The Failures of Bystanders to Prevent or Stop the Genocide in Rwanda

This article is part of an E-IR series marking the twentieth commemoration of the Rwandan Genocide.

The aim of this contribution is to explain the international response to the gross human rights violations in Rwanda. In our view, there is no other situation in the world since the Holocaust/Shoah that encompasses all elements of the meaning of ‘genocide’ in its entirety as the intentional killings of the Tutsi people in Rwanda in 1994 by Hutu. All Tutsi had to be murdered irrespective of their views, age, sex, or place of residence in Rwanda, simply because they were born as Tutsi. In a hundred days, 800,000 were deliberately killed during a genocide, which was planned, prepared ,and organised by the Rwandan state and its institutions, comparable with the Nazi persecutions of all Jews in Europe. The International Criminal Tribunal for Rwanda decided without any doubt that the prime minister of Rwanda at the time, Jean Kambanda, was to be sentenced to life imprisonment for genocide and conspiracy in genocide.

The Shoah was unique, but genocide has been repeated in the second half of the twentieth century. The similarities in the organisation and preparation of the genocide on Jews and Tutsi has been studied and revealed by Mukimbiri,[1] making use of the eight stages of genocide developed by Stanton.[2] These similarities – particularly the speed at which the atrocities unfolded – are remarkable taking into account the modern, technological, and bureaucratic way the Nazis in Germany prepared and performed the destruction of all Jews, whilst Rwanda was at that time an undeveloped society where a high number of Tutsi killings were performed with primitive means such as machetes.

In our study, based on theoretical aspects in the academic field of international relations, we researched some patterns in the behaviours of third actors at the state and international level of analysis. Both the international (i.e. external) influences for a state and the domestic (i.e. internal) influences were studied in the foreign policy-making of the states. The attention on internal influences makes it possible to study the process of decision-making in both states and international organisations, such as rational decision-making, organisational decision-making, and bureaucratic politics decision-making. This focus on decision-making is required in order to explain the gap between the warnings and the actions.

1. Warnings and Non-responses: A Short Background to the Rwandan Genocide

The warnings for an emerging genocide in Rwanda were manifold. These warning not only came from NGOs, but also from third states and the UN military commanders of the peacekeeping forces. Since the spring of 1992, Belgium, France, the United States (including its CIA), many UN experts, and rapporteurs of the Commission of Human Rights, plus Human Rights NGOs like Amnesty International and Human Rights Watch, had warned of the deteriorating situation and the possibility of genocide. But no one reacted to these outspoken warnings.[3] The most reliable information came from UN generals in Rwanda, which were sent directly to their colleagues in New York. The so-called genocide fax of UN Force Commander Dallaire of January 1993 – with the information about the preparation of the Hutu extremists to exterminate the Tutsi – was addressed to the military advisor of UN Secretary General Baril. Dallaire did not send this message to the UN Department of Peace Keeping Operations (DPKO), but chose the direct line to the fellow Canadian military adviser of the Secretary General. Baril immediately informed Annan and Riza of DPKO. Annan and Riza, however, did not inform the Secretary General, and they also did not inform the members of the Security Council on the threatening situation. They only instructed that the heads of diplomatic missions in Rwanda should be informed. The inquiry commission later concluded that “the seriousness of the threats in the cable justified informing the Council as whole.”[4]

2. Theoretical Explanations for Failure of Third Parties to Prevent the Rwandan Genocide

We’ll discuss the behaviour of the bystanders – the third parties at international level – with five different theoretical perspectives on foreign policy making: 1) Domestic Influences, 2) Rational Policy Model, 3) Organisational Process Model, 4) Bureaucratic Politics Model, and 5) Cognitive Dissonance.[5]

Domestic Influences

The genocide in Rwanda was not prevented or stopped by third parties. Domestic influences on the foreign policy-making of the third parties were absent before and during the genocide in Rwanda. There was no public pressure that activated Western governments. At the height of the genocide, the American official who was the head of the African Division of the State Department even tried in vain to mobilize an NGO, such as Human Rights Watch.[6] The little domestic influence there was operated in the opposite direction, in favour of non-involvement in Rwanda. That is to say, when ten Belgian peacekeepers were deliberately murdered at the start of the genocide, it had far-reaching political consequences, because public opinion no longer supported the continuation of the peacekeeping mission. Hence, this sly act of the perpetrators facilitated the – almost total – withdrawal of the peacekeeping force from Rwanda. Public support was not only lost due to the Belgian fatalities, but also due to the still fresh memory and reminder to the Americans of the murder and slaughtering of American (and Pakistani) peacekeepers in Mogadishu, Somalia, one year before. Withdrawing the peacekeepers meant that a buffer was taken away and the genocide could continue without any hindrance of a third party.

Rational Policy Model

The rational policy model, with its focus on cost-benefit analyses, is well placed to explain the warning-response gap and the near absence of efforts by the international community to prevent the genocide. Most analyses of the international response to intra-state conflicts and genocides have focused on the lack of political will and the absence of strategic interests. Such explanations fit well with the rational policy model. Samantha Power concluded in her study on the US response to several genocides in the 20th century that US policy-makers pursued two goals. On the one hand, avoid becoming involved in conflicts that do not or hardly affect American interests. On the other hand, minimise the political costs and moral shame by giving the impression that the maximum achievable is being done. The costs of avoiding engagement were significantly lower than the costs for increasing the involvement.[7]

The rational actor model can also explain the silence of France in the Security Council during the genocide in Rwanda, because a Hutu victory was a French interest in Africa. France had even a rescue operation, Opération Turquoise, for the Hutu militias authorised by the Security Council after the genocide on June 22, 1994. The rational actor model also explains why the other international actors were primarily interested in rescuing only their own nationals and officials. At the start of the genocide in Rwanda, a huge fighting force arrived to evacuate mostly Western foreign nationals. A total of over 1,700 elite troops from the US, France, Italy, and Belgium were either flown in or put on standby in neighbouring countries immediately after the attack on President Habyarimana’s plane when the evacuation of their nationals was ordered.[8] If these 1,700 well-armed and trained elite troops had been added to the 2,500 UNAMIR soldiers, the total number of troops on the ground in Rwanda would have been 4,200 – exactly the number of soldiers all Rwandan parties to the Arusha Peace Accords had asked for in 1993 and the number that was considered realistic by the military who prepared the peacekeeping mission.[9] A possible combination of this strong military force with the weaker UN peacekeeping force was never tabled in the decision-making processes of any Western capital or at UN Headquarters – a clear signal for the génocidaires that no one from the outside world would oppose the genocide.[10]

Organisational Process Model

The organisational process model is very apt in explaining the early-warning-response gap, because it focuses on bureaucratic cultures within states and international organisations. The most prominent example of work that is focused on bureaucratic culture is the writing of Barnett on the genocide in Rwanda.[11] A bureaucratic culture produces powerful and autonomous bureaucrats, whereby rules and procedures, such as the principles of neutrality and impartiality, become ends in themselves rather than the means to realise a certain objective. The model explains very well why the peacekeeping forces maintained the strict and limited standard operating procedures and routines for keeping the peace, even though the situation on the ground was rapidly deteriorating. In the months preceding the genocide in Rwanda, all early warnings were received by the UN, but they failed to effect a change in the decisions of the UN Secretariat. The Secretariat remained unyieldingly committed to thinking in terms of peace and security and the post-conflict transformation process of installing a new multi-ethnic government.[12] At the outbreak of the genocide, the peacekeeper commander general Dallaire phoned five times with the heads of the DPKO in New York (Annan and Riza), but they decided to prohibit the use of force to give safety to the members of the moderate Rwandese government. They reacted according to the organisational process model as if it was a routine decision, even though the Rwandan prime minister and members of his government were killed. In their bureaucratic vision, the UN had to maintain a traditional neutral and impartial peacekeeping role and not take sides against the perpetrators of the impending genocide with forceful action, as was requested by, for instance, the Canadian UNAMIR commander Roméo Dallaire and the Belgian government. The typical answer derived from this organisational process model is of withdrawing almost all peacekeepers when some soldiers were murdered. The soldiers could no longer stay without the consent of the parties as being one of the conditions for peacekeeping operations. Changing one’s views when the situation has changed is not within the limited routinised way of thinking that is characteristic for this organisational process model. Another consideration which prevented the Secretariat to inform the UNSC was the estimation or anticipation that the Security Council would not authorise any (enlarged) intervention anyhow.[13]

Neutrality and impartiality became ends in themselves rather than the means to realise a certain objective. It is not unreasonable to argue that genocide requires a total opposition from the United Nations, because genocide can be seen as threatening everything the UN has stood for, as laid down in its Charter. But this did not happen. Instead, the UN and UN officials were primarily concerned with the interests of the UN. A UN official in New York stated, “I was more committed to the survival of the UN than I was to the Rwandans.”[14]

Bureaucratic Politics Model

This is the model par excellence to explain the behaviour of the UN in the Rwanda genocide. The Security Council did not adopt any resolution or take any measure to stop the atrocities in Rwanda. On the contrary, DPKO, under the direction of the future SG Kofi Annan, did not provide the full information to the members of the Security Council for fear of harming their own administrative unit DPKO. They misinformed the Security Council and they did not provide the perspective that the peace was no longer to be maintained and restored, but an emerging genocide with slaughtered peoples had to be feared.[15] DPKO especially feared that a failure in the operation after Somalia may endanger the existence of this administrative unit and future peacekeeping operations across the world. All instructions on Rwanda were in accordance with the rules of DPKO as just elaborated, and all events and information were coloured to benefit their sub-organisation. They did their utmost to distort all incoming information by rejecting all that might have changed the views that the peace could no longer be kept, that the Rwandese rulers were preparing a genocide.

Cognitive Dissonance

Cognitive dissonance underlines the way actors in the decision-making avoid and discount information that is inconsistent with their prior views and which would require them to adjust their views. There is consequently no adjustment of the original plan of action. UN officials continued to perceive the situation as a peacemaking process involving the installation of a transitional government, and not as an emerging genocide. The UN Secretariat – and in particular the DPKO – therefore favoured a neutral, rather than a confrontational, position. A shift in perception from facilitating the implementation of a peace accord toward preventing an emerging genocide was needed, but no such shift took place.[16] Instead, information inconsistent with the maintained policy of the administrative unit was neglected or changed in such a way to make it consistent with the interests of the unit. For instance, the worsening situation on the ground was not labelled as an emerging genocide, but as a civil war or tribal struggle.

Another example of cognitive dissonance was that Security Council members did not realise and could not believe that their neighbour in the Security Council was a representative of the génocidaire Hutu regime in Rwanda. Actors also tended to filter incoming information through previously held views and expectations. Bystanders find it difficult to believe the unbelievable or imagine the unimaginable. As a result of this disbelief or denial, atrocities were not recognised or were put in a distorted way to fit existing world-views. Similarly, actors may suffer from wishful thinking, convincing themselves that negative developments are less or unlikely to occur, while they are considerably more optimistic about the likelihood of positive outcomes.

Another psychological mechanism used by actors was a normalisation technique. Several actors relied on stereotypical and racial images. Killings in Africa were not seen as especially unusual. US Acting Assistant Secretary of State for African Affairs Bushnell stated, “People didn’t know that it was a genocide. What I was told was, ‘Look, Pru, these people do this from time to time.’”[17] A report of the Inquiry Commission of the AU afterwards also pointed to “the implicit racism” and the “sense that African lives are not valued as high as other lives,” which was visible in the priority given by New York to the peacekeepers in Rwanda to help the rescue of expatriates, even beyond its original mandate.[18] The conflict in Rwanda was also described as ‘chaotic, mad and tribal.’ One consequence of portraying it as an impenetrable Somalia-like chaos with little possibility of resolution is that actors are relieved from any duty to act. Another way of diminishing the urgency to act is by describing the situation in Rwanda as a civil war instead of genocide, which may help bystanders at all levels to disregard the cruel atrocities of the perpetrators.[19]

3. Conclusion

In sum, gross human rights violations were not prevented in Rwanda. In our forthcoming book on Darfur, which also includes a comparative study of Rwanda, Srebrenica, and Darfur, we conclude that Rwanda was not an exception. No effective measures were taken by third parties to stop the atrocities in all three cases. In the book, we depict the role of the bystander in situations of gross human rights violations as burying one’s head in the sand. This happened despite the leading motive that is reiterated after each genocide: that such atrocities should “never again” be allowed to happen. What is the value of such repeated pledges? The statement on Darfur by one of the principle UN decision-makers on Darfur, the former head of the UN Department of Political Affairs Prendergast, is striking:

We don’t mean it when we say that we’re not going to accept other Rwandas, further Rwandas. But I never thought we did mean it. It’s a very sad conclusion, but I don’t think there’s any evidence to sustain the view that we did mean it. We may have meant it at a level of generalized indignation, but when it comes to accepting the consequences of that, we don’t.[20]

Author’s note: This article is based on chapter 3 of the publication Failure to Prevent Gross Human Rights Violations in Darfur; Warnings to and Responses by International Decision Makers (2003-2005) – written by Fred Grünfeld and Wessel N. Vermeulen, in cooperation with Jasper Krommendijk – to be published in June 2014.

[1] Mukimbiri, J. (2005) “The seven stages of the Rwandan Genocide”, Journal of International Criminal Justice, Vol.: 3(4), pp. 823–836.

[2] Stanton, G. H. (1996) The 8 stages of genocide, Genocide Watch, Washington D.C. available at: (last visited November 16, 2013).

[3] Grünfeld, F. and Huijboom, A. (2007) The Failure to Prevent Genocide in Rwanda: The Role of Bystanders. Leiden, Boston: Martinus Nijhoff Publishers, Brill, pp. 61—126.

[4] United Nations Report of the Independent Inquiry into the Actions of the United Nations during the 1994 Genocide in Rwanda, 15 December 1999, United Nations, Ingvar Carlsson a.o., un Doc.: S/1999/1257, p. 33.

[5] For a further elaboration on these theories, see Allison, G.T. and Zelikov, P. (1999) Essence of Decision: Explaining the Cuban Missile Crisis. 2nd edition. Pearson Publishers.Mintz, A. and DeRouen Jr., K., (2010) Understanding foreign policy decision making. Cambridge: Cambridge University Press. Stein, J. (2012), “Foreign policy decision making: rational, psychological, and neurological models” in Foreign policy. Theories, actors, cases, by Smith, S., Hadfield, A. and Dunne, T. (eds.), Oxford: Oxford University Press, pp. 130–137.

[6] Authors’ interview with Prudence Bushnell, US State Department, Africa Desk, 27 May 2005 (Fred Grünfeld).

[7] Power, S. (2007 [2003]) Problem from Hell: America and the Age of Genocide. New York: Perennial, London: Flamingo, pp. 508-509.

[8] Grünfeld and Huijboom (2007), pp. 167—178.

[9] Grünfeld and Huijboom (2007), p. 177.

[10] Smeulers, A. and Grünfeld, F. (eds.), (2011) International Crimes and other Gross Human Rights Violations. A Multiand Interdisciplinary Textbook. Leiden, Boston: Martinus Nijhoff Publishers, Brill, p. 398-399.

[11] Barnett, M. (2003) Eyewitness to a genocide: The United Nations and Rwanda. Ithaca, N.Y.: Cornell University Press.

[12] Ibidem, pp. 166 and 255.

[13] Ibidem, p. 15.

[14] Barnett, M. (1997) “The UN Security Council, indifference and genocide in Rwanda,” Cultural Anthropology, Vol.: 12(4), p. 573.

[15] Grünfeld and Huijboom (2007), pp. 249-261.

[16] Grünfeld, F. and Vermeulen, W.N. (2009) “Failures to Prevent Genocide in Rwanda (1994), Srebrenica (1995) and Darfur (since 2003)”, Journal of Genocide Studies and Prevention, Vol.: 4(2), pp. 221–238, p. 233.

[17] Power (2007), pp. 351, 365. See also Suhrke, A. and Jones, B. (2000) “Preventive diplomacy in Rwanda: Failure to act or failure of actions?”, in Opportunities missed, opportunities seized: Preventive diplomacy in the post-Cold War world, by Jentleson, B. W. (ed.), Lanham: Rowan & Littlefield Publishers, p. 256.

[18] Organization of African Unity, The International Panel of Eminent Personalities to Investigate the 1994 Genocide in Rwanda and the Surrounding Events, available at: [accessed 18 December 2013], para. 21.15.

[19] Piiparinen, T. (2008) “The Rise and Fall of Bureaucratic Rationalization: Exploring the Possibilities and Limitations of the un Secretariat in Conflict Prevention”, European Journal of International Relations, Vol.: 14(4), pp. 697–724, p. 719; Piiparinen, T. (2010) The transformation of un conflict management. Producing images of genocide from Rwanda to Darfur and beyond. London, New York: Routledge, p. 76.

[20] PBS Frontline interview with Kieran Prendergast, 29 June 2007, available at: [accessed 31 August 2013].

Further Reading on E-International Relations

Please Consider Donating

Before you download your free e-book, please consider donating to support open access publishing.

E-IR is an independent non-profit publisher run by an all volunteer team. Your donations allow us to invest in new open access titles and pay our bandwidth bills to ensure we keep our existing titles free to view. Any amount, in any currency, is appreciated. Many thanks!

Donations are voluntary and not required to download the e-book - your link to download is below.


Get our weekly email