any attempt, conspiracy, incitement, instigation, command or procurement to commit [killing, abduction, torture or severe ill-treatment].”37 The fact that programmes have concentrated on these types of violations is not entirely unjustified. When the resources available for reparations are scarce, choices have to be made and, arguably, it makes sense to concentrate on the most serious crimes. The alternative, namely drawing up an exhaustive list of rights whose violation leads to reparations benefits, could lead to an unacceptable dilution of benefits. No programme has explained why the victims of some violations were eligible for reparations and others not. Not surprisingly and at least in part as a consequence of this omission, most programmes have ignored types of violations that perhaps could and should have been included. These exclusions have disproportionately affected women and marginalized groups. So the mere requirement to articulate the principles or at least the grounds for selecting the violation of some rights and not others is likely to remedy at least the gratuitous exclusions.38 It also bears repeating that decisions about which types of violations will be redressed are taken before the reparations programmes are set up, often when the mandate of a truth commission is settled and in that context. No one will have in mind the consequences these decisions will have on subsequent reparations efforts. Some commissions have found themselves needing to interpret their mandates liberally, so as to include violations that, strictly speaking, were not covered, but that could not reasonably be excluded. This was the case in Morocco and in Brazil. If distinct forms of violence were perpetrated against multiple groups, excluding some of the worst or some of the most prevalent forms of violence or some of the targeted groups automatically makes the reparations programme less comprehensive and, consequently, less complete. The problems generated by this are manifold. Firstly, there is a question of justice, of unequal treatment that could undermine the programme’s legitimacy. Secondly, such exclusions merely guarantee that the issue of reparations will remain on the political agenda, which may threaten the stability of the initiative as a whole.39 37 38 39 Promotion of National Unity and Reconciliation Act (1995), sect. 1(1)(xix)(a), 1(1)(ix). It has often been argued that this definition of “victim” and that of the categories leading to reparations were too narrow to adequately redress the enormity of the violations committed under the apartheid regime. For instance, they overlooked the victims of the routine violence that accompanied the operation of the social engineering aspects of apartheid, people who died in forced removals or were detained under provisions of the state of emergency. See Ruth Rubio-Marín, “The gender of reparations in transitional democracies”, in Engendering Reparations, Ruth RubioMarín, ed. (forthcoming). Perhaps the clearest example comes from Chile, where the exclusion of the victims of torture and the political detainees from most reparations programmes until 2004 (see the section on Chile above) meant that the largest group of victims continued to call for reparations benefits. After Augusto Pinochet’s return from detention in England, when it was clear that criminal prosecutions against members of the military were on the cards in Chile, the Unión Demócrata Independiente (UDI), traditionally supportive of Augusto Pinochet and usually lukewarm on reparations, suggested a major restructuring of the reparations programmes that had been in existence for more than 10 years, so as to expand their coverage and, importantly, significantly increase their benefits. But there was one catch: beneficiaries had to waive all claims against perpetrators. (See, for example, 21