus even if such a list existed, exactly who would use it would be limited.
Essentially, research on the topic of Rwandan casualties has relied upon eye-witness testimony via survey, census and interviews. While insightful in many ways, none have released the original/raw data collected during efforts for external validation accept for isolated cases and/or a relative small number of individual’s stories. Generally, individuals have used compilations of these eye-witness testimonies, which while not as good in terms of quality as the raw data does allow one some evidenced-based investigation that can be evaluated in a rigorous manner. We encourage that all conversations about casualty counts should be based on evidence that is available to individuals so that the estimates can be validated as well as replicated.
One reason for not releasing such information is to protect the memories of those that have been lost. This said, other large-scale mass killings (e.g., the Holocaust, the killing fields in Cambodia and even the Stalinist purges and Mao’s Cultural Revolution) have released data and/or had data discovered that has been evaluated by a variety of different scholars. The public availability of the raw material has assured not only continued, rigorous interest – yielding important and new insights (as different researchers have continually found new things), but it has also assured quality control as the diverse scholars can look over each other’s work and make sure that nothing was done incorrectly. This is how the scientific enterprise works best. This meets the highest standards of social science research and this is the standard that the topic of mass killing deserves. Alternatively, one could anonymize the records to protect individuals but still allow the remaining material to assist individuals with analyzing what happened, where and why. Toward this end, we have several thousand eye-witness accounts from the ICTR and will be making these publicly available after we have finished the redaction of said documents in order to protect named individuals.
A different response to the claim regarding our estimation of Tutsi victims is that almost all individuals acknowledge (implicitly and sometimes explicitly) that everyone “knew” what ethnicity everyone was but this is clearly a local phenomenon. You know who your neighbors were but not those from several villages over. This last point is important because there is another (often unstated) presumption that everyone was killed where they were from but given the large number of internally displaced persons (potentially several million), refugees (potentially several million) and repeated discussions of running as well as hiding, this is unlikely. This also misses the problematic assumption that individuals can identify ethnic others while on the run for their lives, which current research suggests is very difficult to do.
With these concerns in mind, we took information from before the questionable census of 1991 and projected forward diverse population growth rates (which is standard practice in demography) and found figures that were comparable to what was discussed in 1991 therefore allowing us to use it in our estimations. We also employed survivor figures from the Tutsi survivor association Ibuka as well as the census of genocide survivors. We used this information to generate different estimates and then discussed what we believed to be the most reasonable among them given the information that was available. Within this work, there was no point estimate with a single figure – nor should there be. Our research identifies ranges (estimations with +/- error) and typically when the media covers our work they report single figures eliminating nuance, accuracy as well as uncertainty. Across most of our estimations, we generally do not find results that suggest that the Tutsi death toll was higher than the Hutu death toll because of the initial population totals as well as the varied ability to identify ethnicity within a largely fleeing population. All of this information (including the raw source material utilized in these calculations) is publicly available on our genodynamics webpage – a practice that has not been matched by any of those that have questioned our findings.
Evidence to counter our estimation would involve a more definitive population count of Tutsi and Hutu for 1994 (which has not yet been presented), a compilation of id cards distributed among the victims at genocide sites (which was never reported) and a systematic collection as well as evaluation of the witness statements by a neutral investigatory body of all victims and perpetrators both in and outside of Rwanda that could then be evaluated by scholars from around the world to assure competency and quality. To be more definitive on this issue, what is needed in order to provide an accurate assessment of the death toll (by ethnicity) is information on every single killing in Rwanda. This information would identify the perpetrator (by name, ethnicity and organization), victim (by name and ethnicity), space/place (e.g., village) and time (e.g., hour, day and week). With this information, we can more definitively provide an estimate for there will always be error in such calculations as not all of the information noted above would be available. Without this information, however, then the type of estimation that we provide follows the best practices currently available in the social sciences. Counter arguments should at a minimum make their raw material available so any concerned party can work through what was done in a reasonable manner. This would facilitate discussion as well as understanding. If someone has the will but not the resources to share such information, GenoDynamics will provide assistance in scanning, uploading, shipping or programming toward this end.
Christian Davenport – University of Michigan
Allan Stam – University of Virginia