. 4
( 11)


The occasional involvement of Belgian courts in punishing alleged perpetra-
tors of genocide in Rwanda continues under Belgian legislation that permits its
courts jurisdiction to prosecute certain extraordinary international crimes com-
mitted outside Belgium when the accused is a resident of Belgium. On June 29,
2005, two Rwandan businessmen were convicted by a Belgian jury of aiding and
abetting war crimes and were sentenced to twelve and ten years.
A silent irony lurks in the Belgian judgments in that they do not demonstrate
introspection regarding Belgium™s colonial involvement in Rwanda, in which
Belgium was a perpetrator of systematic rights abuses, and through which it also
created conditions that eventually facilitated genocide in 1994. Belgian colonial
administrators took a liking to the Tutsi, whom they treated preferentially. This
angered the majority Hutu. In 1933, the Belgian colonial administration passed a
law requiring every Rwandan to carry an ethnic identity card. The lines between
Tutsi and Hutu, which traditionally had been porous and informal, suddenly
became permanent and legalized. The ethnic identity card requirement per-
sisted after Rwandan independence in 1960. Tragically, the continued presence
of this requirement accelerated the genocide, insofar as persons unable to pro-
duce a Hutu card simply were slaughtered. Nor do the Belgian courts refer
in their judgments to Belgium™s role in failed international peacekeeping in
Rwanda during the genocide.
The Belgian courts do not justify their interventions on the basis of Belgium™s
repaying a debt to Rwanda for historical events. The ¬‚attened historical narra-
tives that emerge from individualized criminal trials, especially those conducted
far away from Rwanda, do not elucidate the gnarled, and deeply complex, roots
of responsibility for genocide. Instead, criminal trials permit the former colonial
state to cleanse its wrongdoing and appear heroic in its quest for justice. These
monodimensional and partial narratives pose a challenge to the expressive value
of trials and punishment.
To be sure, there are times where proceedings conducted far away can be
catalysts in the process of accountability at home, where they may help pry loose
information that is deeply buried.75 Arguably, this was the case with regard to
extraterritorial prosecution of General Pinochet of Chile. Alternately, extrater-
ritorial prosecutions can provide some justice when the territorial state™s appa-
ratus remains repressive, which is the case with regard to Spanish proceedings
involving atrocity in Guatemala. In the case of Rwanda, however, the upshot
Punishment of International Crimes Nationally and Locally 85

of extraterritorial prosecutions is more dif¬cult to discern and the interface of
these proceedings with broader notions of justice remains quite complex.

c. Gacaca
Gacaca, which means “justice on the grass” in Kinyarwanda,76 is a traditional
method of dispute resolution. The Rwandan government has turned to this tra-
dition, which it has signi¬cantly adapted, to promote accountability for offenses
related to genocide and crimes against humanity committed between October 1,
1990, and December 31, 1994. With regard to these adapted proceedings, gacaca
judges are elders and “people of integrity” (Inyangamugayo) elected from local
communities throughout Rwanda. In toto, 170,000 judges sit on approximately
10,000 panels. The panels are composed at the lowest level (that of the cellule)
of nine judges with ¬ve deputies.77 There are two higher levels of panels at the
secteur and appellate levels (each of these two levels has about 1,500 panels).
All panels are to apply the same substantive criminal law that is applied by the
national courts in proceedings related to genocide and crimes against humanity.
Suspects are brought to the communities where they are alleged to have commit-
ted their crimes to face villagers and judges elected from the community. The
notion of community in postgenocide Rwanda is dynamic, insofar as the com-
munity in the village that adjudges perpetrators generally does not correspond
to the community that had been present at the time the crimes allegedly had
been perpetrated. Many communities have become recomposed in the wake
of the genocide owing to refugee movements, mass killings, internal displace-
ments, immigration of Tutsi from Uganda, and government-driven resettlement
programs.78 This does not denude gacaca for genocide of its communitarian
ethos, but suggests that the relationship of local gacaca initiatives with locally
assembled populations certainly is nuanced.
Practically speaking, the decentralized nature of the gacaca process facili-
tates access to justice by reducing transportation costs for witnesses and victims,
which has been cited as a shortcoming for the national trials and, especially, for
the ICTR. Public involvement also is encouraged insofar as the proceedings are
conducted in Kinyarwanda and businesses close (albeit by governmental order)
on the days gacaca is in session. At the proceedings, the public (the General
Assembly) can raise issues “ discursively “ that exceed the microscopic truths
that would arise at trials. Members of the public can ask questions of suspects,
to which suspects are permitted to reply. However, the judges are empowered
to control the discussion, the ¬‚ow of evidence, and maintain order at the pro-
ceedings. In the end, although the judges primarily adjudicate, they also act as
mediators to help the gathered community attain both legal and extralegal goals.
Lawyers are excluded, purportedly to ensure the open, participatory nature of
the proceedings. Judges are laypersons who do receive limited legal training.
In practice, gacaca for genocide ¬rst began haltingly. The process was subject
to numerous delays. A number of panels, however, began proceedings in March
2005.79 In January 2006, it was reported that 4,162 individuals had been adjudged,
86 Atrocity, Punishment, and International Law

142 of whom were women.80 Prior to appearing in gacaca courts, certain parolees
had spent time in reeducation camps. All gacaca panels have undertaken inves-
tigations. Once proceedings begin in earnest throughout Rwanda, estimates
vary widely as to how long it would take to process all detainees. Realistically
speaking, it may take up to ¬ve or six years.
The government established gacaca courts for several reasons. One is man-
agerial. After all, many detainees have been incarcerated since 1994 awaiting
a putative trial date; given the pace of trials in the national courts, it could
take as long as a century to clear all the cases. A second stated reason is to
diversify the legal response to genocide by invoking mechanisms more steeped
in reconciliation, reconstitution, and reintegration “ each of which resonates
in Rwandan sociolegal culture. This diversi¬cation, in theory at least, would
move the focus away from retribution, in particular with regard to lower-level
offenders. Consequently, the Rwandan government touts both retribution and
reconciliation as goals of gacaca adjudication; and the gacaca framework notes
the importance of penalties that permit convicts to “amend themselves” and
reintegrate into Rwandan society.81 A third reason is participatory “ to involve
the public in adjudication and discussion of genocide. A fourth reason is to “dis-
close the truth” (although the gacaca process, whether traditional or in modi¬ed
form for the genocide, is not a truth and reconciliation commission).82 And a
¬fth reason is one of sovereignty, namely the Rwandan government™s percep-
tion that Rwanda needs to develop “by itself” solutions to the genocide and its
These neotraditional gacaca courts initially were established in 2001 by virtue
of an Organic Law.84 In 2004, the 2001 Organic Law for gacaca courts was signi¬-
cantly amended.85 The amended law, which collapsed and simpli¬ed elements
of the preceding framework, categorizes offenders and punishments. In this
regard, the approach is similar to that of the 1996 Organic Law for the Special-
ized Chambers of national courts, although there are important differences. For
example, the form of punishment under gacaca includes incarceration as well
as community service (travaux d™int´ rˆ t general), the length of sentence overall
is slightly shorter than that set out in the 1996 Organic Law, and sentencers are
given somewhat broader discretion. These changes “ in particular with regard
to punishment “ are important insofar as it appears that, from now on, detainees
will be processed through the 2004 gacaca legislation.
Article 51 of the 2004 gacaca legislation creates three categories of offenders.
These are:

r Category 1: planners, leaders, notorious murderers, torturers (even when not
resulting in death), rapists and sexual torturers, and those who committed
dehumanizing acts against a dead body (in all cases, actual perpetrators and
accomplices are implicated);
r Category 2: (1) murderers; (2) those who committed attacks with the intention
to kill but did not succeed; and (3) those who committed other offenses against
the person without the intention to kill;
Punishment of International Crimes Nationally and Locally 87

r Category 3: those who committed property offenses (an offender in this cate-
gory cannot be prosecuted if there is an agreement between the offender and
the victim to settle the property harms caused).

Category 1 offenders are excluded from the local gacaca panels.86 They are
to be prosecuted more formally. Initially these prosecutions were to occur in
the ordinary national court system. However, in November 2005, a new special
court was created to hear these prosecutions. Local gacaca panels will only hear
those cases involving Categories 2 and 3.87 They will therefore have jurisdiction
over those who killed (even intentionally), who assaulted persons, and who
committed property offenses. That said, the gacaca law does create punishments
for Category 1 offenders. This is because the determination of new Category 1
offenders (and the corroboration of evidence against others) can be made in the
information-gathering pretrial stages of gacaca. Looking ahead, those individuals
thusly determined to be in Category 1 shall, although processed by the new
special court reconstituted following major legal reforms, apparently be entitled
to the sentencing scheme established for them by the gacaca legislation, which
differs from the sentencing scheme established by the 1996 Organic Law for the
Specialized Chambers of national courts.
In all cases, investigations and compiling of evidence are carried out by the
lowest-level gacaca panel, namely that of the cellule.88 The community thereby
becomes involved in developing lists of individuals accused of crimes and also in
corroborating or removing charges the prosecution may have previously brought
against parolees. It is through this process that an accused can be placed in a
certain category (or removed therefrom). It is only when the investigatory and
pretrial stages are completed that the gacaca panels adjudicate wrongdoing.
Adjudication of Category 3 suspects occurs at the cellule level and Category 2
suspects at the secteur level.
The 2004 gacaca legislation provides a very detailed punishment schematic.
It also meshes punishment with a confession and plea bargain regime that bears
some similarities with, although also expands upon, that of the 1996 Organic Law
for the Specialized Chambers. As set out in article 54 of the 2004 legislation, the
focus is on confessions, pleading guilty, apologies (made publicly to surviving
victims and to Rwandan society), and repentance; there also is a requirement to
provide information regarding the whereabouts of victims™ remains. The extent
of the sentence discount is motored by when the accused confesses: namely,
whether the confession is approved before the accused™s name appears on a
list drawn up by the gacaca courts in their investigative functions, or after. The
General Assembly can reject an incomplete or insincere confession.
Article 72 states that Category 1 offenders who refuse to confess, or
whose confessions have been rejected, incur either the death penalty or life
imprisonment.89 Given the current attitude of the Rwandan authorities toward
the death penalty, it is likely that such offenders de facto will face life imprison-
ment. Category 1 offenders who confess as provided by the law incur a prison
sentence ranging from twenty-¬ve to thirty years.
88 Atrocity, Punishment, and International Law

Category 2 offenders who kill or who commit serious attacks with the intent
to kill, and who either refuse to confess or whose confessions have been rejected,
incur a sentence ranging from twenty-¬ve to thirty years.90 Those who confess
after their names have appeared on the list compiled by the relevant cellule-level
gacaca court incur a sentence from twelve to ¬fteen years, but out of this sentence
they only serve half of their time in custody and the remainder is commuted
into travaux d™int´ rˆ t general (community service). Those who confess before
the list is drawn up incur a prison sentence ranging from seven to twelve years,
half served in prison and half in community service. These sentence discounts
are quite striking. The purpose of discounting sentence for persons who come
forth and turn themselves in before investigations implicate them in atrocity and
place them on the list is to save resources and encourage truthfulness.
Category 2 defendants who committed offenses against the person without
the intention to kill face a term of imprisonment within the following ranges:
¬ve to seven years if they refuse to confess, or if the confession is rejected, half of
which is in community service; three to ¬ve years if they confess after the list is
drawn up, half of which is in community service; and one to three years if they
confess before the list is drawn up, again half of which is in community service.
Members of Category 3 “ those accused of property offenses “ only can be
sentenced to civil reparations for the damages they caused.91 In the absence of
an agreement concluded between the perpetrator and the aggrieved parties, the
gacaca court quanti¬es the reparation due.
In addition to the recategorization and the simpli¬cation of the panel struc-
ture, the 2004 gacaca legislation effected two notable changes to the 2001 gacaca
legislation that pertain directly to punishment. Whereas article 69 of the 2001
legislation sentenced a Category 2 offender who either did not plead guilty or
whose plea was rejected to a sentence of twenty-¬ve years or life, the 2004 legis-
lation sets a maximum sentence of thirty years™ imprisonment for a Category 2
offender. Second, under article 75 of the 2001 legislation, community service
was cast as an option for the convict, namely, something that the convict could
choose to do; in fact, the convict was free not to elect to serve half the sentence
in community service but spend the whole sentence in prison. The 2004 legisla-
tion appears to eliminate the hitherto optional character of community service,
thereby making it a mandatory component of many sentences. The motivations
for these two changes to the punishment scheme are not readily ascertainable
from the text or preamble to the legislation. They may re¬‚ect, on the ¬rst part,
a move away from retribution and, on the second part, a desire to coax offender
reintegration and victim restoration through labor.
In sum, gacaca offers a more diversi¬ed array of punishment than the Special-
ized Chambers of the Rwandan conventional courts, and certainly more so than
international criminal tribunals. The community service aspect is the central
vehicle for this diversity. In theory, this service might include tilling the ¬elds of
victims, donating produce or labor, obliging other members of the perpetrator™s
family to help the aggrieved family, constructing roads, and renovating houses
partially destroyed during the genocide or building new houses for survivors.92
Punishment of International Crimes Nationally and Locally 89

That said, it is too early in the practice of gacaca to make de¬nitive assessments of
the quality and form of community service projects, although the compensatory
value appears to be underactualized.
Time spent in community service is conditioned on the convict™s not com-
mitting another crime. Moreover, if a convict defaults on the community ser-
vice commitments, then the time remaining on the sentence is to be served
in custody.93 It remains unclear exactly how the service shall be monitored
and default determined. Monitoring costs could in fact be quite high. Default
claims could tie up the gacaca system. On the other hand, if gacaca judges
impose lengthy jail sentences (and too readily ¬nd default), then the problems
of prison overcrowding and endless proceedings that plague the Rwandan legal
and correctional system simply will reappear. In terms of the law-in-practice of
gacaca, it is important to recognize that many suspects have been detained “
some for over a decade “ awaiting adjudication. In the event this pretrial deten-
tion counts toward any eventual sentence, a large number of individuals would
simply be released for time served.
D´ gradation civique also is contemplated as a sentence. According to arti-
cle 76 of the 2004 Organic Law, for those convicted of Category 1 offenses
this is perp´ tuelle et totale. Certain Category 2 offenders also are subject to
d´ gradation civique. The 2004 Organic Law narrows the scope of convicts sub-
ject to d´ gradation civique from the 2001 Organic Law.94
The gacaca law also provides for restitution or repayment of looted or
ransacked property, or carrying out the work required for the property to be
repaired.95 This is distinct from community service. At the initial gacaca pro-
ceedings, genocide survivors ¬lled out forms requesting compensation, although
perpetrators largely are illiquid and the Rwandan government has proven reluc-
tant to commit funds. Given the great dif¬culty in enforcing the partie civile
damage awards that emanate from the conventional court system, it may well
be that the gacaca legislation™s permitting offenders to carry out work to
repair what they had destroyed becomes a more realistic method of restora-
tion (although this, too, may be subject to tremendous monitoring costs, along
with the potential for corruption and the possibility of involuntary servitude to
private parties). For a society such as Rwanda, in which tens of thousands of
families have been orphaned and for many years have been headed by children,
¬nancial reparation is not just a matter of commemoration or symbolic justice.
It also could prove essential to rudimentary quality of life. That said, not all
Rwandans wish to receive money or property as some sort of compensation for
the loss of their loved ones.96 This does not mean that they necessarily eschew
the civil liability process. They may welcome its expressive and didactic value,
as discussed earlier.
The 2004 Organic Law makes some reference, albeit quite parsimonious, to
aggravating and mitigating factors. As for aggravating factors, article 52 states that
position of leadership, which is a constitutive element of the categorization of
the offender™s degree of criminal liability, also is a factor that could expose the
offender to the most severe punishment within the appropriate category. With
90 Atrocity, Punishment, and International Law

regard to mitigating factors, the statutory framework makes great allowance for
the process of confessing, pleading guilty, and apologizing. Minority, too, leads
to discount and is explicitly referenced in a manner that is much more detailed
than the statutory framework for the Specialized Chambers. Article 78 provides
specially reduced punishment for minors between the ages of fourteen and
eighteen at the time of the events.97 Minors under the age of fourteen at the
time of the events cannot face prosecution but can be placed in special solidarity
The Organic Law is supplemented by a special manual that the Rwandan
government has created for gacaca judges (Manuel explicatif sur la loi organique
portant cr´ ation des juridictions gacaca).98 This document, initially created for
the 2001 Organic Law, provides additional instruction for judges on how to
conduct hearings. It also summarizes penalties for offenders and enumerates
certain aggravating and mitigating factors in sentence (only 3 of over 100 pages
are devoted to sentence). Having an authority position in the cellule is listed as
an aggravating factor.99 The Manuel explicatif also mentions as aggravating the
fact that the offender may have been sentenced to concurrent convictions: in
such a case, the sentencer is to exercise discretion (if available) to award the most
severe sentence in the permissible range.100 As for mitigating factors, the Manuel
explicatif mentions vulnerability (faible capacit´ d™esprit), undue in¬‚uence (forte
in¬‚uence subie), and whether the accused may have saved the lives of other
victims.101 In all cases other than what the Organic Law provides for minority and
confessions, the incorporation of mitigating factors is a discretionary exercise.
This discretion is fettered by article 81 of the 2004 Organic Law, which precludes
judges from dipping below the minimum sentence statutorily provided.
Other than these factors, the positive law frameworks provide no explicit
guidance to gacaca judges as to how they ought to exercise their considerable
discretion in sentencing within the prescribed ranges (a discretion that exceeds
that provided by the 1996 Organic Law). Furthermore, other than a few lines in
the preamble, and extrinsic sources, the gacaca legislation makes no mention
of the rationales for punishment generally. Although the gacaca court must
present its reasons for judgment, it is under no obligation to present reasons
for the sentence; the only requirement is for the penalties to be pronounced.102
Despite the importance of community service, it is unclear whether judges
can give voice to penological rationales, say retribution or restoration, through
the choice of which kinds of service projects to assign to particular offenders.
Nor is there any guidance regarding which sorts of projects intrinsically are more
restorative, reparative, deterrent, or retributive; or how to differentiate modalities
of community service from each other based on these different aspirations (in
other words, what factors make a particular project more restorative than another,
or more retributive than another). In the absence of such a schematic, it is
unclear how the purposes of sentencing can be coherently individuated for the
actual convict or victims implicated in a speci¬c case.
Insofar as gacaca courts have only just begun their operations, there is
limited law-in-practice. Statistics compiled by the Rwandan government in
Punishment of International Crimes Nationally and Locally 91

June 2005 from initial gacaca adjudications throughout the country reveal that in
all but one of twelve jurisdictions the maximum penalty of thirty years had been
issued.103 This is unsurprising insofar as the early gacaca proceedings involved
more serious offenders, although many of them had confessed.104 Minimum
penalties issued range from one year to ¬ve years. Community service was
ordered in about 45 percent of all judgments. In cases that have been adjudged,
approximately 12 percent of defendants experienced a change in their category
classi¬cation for reasons that remain unclear. The acquittal rate, reported in
January 2006, was 12 percent (caveat: many initial cases involved confessions
and guilty pleas) and about 25 percent of all judgments were appealed.105
ASF reports limited discussion of the rationales for punishment at the initial
gacaca sessions its observers attended. It also reports that judges sentenced at the
upper bands of the permissible range, which suggests that they exercised their
discretion to impose the longest sentences possible and underplay mitigating
factors.106 Gacaca judges did not regularly award d´ gradation civique in the
initial proceedings. With regard to those sentenced by gacaca to community
service, Human Rights Watch reports that the 750 individuals sentenced by
September 2005 all were brought to work in one place (akin to a labor camp).108
This seems to be at odds with traditional gacaca™s goal of diffuse local restitution
or direct victim compensation.
It is now estimated that at least 760,000 individuals “ perhaps even 1,000,000
(which would be over ten times the extant prison population) “ eventually
may end up facing a gacaca court.109 Evidence emerging from the investigatory
phases of gacaca proceedings suggests much broader levels of public partici-
pation in the Rwandan genocide than what many had previously believed (or
wanted to believe), although a number of observers “ myself included “ consis-
tently have maintained the populist nature of atrocity in Rwanda.110 Through
its investigations, gacaca may be unmasking these broad levels of complicity
and the identi¬cation of perpetrators through public denunciations. Assuredly,
some of the testimonial evidence proffered by detainees and accusers is likely
unreliable, dated, uncorroborated, untruthful, and motored by ulterior motiva-
tion. But not all of it is so, and obviously the gacaca process will afford some
occasion to verify the veracity of this evidence. In this regard, by permitting
the adduction of evidence that expands the breadth of accountability for the
Rwandan genocide, gacaca could distribute blame more evenly among those
The prospect that gacaca might implicate an additional one million people
in genocide in Rwanda is deeply troubling to some observers. To be sure, this
prospect presents signi¬cant administrative and bureaucratic challenges. It is
unclear whether any system can accommodate such a volume of cases, espe-
cially in Rwanda where limited resources already are strained by the drive for
accountability and where many suspects face substandard conditions of deten-
tion. Moreover, there is cause to fear that the state may turn to gacaca as an
instrument to intimidate opposition. However, some of the concerns voiced by
observers transcend the managerial or political. William Schabas, for example,
92 Atrocity, Punishment, and International Law

¬nds that implicating one million people in genocide is a “terrible and totally
unexpected result [ . . . that . . . ] ha[s] opened a Pandora™s box [ . . . ] Charging
1,000,000 Rwandans with genocide amounts to an indictment of perhaps one-
third of the country™s adult population.”111
But what if one-third of the country™s adult population actually was involved “
whether as perpetrator, accomplice, pro¬teer, or bene¬ting bystander “ in the
1994 genocide? It may well be that expanding the array of suspects more accu-
rately re¬‚ects the popular nature of genocide in Rwanda especially when, as
gacaca does, property crimes and pro¬teering are included. I agree with Schabas
that detaining suspects in “appalling conditions” is deeply problematic. I also
acknowledge that managerial concerns arise with such a volume of cases and
that there is evidence the government is deploying gacaca for social control
purposes. Certain of the denunciations are politically motivated and, as is the
case with all legal matters, some are unsubstantiated. However, these pragmatic
concerns should not dissuade the contemplation of processes, such as gacaca,
that have some potential to deracinate and examine the structural nature of
genocide in Rwanda.
Frankly, one of the reasons why many observers are fearful of one million
Rwandans facing gacaca is because gacaca for genocide looks and acts more
like a liberal criminal court than what it traditionally is, namely a communal
restorative mechanism. Community-based informal justice focusing on restora-
tive and reintegrative shaming has a long history throughout Africa and takes
various names and forms (e.g., lekgotla and inkundla in South Africa).112 Gacaca
shares in this history. However, gacaca as set out in the 2004 Organic Law differs
from traditional gacaca:

Charles Ntampaka, one of the leading experts on Rwandan customary law,
observed that the traditional system of con¬‚ict resolution did not include
any written rules; remained wary of legal prescriptions that adjudicate and
convict; was closely related to the family unit; favored the role of the “head
of the family”; involved forms of collective responsibility; did not promote
equality; gave priority to community interests over individual rights; often
deemed confessions to be a form of provocation; and drew on the sacred and
the religious. [ . . . ] Such characteristics are in stark contrast to the present
gacaca courts and their functioning.113

The structure of the genocide gacaca tribunals and the conduct of their
trials therefore operates somewhere between traditionalism and liberal legal-
ism. The movement, though, has been from the former to the latter. In effect,
gacaca for the genocide is more like a court than its customary nature of com-
munal gathering. It is more formal than informal. Article 39 of the Organic
Law explicitly states that “[g]acaca courts have competences similar to those of
ordinary courts,” including the power to issue subpoenas and search warrants,
summon witnesses, and con¬scate goods. Although lawyers do not represent par-
ticipants, the gacaca tribunals are counseled by appointed conseillers juridiques
(legal advisers). A detailed appellate structure, including for sentencing appeals,
Punishment of International Crimes Nationally and Locally 93

is created.114 Moreover, public participation in gacaca is not really voluntary.
According to article 29 of the 2004 Organic Law, every Rwandan citizen has the
duty to participate in the gacaca courts. A citizen can be sanctioned for refusing
to testify.115 In some cases, local people who fail to attend gacaca hearings have
been punished. Whereas traditional gacaca excluded women from decision-
making, and thereby was a deeply patriarchal institution, gacaca as envisioned
by the 2004 Organic Law is encouragingly inclusive of women, including in the
capacity of judges. That said, other aspects of traditional gacaca of which liberal
legalism might be suspicious “ such as its sacred, transcendental, revivalist, and
religious aspects “ might serve important transformative functions. Although
differences persist between gacaca as contemplated by the 2004 Organic Law
and the Specialized Chambers of the national courts (especially with regard to
community service as a sentence and somewhat shorter terms of imprisonment),
gacaca for genocide remains “tradition [ . . . ] cloaked in the mantle of a criminal
trial, with a strict and written procedure.”116
Traditional gacaca was not designed for mass atrocity (in fact it was geared for
property, inheritance, and family law matters, but it did exceptionally encompass
violent and serious crimes). However, just as it did not contemplate genocide,
it did not contemplate long-term isolated imprisonment either. And, with a
maximum sentence of thirty years™ incarceration without community service
for Category 2 offenders, there is something jarringly punitive about gacaca
as contemplated by the 2004 Organic Law. A need therefore arises for vigi-
lant monitoring insofar as the harsh sentences may prod the initiation of false
denunciations lodged for ulterior motivations of acquiring land or dispossessing
neighbors. The more gacaca trends toward punitiveness, the more due process
ought to attach to its processes “ else incarceration may be meted out to a person
with insuf¬cient guard against bias, error, or manipulation. On the other hand,
were gacaca for genocide to focus on traditional restoration and reintegrative
shaming, it would seem less terrifying even if it implicated the same vast num-
ber of people. If accountability were operationalized through remedies akin to
those of traditional gacaca, then its implication of the many acts of lower-level
offenders and bene¬ting bystanders that are necessary for atrocity to become
truly massive might not be so disquieting. International lawyers™ fears of gacaca™s
capaciousness might diminish to the extent that the system contemplated tradi-
tional sanctions.
Notwithstanding my many concerns regarding gacaca for genocide, it
remains an innovative approach to accountability and, in this vein, deserves
some deference, particularly in the initial stages of its operation. That said, I
believe gacaca had the potential to constitute a truly revolutionary approach to
accountability for mass violence, but as time passes it is not fully actualizing this
potential.117 It could have been a locus for the revitalization of indigenous, local,
and restorative mechanisms to stimulate a deeper accountability dynamic. How-
ever, attempts to diversify the accountability paradigm in Rwanda through pop-
ular measures such as gacaca, although partly successful, underachieve their
restorative, cathartic, and reconciliatory potential. In effect, gacaca for the
94 Atrocity, Punishment, and International Law

genocide is more akin to judicialized proceedings than the informal, ¬‚exible
extralegal methods of traditional gacaca from whence it was inspired. This gives
rise to two important questions. How did this come to be? And in this process of
becoming, which international lawyers may equate with progress, has something
been lost amid the gains?
I argue that a variety of pressures, some exogenous, have moved gacaca away
from its restorative and reconciliatory goals and structures to something that is
much more punitive and retributive. These include: (1) pressures brought by
the international community, in particular rights monitors and donors, to push
gacaca in the direction of criminal trials; (2) pressures by the Rwandan govern-
ment to centralize and bureaucratize gacaca, thereby removing local autonomy
and control, to suit its own ulterior motives; and (3) the reality that the gacaca
system was not initially designed to prosecute perpetrators of extreme evil and
the prospect of provisionally releasing, shaming, and rehabilitating murderers is
daunting. The ¬rst two of these reasons are of particular concern. They also are
interrelated insofar as one of the ways in which a process can look more lawlike
is if local discretion and particularities are sti¬‚ed through deep bureaucratiza-
tion by the state. In the end, pressures exerted upon gacaca have inhibited the
development of penological rationales that truly operationalize restoration and
reintegration as goals of sanction; insofar as some of these pressures could be
corrected over time, any assessment of the ultimate merit of gacaca for genocide
is an ongoing relational one.
Turning to the ¬rst factor: the gacaca proposal has been subject to criticism
by international lawyers, Western governments, and human rights activists, in
particular regarding its lack of conformity with dominant understandings of due
process.118 Criticism was ¬ercest when gacaca ¬rst was introduced. Insofar as
gacaca for genocide has responded to some of these criticisms, the outcry has
moderated, but certain speci¬c criticisms remain: the unavailability of defense
counsel; limited appeal rights; that the process of gathering evidence is com-
munal; and poor education and training (and often none in law), and potential
partiality, of judges. The international community consistently has urged gacaca
to resemble liberal legalist process and sanction, in which guilt instead of respon-
sibility is the goal.
In 2003, the Rwandan government adopted a new Constitution that “draws on
the main human rights treaties and institutions of Western democracies.”119 Cer-
tain constitutional provisions encompass due process protections.120 Article 190
of the 2003 Constitution provides that international treaties and agreements that
have been conclusively adopted are superior to organic and ordinary laws. The
government has thereby recommitted Rwanda to universalized human rights as
articulated in the major international human rights instruments (Rwanda had
been a party to a number of these instruments prior to the genocide). These
(re)commitments are motivated by a variety of goals, which include ideologi-
cal buy-in, standing in the international community, credibility, and the desire
to receive cases on referral from the ICTR (paradoxically, while committing
to these rights-bearing instruments the Rwandan government also has been
Punishment of International Crimes Nationally and Locally 95

exasperated by their application to ICTR defendants, especially when occasion-
ing an acquittal). The Rwandan government has been less successful at garner-
ing buy-in for these constitutional and human rights instruments at the local
level in Rwanda. This indicates a divide between state and society, central and
local authority, and core and periphery that is common to legal reform in many
developing nations. This divide can constitute an obstacle to the credibility of
national modalities for justice when operationalized within local communities,
particularly when communities remain ethnically divided. Justice as ordered
by the state and state elites may be externalized from justice as understood by
individuals whose lives primarily are lived locally.
Donor nations have been uninspired by, skeptical of, and in some cases
hostile to gacaca,121 especially without gacaca looking somewhat familiar to
donor nations and corresponding to their expectations of what justice norma-
tively should be. Rwanda, a very poor developing nation, cannot realistically
remain impassive in the face of these expectations. There is a striking imbalance
between, on the one hand, the response of the international community to the
ICTR (quite generous, where international funding supports an average price
tag of approximately U.S. $25 million to 30 million per verdict) and, on the other,
to justice mechanisms in Rwanda itself (much more modest). One de¬ciency of
these good faith criticisms by outsiders is that they construct gacaca primarily “
if not exclusively “ as a legal institution when, traditionally, and still to some
extent neotraditionally, it wears many hats. These include the legal, but also
extend to the political and the social. However, as gacaca becomes increasingly
adversarial, it focuses more on the fate of the accused rather than the fate of the
victim or the community.
Although, positivistically speaking, it is within the purview of the Rwandan
government to decide how it wishes to respond to genocide, and how it wishes to
react to international pressures, the Rwandan government is not without its own
set of political motivations. This moves the discussion to the second of the three
factors: pressures by Rwanda™s RPF government to centralize and bureaucratize
gacaca undermine local control while promoting the government™s political
agenda. Traditional gacaca fell outside the grasp of the formal state apparatus
and did not occur in state-created institutions. This no longer is the case under
the 2004 Organic Law. The centralization of gacaca has diminished popular
ownership over the process and has permitted the government access to gacaca
as a tool of social control. Some observers claim that gacaca courts are accus-
ing individuals based not on what they did during the genocide, but based on
their opposition to certain governmental policies.122 Rwanda remains an author-
itarian state under the auspices of the RPF. The government does not permit
much criticism. Rwanda™s President Paul Kagame received over 90 percent of
the popular vote in the last election, which was marred by allegations of vot-
ing irregularities and intimidation. Understandably, Rwanda faces a precarious
international and internal security situation, but RPF dominance is not a long-
term solution. Whereas decentralizing authority could diversify the loci of power
in Rwanda and, thereby, structurally mitigate the consolidation of in¬‚uence that
96 Atrocity, Punishment, and International Law

was one factor among many that accelerated genocide in 1994, the reality on the
ground is that gacaca has not actualized meaningful decentralization. There are
divides between the Tutsi elites governing the county and Tutsi survivors in the
In its traditional form, gacaca had considerable potential to serve goals of
communal empowerment, “to reincorporate the person who was the source of
the disorder”123 and “thereby restore the balance of the community.”124 It still
shares in this potential, which to my mind should be nurtured, but this has been
whittled down quite deliberately. Assuredly, gacaca for genocide does open “a
small, but real democratic space that creates the possibility for unforeseen, non-
hegemonic discussions. . . . ”125 This is one of its most valuable aspects. These dis-
cussions could involve issues of accountability for genocide, but also could spill
over into other areas unrelated to the genocide, thereby promoting political par-
ticipation generally. There is thus a discursive aspect to gacaca. Phil Clark, noting
that this discursive view (and the value of communal dialogue in and of itself)
accords with the expectations of much of the general population, elaborates:

According to the discursive view, participants in gacaca should feel free to
discuss issues which are crucial to their personal and communal experiences
during and after the genocide. Whatever “truth” may be discovered in gacaca
will be reached through communal dialogue, not through the views of elites
which they impart to the population. Such dialogue may be messy, may take
a long time and may in the end produce rather inconclusive results. [ . . . ] In
this view, gacaca encourages participants to discuss crucial issues in an open
environment where the community as a whole may bene¬t from hearing, and
contributing to, such dialogue.126

In practice, this discursive potential remains underexploited. For instance,
the content of the conversations are regulated by the government: off the table is
any discussion of human rights abuses by the government, or the reality that, in
ousting the genocidal regime, the RPA massacred thousands of Hutu civilians.
By eliminating jurisdiction over war crimes, the 2004 Organic Law cuts out
much of the alleged RPA and RPF criminality.127 Needless to say, discussion
of RPF activity also is off the table in the national courts and at the ICTR. In
fact, the Rwandan government lobbied against the reappointment of Carla Del
Ponte as ICTR Chief Prosecutor in part due to her insistence that allegations
of RPA crimes be investigated. The Security Council complied by deciding not
to renew her mandate.128 So it appears safe to say that the ICTR will not pursue
this line of investigation.
Moreover, there are reports that indicate that some members of the public
participate in the gacaca process out of a sense of coercion: they liken attendance
at gacaca events to duties they owe the government and express fears of being
branded as divisive should they not be seen as supporting the process.129 These
comments suggest the deep penetration of the formalized state apparatus into
gacaca; they also hearken back to chilling talk of “work” for the state (umuganda)
by which many Rwandans accounted for their participation in genocidal attacks
Punishment of International Crimes Nationally and Locally 97

in 1994. The formalism of the process has inhibited popular participation by
both the Hutu and Tutsi populations. Furthermore, there is evidence that the
RPF advances certain of its own ideologies at gacaca hearings.
Turning to the third factor: there is no denying that traditional gacaca was
not designed with mass atrocity in mind. Understandably, the prospect of pro-
visionally releasing extraordinary international criminals into the community,
after many years of incapacitation, with a goal of reintegrating them is daunting.
Unsurprisingly, there is concern among victims that gacaca too easily permits
reintegration and punishes too lightly. Certain victims believe that the use of
gacaca minimizes the seriousness of the underlying offense. From the victims™
perspectives, gacaca will have to strike a dif¬cult balance. It will have to main-
tain its distinctiveness without trivializing the wrongdoing that took place. That
said, surveys of the Rwandan population, although revealing mixed attitudes
among Hutu and Tutsi regarding gacaca (Hutu more favorably disposed than
the Tutsi), also demonstrate that both groups are more supportive of gacaca than
of ICTR trials, to which Rwandans as a whole remain relatively ambivalent and
uninformed, and of trials in the national courts.130
The fact that restorative initiatives, such as those envisioned by traditional
gacaca, have been downgraded in the justice matrix for mid- to lower-level
offenders is of concern to me.131 My concerns stem in large part from the limited
success of retributive criminal trials in Rwanda in propounding acknowledge-
ment of responsibility and atonement for genocide among detainees and defen-
dants. The judicialization of atrocity in Rwanda has not sparked these important
precursors to genuine reconciliation. In the Rwandan context, underdevelop-
ment of these catalytic precursors is troubling insofar as, owing to two salient
characteristics of the country™s social geography, collective reconciliation is vital.
These two characteristics are: (1) an ongoing need for victim and perpetrator to
live together; and (2) massive popular involvement in terms of perpetration and
Criminal trials in Rwanda have produced a limited sense of individual respon-
sibility or blameworthiness among detainees. I ¬rst noted this disconnect in
1998, when I interviewed hundreds of genocide suspects in the central prison of
Kigali.133 Nearly every interviewee did not believe he or she had done anything
“wrong,” or that anything really “wrong” had happened, in the summer of 1994.
Detainees who acknowledged that violence had occurred generally believed it
was necessary out of self-defense. These detainees did not perceive the massacres
as genocidal or in any way manifestly illegal. They saw themselves as honorable
citizens tasked to do the dirty work of furthering the interests of the state. Even
after years in jail, these detainees had not been disabused of the propaganda
fed to them by extremist Hutu leaders, according to which the Tutsi were out to
attack them, so, therefore, this attack had to be preempted by killing all the Tutsi.
This violence therefore became legitimized as a preemptive war of survival, not
condemned as genocide. Unsurprisingly, then, many detainees saw themselves
as prisoners of war, simply ending up on the losing side. As a general rule, the
trials, or the prospect of facing trial, failed to produce shame, contrition, regret,
98 Atrocity, Punishment, and International Law

or remorse among the prisoners. They instead produced emphatic denial, but-
tressed by the group solidarity that then pervaded Rwandan prisons. Needless
to say, this solidarity continues well past the 1998 period in which I conducted
my interviews: some detainees continue to refer to their fellow detainees as “a
Although the denials among detainees have thawed since the time I con-
ducted my interviews, other researchers whose work postdates mine note that
denials still persist.135 I believe that the large number of confessions and guilty
pleas “ involving tens of thousands of detainees “ that have occurred in recent
years demonstrates that the disavowal of responsibility, although still a thread
running through the Hutu detainee population, is dissipating.136 The prospect
of facing sentence by neotraditional gacaca panels prompted many of these
confessions.137 However, were gacaca for genocide to be more like traditional
gacaca, and less like formal criminal trial proceedings, I posit that a larger
number of individuals would come forth and confess “ and would do so more
sincerely “ and acknowledge the harms caused and their unacceptability, offer
apologies, and make amends. Accordingly, although neotraditional gacaca has
encouraged detainees and defendants to become more contrite over time, and
thereby has made a valuable contribution to reconciliation in Rwanda (much
more so than the criminal trials in the Specialized Chambers, although these
have advanced toward other goals), the extent to which gacaca has become judi-
cialized and subject to governmental control has dampened its reconciliatory
Resistance to fully operationalizing restorative and reconciliatory measures
in Rwanda appears misguided given the country™s social geography. Rwanda is a
dualist postgenocidal society, where in the aftermath of genocide both victim and
aggressor must live unavoidably side by side within the same nation-state, occupy
the same territory, and share common public spaces.138 In today™s Rwanda (as
has been the case throughout its history), Hutu and Tutsi live geographically
intermingled and in close economic interdependence. There is no separate
Hutuland or Tutsiland, nor any possibility for such separation. Hutu and Tutsi
speak the same language. Religious af¬liations are not ethnically driven.
This commingling between Hutu and Tutsi operates in tandem with the high
degree of public participation and complicity in the genocide, together with the
pronounced level of victimization. Violence often was committed by neighbors
upon other neighbors within local communities. Killings were committed pub-
licly and were known to all. No attempt was made to conceal them. They were
not sanitized through technology nor sterilized through anonymity. The killing
was grueling, dirty, labor-intensive work “ it takes many blows to kill someone
with a machete, hoe, or stick. Many Rwandans provided lists of Tutsi in their
region to the killers. Teachers identi¬ed students, physicians identi¬ed patients,
and pastors identi¬ed the faithful. Signi¬cant numbers of Rwandans acquiesced
in the face of genocide. Many of these individuals stood silent as murder plagued
their streets, only to promptly move into a suddenly vacant home.
These characteristics, in turn, suggest that when considering objectives of
punishment in the Rwandan context, reconciliation and reintegration ought to
Punishment of International Crimes Nationally and Locally 99

be given high priority. This does not mean that individuals should be spared
accountability “ quite the contrary, I believe the net should be cast broadly “
but that the processes of accountability should encourage acknowledgement
of responsibility and reconstruction of social norms. Until reconciliation and
reintegration are deeply operationalized, Rwanda likely shall remain an eth-
nocracy with the fears of the minority Tutsi consolidating, instead of relaxing,
their grip on the levers of power. When postgenocidal accountability measures
consolidate instead of pluralize authority, and serve as conduits for state power
at the expense of local empowerment, they remain inherently limited in the
kind of transformation they can effect in removing preconditions to future vio-
lence. These concerns do not vitiate gacaca for genocide™s innovative relevance
in the accountability process, nor strip it of its entitlement to quali¬ed deference
on the part of the international community, but, rather, suggest that, as gacaca
begins in earnest throughout Rwanda, its architects re¬‚ect upon how instanti-
ating some of its informal and communal aspects could boost its restorative and
reconciliatory potential.

(ii) former yugoslavia
The collapse of the Socialist Federative Republic of Yugoslavia (SFRY) triggered
the dissolution of the union of its constituent entities. Croatia and Slovenia were
the ¬rst to proclaim independence in 1991, followed by Bosnia and Herzegov-
ina (in which there was a sizeable population of Bosnian Muslims, but also
Bosnian Serbs and Bosnian Croats). This left the remaining provinces as the
Serb-dominated Federal Republic of Yugoslavia (FRY). Fighting then began
among Serbs, Croats, and Bosnian Muslims, in particular among militia forces,
but civilians were deliberately targeted in pervasive violations of international
humanitarian law. This violence raged until 1995, at which point a peace set-
tlement was brokered. In 1998 and 1999, the FRY commenced a campaign
of systemic human rights abuses against ethnic Albanians in Kosovo. In total,
these con¬‚icts among Serbs, Croats, Bosnian Muslims, and Kosovo Albanians
claimed the lives of 200,000 to 250,000 individuals. These con¬‚icts also saw
the worst atrocities in Europe since World War II. Particularly egregious were
brutalities committed in detention camps run by Serbs pursuant to policies of
ethnic cleansing and, as has been found by the ICTY, genocide. Many of these
brutalities were committed in Bosnia and Herzegovina.
Judicialization of atrocity in the former Yugoslavia proceeds through a number
of institutions: the ICTY, a hybrid court (Kosovo), and national courts in several
states within and outside of the states that emerged from the former Yugoslavia.

a. Positive Law Frameworks
The legislative framework for the punishment of extraordinary international
criminals in the national courts of the states of the former Yugoslavia is extremely
complex. This is so because several legal instruments might apply to the crimes.
The SFRY Criminal Code is one such instrument. Each of the states that
100 Atrocity, Punishment, and International Law

emerged out of the SFRY, however, has adopted its own domestic criminal
code. And, what is more, many of these codes have over time been subject
to amendment and, in some cases, signi¬cant reform. The legal framework in
effect today, when a convict may be punished, differs from the framework in
effect at the time the offense was committed, which itself was a period of rapid
dissolution and reconstitution in terms of operative legal structures. Although
principles of retroactivity suggest that the law in force at the time of committing
the offense should govern (which is the approach the ICTY has taken), prin-
ciples of lenity intimate that, if the current punishment is more lenient than
the former punishment, then perhaps current punishment schemes ought to
apply.139 The ICTY is especially leery of being in any way bound by changes
in domestic law that make punishment more lenient. Its fears are that “[i]n
passing a national law setting low maximum penalties [ . . . ] States could then
prevent their citizens from being properly sentenced by [the ICTY]. This is not
compatible with the [ICTY™s] primacy [ . . . ] and its overall mandate.”140
The SFRY Criminal Code came into force in 1977.141 At the time, regional
criminal codes also were enacted in the SFRY™s constitutive republics and
autonomous regions. These legal frameworks coexisted in the federated system.
When the republics and regions split from the SFRY, they originally retained
both the SFRY Criminal Code and their regional codes, but in the following
years enacted new codes that became the governing law. For example, in 2003,
Bosnia and Herzegovina enacted a new criminal code on the state level; its
main constitutive entities (the Serb-led Republika Srpska and the Federation of
Bosnia and Herzegovina [Bosniak/Croat led]) also enacted their own criminal
The 1977 SFRY Criminal Code permitted the imposition of capital punish-
ment, imprisonment, con¬scation of property, and ¬nes.142 The death penalty
could be imposed only for the most serious criminal acts; the general range
of imprisonment was from ¬fteen days to ¬fteen years. However, if a criminal
offense was eligible for the death penalty, was perpetrated under particularly
aggravating circumstances, or caused especially grave consequences, then a
sentence of twenty years could be given.143 In other words, the death penalty
could be transformed by the court into a twenty-year prison sentence. In terms
of speci¬c extraordinary international crimes, article 141 of the SFRY Criminal
Code punished genocide with imprisonment for not less than ¬ve years or by
the death penalty; war crimes were subject to a similar scale.144 Although in the
former Yugoslavia the death penalty could attach to extraordinary international
crimes, over time this penalty became abolished, thereby leaving the alternative
punishment of imprisonment for a term of twenty years for criminal acts eligible
for the death penalty.145 The purposes of punishment under the SFRY Criminal
Code were to prevent the offender from committing criminal acts, rehabili-
tation, to in¬‚uence others not to commit criminal acts, and “to strengthen the
moral ¬bre of the socialist society and to in¬‚uence the development of the social
responsibility and discipline of the citizenry.”146 The SFRY Criminal Code also
speci¬ed aggravating and mitigating factors.
Punishment of International Crimes Nationally and Locally 101

When the SFRY broke apart in 1992, the FRY maintained that it was the
successor to the SFRY. This view was not shared by most in the international
community, who instead maintained that the SFRY had dissolved and that the
FRY was a newly emergent state. This question of status, however, was a matter
relevant to the external relations of the FRY and not its internal legal structure.
The 1992 FRY Constitution abolished the death penalty, although this did not
affect the death penalty for offenses regulated by regional criminal codes in
Serbia and Montenegro (although in the former case, the Constitutional Court
of Serbia had declared the death penalty unconstitutional and, in 2002, the Ser-
bian Parliament formally abolished it and replaced it with a term of forty years™
imprisonment). For the most part, the FRY kept the SFRY Criminal Code in
force, which it simply renamed the FRY Criminal Code, and which remained
applicable to extraordinary international crimes. In 2002, an amendment to the
FRY Criminal Code replaced the death penalty for offenses regulated in the FRY
Criminal Code with a punishment of long-term imprisonment for forty years,
although as of 1992 the imposition of the death penalty had already become
impermissible for FRY Criminal Code crimes owing to the constitutional abo-
lition thereof.147
In 2003, the FRY was transformed into the Confederation of Serbia and Mon-
tenegro. A further complicating development occurred in 2006, when Montene-
gro proclaimed independence following a plebiscite in which 55.4 percent of
Montenegrans voted to end the confederation with Serbia.148 This proclamation
should not repercute strongly on the administration of criminal justice, insofar
as the Confederation of Serbia and Montenegro did not have a federal crimi-
nal code: the competence to legislate in criminal matters operated at the level
of Serbia and Montenegro. Serbia essentially reproduced the FRY Criminal
Code as the Basic Criminal Code of Serbia, whereas Montenegro adopted a
comprehensive new criminal code that entered into force in 2004. In Serbia,
grave crimes are punishable by a ¬xed term of forty years™ imprisonment, with
possibility for parole after service of half the sentence (in exceptional cases, one-
third).149 War crimes and genocide are punishable in Serbia by imprisonment
for not less than ¬ve years or by long-term imprisonment of forty years, thereby
revealing vast judicial discretion.150
The Bosnia and Herzegovina Criminal Code contemplates long-term pun-
ishment in a range from twenty to forty-¬ve years for the gravest forms of crimi-
nal offenses. The sentencing factors applicable to all crimes, even extraordinary
international crimes, include: degree of criminal liability, motives for perpetrat-
ing the offense, degree of danger to the protected object, circumstances in which
the offense was perpetrated, past conduct of the perpetrator, personal situation of
the perpetrator, and conduct after the offense.151 The goals of sentencing all types
of offenses include the expression of the community™s condemnation, reform of
the perpetrator, deterrence, and raising the public™s awareness of the danger of
crime and the fairness of punishment.152 Early release is contemplated, gener-
ally after service of one-half of the sentence, but a person punished by long-term
imprisonment may be granted conditional release only after three-¬fths of the
102 Atrocity, Punishment, and International Law

sentence has been served.153 Speci¬cally enumerated ranges for certain extraor-
dinary international crimes in the Bosnia and Herzegovina Criminal Code are
a term of not less than ten years or long-term imprisonment, thereby investing
considerable discretion in the judge,154 and apparently setting a range of pun-
ishment that begins ten years lower than for long-term punishment of serious
ordinary crimes. As for outer limits, as the ICTY has recognized in recent referral
decisions, the maximum sentence for serious international crimes in Bosnia and
Herzegovina is forty-¬ve years.155 A similar structure operates in the subnational
entities within Bosnia and Herzegovina for long-term punishment, although
extraordinary international crimes are not speci¬cally addressed in these subna-
tional codes. Bosnia and Herzegovina has the highest possible prison sentence
currently operative in the states of the former Yugoslavia for extraordinary inter-
national crimes.
The Croatian Criminal Code entered into force in 1998, but has since been
amended and revised on a number of occasions. In 2003 and 2004, Croatia
adopted legislation to implement the Rome Statute of the International Crimi-
nal Court, resulting in the integration of new criminal offenses and procedures
to national law. This law does not apply to the violence in the Balkans Wars of
1991“1995. The Croatian Criminal Code contemplates long-term imprisonment
of between twenty and forty years for the most serious offenses, which is similar to
the scheme in Bosnia and Herzegovina, but with an upper edge that is ¬ve years
less. Although the 1998 Croatian Criminal Code does not speci¬cally reference
crimes against humanity, and hence differs from the Bosnia and Herzegovina
Criminal Code, it turns to the similar structure of not less than ten years or
long-term imprisonment for certain war crimes (although for most war crimes
the minimum sentence is not less than ¬ve years) and genocide.156 The lower
range of the sentencing threshold for these extraordinary international crimes is
ten years below (and in the case of many war crimes, ¬fteen years below) that
for the most serious ordinary crimes.
Criminal code legislation enacted in Kosovo in 2004 in the hope of regu-
larizing and standardizing the prosecution and punishment of ordinary crime
established a sentence of long-term imprisonment of twenty-one to forty years
for particularly serious offenses committed under aggravating circumstances.157
Alternative measures such as suspended sentences, ¬nes, and community ser-
vice work also are contemplated. The Kosovo criminal justice system has pro-
cessed, in ordinary courts, crimes of ethnically motivated violence. Initially, these
prosecutions were deeply marred by ethnic bias.158 This prompted the United
Nations Mission in Kosovo to create (through Regulation 2000/64) internation-
alized hybrid panels to adjudge extraordinary international crimes, whose work
I brie¬‚y mentioned in Chapter 3. The ICTY retains primary jurisdiction over
serious international crimes committed in Kosovo but may begin to transfer
cases to hybrid courts in Kosovo. Ethically motivated violence adjudged in the
ordinary courts is not treated as extraordinary international crime. The ordinary
courts in Kosovo continue to be faulted and the system is in disarray. Partic-
ular concerns have been noted regarding sentencing. These include lack of
Punishment of International Crimes Nationally and Locally 103

appropriate reasoning or substantiation, inconsistency, excessive use of custo-
dial measures, and lack of institutional capacity.159 Some local judges in Kosovo
reference the customary law of the Code of Lek¨ Dukagjini (otherwise known
as Kanun), ¬rst codi¬ed in the ¬fteenth century. The Kanun contains “detailed
rules for governing daily life and prescribes rights, obligations, duties, levies and
punishment.”160 It makes mention of retribution (lex talonis) and also recon-
ciliation. The Kanun is of some in¬‚uence in the determination of sentence in
Kosovo, especially by Kosovo Albanian judges.161 It is not encouraged by the
new Kosovo criminal code.
The ICTY has exerted considerable in¬‚uence on the legal systems of the
states of the former Yugoslavia. Much of the recent law reform, for example
reenactments of criminal codes (in particular procedural elements) in Bosnia
and Herzegovina and Kosovo, radically moved these systems to an adversar-
ial model from what had hitherto been an investigatory/inquisitorial model.162
The 2003 reforms in Bosnia and Herzegovina constituted a “shift to a broadly
adversarial criminal justice system where “ in contrast to the previous judge-led
mixed system “ the trial is moved forward by the prosecutor and the defendant,
and the judge represents the neutral arbiter of the disputed issue.”163 A major
impetus in this process of law reform throughout the former Yugoslavia is the
reality that, by aligning domestic structures to those of the ICTY, these domes-
tic structures become better positioned to receive cases from the ICTY, along
with international support, expertise, and resources. Furthermore, coopera-
tion with the ICTY has become a central criterion on which Serbia™s relationship
with the European Union has become contingent.
These developments attest to the in¬‚uence of newly crystallized interna-
tional processes to prosecute and punish perpetrators of mass atrocity, as well
as the migration of these operational norms back to the national level in places
that, heretofore, had not adhered to such methodologies. In the case of the for-
mer Yugoslavia, there is some evidence that these transplants are improving the
quality of justice by dissipating ethnic bias and promoting transparency in the
administration of justice. In terms of sentencing, many of the domestic reforms
mirror the ICTY™s grant of considerable discretion to judges, although are not as
permissive. Insofar as national judges have few guidelines to fetter their discre-
tion, a review of the case law, to which this discussion now turns, demonstrates
considerable variance in terms of sentence issued.
To be sure, some factors routinely increase or decrease the length of sen-
tence in the national courts of states emergent from the SFRY. Guilty pleas
appear to be one factor consistently considered in mitigation. In the case
of Bosnia and Herzegovina, observers have noted that many plea-bargained
sentences dip below the proscribed minimum sentence, which is permissible
only in highly extenuating circumstances (article 49, Bosnia and Herzegovina
Criminal Code). This phenomenon particularly arises in cases where long-term
sentences are issued. There also is evidence of divergent and inconsistent judi-
cial practice regarding sentencing individuals who plead guilty that, in turn,
threatens the principle that like cases are to be treated alike.164 In terms of
104 Atrocity, Punishment, and International Law

aggravating factors, consistent reference is made to the heinous nature of the
Ulrich Sieber and a team of experts interviewed judges in the former
Yugoslavia.165 In these interviews, commissioned in 2003, the judges stated that
the fact that an offense was committed in times of war, as opposed to “normal cir-
cumstances,” leads them to sentence more severely. As such, the extraordinary
nature of the crime would constitute an aggravating factor (contrary to other
situations, for example, in the East Timor Special Panels and in some of the
domestic Rwandan cases, where the chaos of war has been seen as a mitigating
factor).166 At ¬rst blush, these interviews suggest that extraordinary international
crimes are viewed as more serious than ordinary common crimes even when
they embody a similar actus reus (e.g., rape, murder, or torture when committed
as ordinary offenses or as war crimes). To this end, a penology for extraordinary
international crimes might be emerging that grounds the differential sanction of
these crimes in retributive or expressive goals tethered to the inherently greater
gravity of offenses when committed in group con¬‚ict situations.
However, there are a number of wrinkles and limitations to the extrapolations
that can be inferred from the interviews of judges in the former Yugoslavia
published in Sieber™s report. First, the interviews suggest that the differences
between the punishment of wartime offenses and ordinary offenses are most
distinct in cases of the commission of a single offense and drop sharply in cases of
combined offenses of ¬ve to ten victims (where at the federal level in Bosnia and
Herzegovina and in Croatia no differences were reported and, in fact, there was
some evidence of more severe sanction for ordinary crimes).167 Given the nature
of extraordinary international crimes, a single offense is more an aberration than
the norm. Second, the positive law frameworks in Croatia set a lower minimum
punishment for war crimes than for serious ordinary crimes. So, too, do those
in Bosnia and Herzegovina. Third, recourse by judges in the former Yugoslavia
to the “special circumstances of war” as an aggravating factor is not in any way
predictable. In fact, the researchers conducting the judges™ interviews found
that “judges had trouble explicitly considering concrete factors” in sentencing
and that this, in turn, prodded a “retreat to the general sentencing criteria”
and a replication of those mitigating factors explicitly stipulated in the general
criminal legislation drafted with ordinary common crimes in mind.168 Fourth,
and most pertinent, these data emerged from model cases presented to a small
sample of judges for them to determine sentences hypothetically based on their
experience; this is quite a different exercise than sentencing actual perpetrators
following actual convictions. In fact, there is a difference between what judges
may in interviews say they are doing and what they actually do.
In the case law, the differentiation between crimes committed as extraordinary
international crimes and ordinary common crimes is more inconclusive. Extant
Croatian case law suggests a tendency to prosecute (and judge) ordinary crimes
as extraordinary international crimes and then award them the lowest possible
sentence. Ethnic bias and politicization corrode the retributive, expressive, and
deterrent value of punishment in Serbia. There is frequent quashing of lower
Punishment of International Crimes Nationally and Locally 105

court decisions and remand by appellate courts in all jurisdictions. Although
there are indications that, in very recent years, national courts are beginning
to sentence perpetrators of international crimes to longer prison terms and are
shedding the distorting effects of ethnic bias, there is no predictably conclusive
movement in this direction.

b. Courts in Bosnia and Herzegovina
The Organization for Security and Cooperation in Europe (OSCE) has reported
on the prosecution of extraordinary international crimes within the domestic
(cantonal and district) courts of Bosnia and Herzegovina.169 The OSCE Report
focuses on proceedings before the ordinary courts, as the War Crimes Chamber
of the State Court of Bosnia and Herzegovina had not yet come into existence
at the time of its preparation.
The War Crimes Chamber, which focuses on serious extraordinary interna-
tional crimes, was created in January 2005 primarily in response to the referral
procedure initiated by the ICTY.170 The Chamber formally opened in Sarajevo
in March 2005.171 Proceedings have commenced. The ICTY has referred cases.
At the time of writing, the War Crimes Chamber had issued its ¬rst sentence,

13 1/3 years™ imprisonment, to Nedo Samardˇ i´ , a Bosnian Serb convicted of
crimes against humanity. In May 2006, it initiated its ¬rst genocide trial; these
proceedings involve eleven Bosnian Serbs associated with the Srebrenica mas-
sacre. Looking ahead, the War Crimes Chamber shall be better equipped in
terms of expertise to deal with the prosecution of extraordinary crimes than the
ordinary domestic courts. That said, it remains unclear whether the Chamber™s
judges shall develop independent rationales of punishment insofar as pressures
to conform to ICTY expectations in order to keep receiving referred cases will
likely diminish the development of any sui generis approach, unless the ICTY
itself moves in this direction.
Because the War Crimes Chamber is tasked only with the more serious cases,
many charges of extraordinary international crimes will remain within the ordi-
nary courts of Bosnia and Herzegovina.172 The OSCE report determined that,
in 2004, Bosnia and Herzegovina cantonal courts acquitted ¬fteen defendants,
found nine guilty and, for these nine, issued sentences ranging from eighteen
months to ¬fteen years of imprisonment.173 The twenty-four defendants impli-
cated in these proceedings divide into seven Bosniaks, ten Croats, and seven
Serbs.174 In 2005, the OSCE Report noted two additional convictions with sen-
tences of 7 and 4.5 years.175
The OSCE Report also noted a number of important trends. First, it found
that some courts and prosecutors had “made conscientious efforts to bring those
responsible for war crimes to justice.”176 But numerous shortcomings were noted,
principally ethnic bias, fear among judges and prosecutors for their safety, dif-
¬culties in terms of securing witnesses, and lack of structures for transborder
cooperation.177 Lack of coordination among courts and prosecutors also ham-
pers efforts to obtain custody over and prosecute suspects. Observers also have
106 Atrocity, Punishment, and International Law

voiced concern with regard to due process, although these criticisms have qui-
eted over time. In terms of sentencing rationales, the OSCE Report is helpful
in delineating the operationalization of penality within these domestic courts,
and con¬rms an apparent lack of independent or cogent rationales for sentenc-
ing extraordinary international criminals (or, in many cases, an absence of any
stated reasons for aggravation or mitigation). The mean sentence of the cases
documented by the OSCE is slightly under nine years™ imprisonment.178
On a different note, in July 2006 a reparations system for rape victims from the
Bosnian Wars was for the ¬rst time being considered by legislators in Sarajevo.
Also, a private Bosnian nongovernmental association is organizing a lawsuit
against the Republic of Serbia, in which it seeks reparations for women who
were raped or abused or who had family members killed during the con¬‚ict.
In both cases, these initiatives demonstrate attempts to pluralize modalities of

c. Courts in Serbia
Ernesto Kiza reports that national courts in Serbia are not approaching the pun-
ishment of extraordinary international criminals in a predictable or structured
manner.179 For the most part, the sentences issued remain quite lenient when
compared to ICTY sentences. This arises, in Kiza™s estimation, for two reasons.
First, although there is evidence of the in¬‚uence of the ICTY model of justice
in terms of affecting the perceived need to judicialize mass atrocity, and the for-
mula of what that judicialization should resemble, what is lacking is reference to
ICTY sentencing jurisprudence (which is itself already unsystematized in terms
of the relationship between gravity of crime and severity of punishment).180
This lack of systematization is compounded by the inability of Serbian judges to
develop a comfort with or a methodology to punish extraordinary international
criminals. Kiza™s empirical research leaves him with the sense that “domestic
judges [in Serbia] were simply overstrained by the prospect of punishing offend-
ers of the gravest crimes committed during a state of widespread anomie. They
were just not sure how to handle the cases, although most of them were very
experienced concerning ˜regular™ murder, rape, and other violent crimes.”181
What is more, there is considerable evidence of ethnic bias in the Serbian
judgments, pervasive clientelism, discomfort with analyzing the responsibility
of the political and military leadership, and a lack of support from the Serbian
International and foreign pressures upon the Serbian justice system, includ-
ing demands explicitly raised as part of extradition negotiations, have dissipated
certain of the crudest manifestations of bias.183 This result is desirable. These
pressures also have led to the establishment of a Special Court for War Crimes
in the Belgrade District Court. A War Crimes Prosecutor has been appointed.
Arrests have been undertaken, including for atrocity in Kosovo. No referrals
have yet been made from the ICTY to Serbia, unlike the case with Croatia, and
Bosnia and Herzegovina.
Punishment of International Crimes Nationally and Locally 107

High-pro¬le trials have begun at various levels in the Serb judicial system. In
December 2005, a Serbian court convicted fourteen former Serb militia ¬ghters
for the killing of nearly 200 Croat prisoners in Vukovar; it issued sentences
ranging from two to twenty years.184 Trials have been undertaken in Serbia
with regard to the Srebrenica massacre.185 Proceedings have been initiated with
regard to massacre in Kosovo.
That said, one thing these trials and convictions have not accomplished is
to generate widespread acknowledgement within the Serbian population that
Serb forces committed atrocities throughout the former Yugoslavia. A contrario:
“[M]any Serbs say they are either unaware of war crimes or refuse to accept that
their police or security forces could have committed them.”186 In the words of
the Humanitarian Law Center, “Serbia is still stumbling under the burden of
war crimes committed in the name of alleged patriotism.”187

d. Courts in Croatia
Developments related to domestic prosecutions for extraordinary international
crimes have been in¬‚uenced by Croatia™s relations with the ICTY, in particular
the prospect of referral of cases to Croatia as part of the ICTY™s completion
strategy.188 This prospect has encouraged Croatia to create a Special Court for
War Crimes. At the time of writing, the ICTY has referred one case to Croatia.
However, ordinary courts in Croatia have for some time prosecuted atrocity.
Since 1991, Croatian courts have entered guilty verdicts against approximately
800 persons (many convicted in absentia) for war crimes (against civilians or pris-
oners of war) and genocide; at the time of writing, proceedings remain pending
against another 1,400 to 1,500 individuals, with other investigations outstand-
ing.189 The OSCE has issued a number of detailed reports regarding domestic
trials for extraordinary international crimes in Croatian courts.190 These reports
shed some light on sentencing practices and rationales.
The OSCE Report published in 2005 noted pervasive, albeit observably dissi-
pating, ethnic bias resulting in Serb defendants™ being disproportionately subject
to investigation and prosecution for extraordinary international crimes. In 2004,
Croatian prosecutors eliminated large numbers of unsubstantiated proceedings
against Serbs.191 This is an important step, in the OSCE™s view, toward reme-
dying a situation in which thousands of cases had been initiated against Serbs
and only tens of cases against Croats, an extreme disproportion that “cannot be
attributed only to different levels of criminality of certain members of the war-
ring parties.”192 Whereas Serbs have been prosecuted for war crimes based on
allegations of physical or psychological abuse, Croats have been almost exclu-
sively prosecuted for conduct that involved killings; what is more, the Croatian
judiciary “appears to apply a broader de¬nition of genocide for which only
Serbs have been convicted [ . . . ].”193 Even when prosecutions have been initi-
ated against Croats for killings and torture, ethnic bias has pervaded the initial
judgments “ requiring in some cases corrective action by the Croatian Supreme
Court. On September 13, 2005, a domestic retrial of Croat military policemen
108 Atrocity, Punishment, and International Law

accused of torturing and killing Serb prisoners of war in 1992 began following
the overturning of acquittals by the Croatian Supreme Court in 2004. However,
of the original eight defendants, only four reappeared in court, with the other
four having ¬‚ed into hiding following acquittal at the initial trial in 2002. The
Croatian Supreme Court plays an important corrective function by reversing
55 percent of the trial court verdicts and ordering retrials and, in some of the
af¬rmed verdicts, adjusting the sentence.194 The ¬gure of 55 percent, which is
from 2004, is down from a reversal rate of 95 percent in 2002 and 60 percent in
The OSCE Report also notes that, because so many lower-level Serbs are pros-
ecuted, a large number of convictions become issued for “less serious offenses”
that, when aggregated, result in the widespread imposition of less onerous pun-
ishment. For example, nearly 60 percent of the Serbs convicted of war crimes in
2004 received a sentence less than the statutory minimum of ¬ve years (the sen-
tencing range for war crimes [against civilians or prisoners of war] and genocide
is ¬ve to twenty years).196 This creates an arti¬cial downward pressure on the
quantum of sentence in Croatian national courts for war crimes and genocide,
which places this quantum out of proportion to that of the ICTY. The less seri-
ous charges pursued domestically (apparently done deliberately so as to convict
large numbers of Serbs) account for the discrepancy.
Among the large number of cases (76 cases, covering 211 individuals) moni-
tored by the OSCE Croatia Mission in 2004, 24 trials involving 47 individuals
(42 Serbs, 4 Croats, and 1 Hungarian) were concluded that year.197 Thirty indi-
viduals were found guilty, twelve were acquitted, and charges were abandoned
against ¬ve.198 More than half of those convicted received sentences less than
the stipulated minimum of ¬ve years (it is permissible under Croatian law to
dip below the minimum only if particularly obvious mitigating circumstances
exist).199 Overall, sentences ranged from a low of 1.5 years to a high of 15 years
with an average sentence of approximately 5.5 years200 (this is a decrease from
an average sentence in 2003 of 9 years). Fifteen individuals received sentences
in the one- to four-year range; eleven individuals in the ¬ve- to nine-year range;
three in the ten- to fourteen-year range; and one in the ¬fteen- to twenty-year
range.201 In 2003, three monitored defendants received the maximum punish-
ment of twenty years.202 More speci¬cally, in 2003, two individuals were sen-
tenced to terms in the one- to four-year range; twenty-three to terms in the ¬ve-
to nine-year range; ¬ve to terms in the ten- to fourteen-year range; and seven
to terms in the ¬fteen- to twenty-year range.203 Of these convictions, twenty-
six were for war crimes against civilians, three for war crimes against prisoners
of war, and eight for genocide.204 As an aside, the ICTY has not held that
genocide took place in Croatia. In 2002, fourteen individuals were sentenced
in the one- to four-year range, eleven to the ¬ve- to nine-year range, fourteen
to the ten- to fourteen-year range, and thirteen to the ¬fteen- to twenty-year
Mitigating circumstances are frequently invoked to justify the lowering of sen-
tences below the statutory minima.206 These mitigating circumstances include
the following, some of which are contradictory and none of which seem to be
Punishment of International Crimes Nationally and Locally 109

of the requisite importance to justify dipping below the statutory minima: the
defendant is married; the defendant is divorced; the defendant has children;
the defendant does not have a criminal record; the defendant is poor and does
not own property; defendant™s health and physical constitution, employment
status, social status, susceptibility to coercion; and following orders or convey-
ing orders.207 Paradoxically, although conveying orders has been found by some
Croatian courts to constitute a mitigating factor, others refuse to consider it as
such.208 Whereas service in the Croatian armed forces is a mitigating factor, ser-
vice in the Yugoslav armed forces is an aggravating factor.209 Whereas Croatian
attacks are seen as defensive, Serb attacks are seen as offensive, the latter being
graver for purposes of sentencing. Other aggravating factors include persistence
in committing the act and groundless maltreatment,210 and motivation to create
a greater Serbia.211
The OSCE Report also concludes that the punishment imposed for compa-
rable conduct differed drastically, thereby leading to inconsistent sentencing.212
It cites among examples the following: “In the ˜Virovitica™ case, the Bjelovar
County Court sentenced three Croats to one year each for having beaten two
civilians, one of whom subsequently died. In contrast, the Osijek County Court
sentenced Branko Stankovi´ , a Serb, to 6 years™ imprisonment for arresting and
beating a civilian until he fainted.”213 The ¬rst of these decisions, however, was
reversed by the Supreme Court and, as of 2005, a retrial is pending. There
has been a trend toward improving the predictability and integrity of justice
in Croatia, as in other jurisdictions in the former Yugoslavia. The increasing
regularization of the activity of the Croatian courts through appellate review
may lead to more principled systematicity in sentencing. However, the Croatian
courts have not yet reached this point.214 The main source of predictability in
sentencing remains ethnic bias (i.e., violence being less grave when committed
by Croats than by Serbs). Moreover, there is no indication of the development
of broader-based remedies or theories of punishment speci¬cally attuned to the
atrocity perpetrator.

e. Foreign Courts
Trials have been conducted in national courts outside the former Yugoslavia.
In some of these cases, principles of universal jurisdiction have been invoked.
German courts have adjudged a number of defendants: Djaji´ (1997), Jorgi´
c c
215 216
(1997), Sokolovi´ (1999), and Kuslji´ (1999). Jorgi´ and Kuslji´ received
c c c c
life sentences for genocide. Djaji´ , convicted of war crimes for fourteen cases
of aiding and abetting murder and one case of attempted murder, received ¬ve
years.217 Sokolovi´ , convicted of aiding and abetting genocide and war crimes
and of committing murder as a war crime, received nine years.218 There is little
discussion of sentencing considerations, although the Jorgi´ court found no
elements of justi¬cation or exclusion of responsibility that would ordinarily serve
to reduce a life sentence under German law.219 Quite the contrary: the Jorgi´ c
court underscored the gravity of the crime, thereby implying the importance
of retribution.
110 Atrocity, Punishment, and International Law

In 1994, a Danish jury sentenced Re¬k Sari´ , a Bosnian Muslim, to eight
years™ imprisonment for fourteen counts of serious bodily harm as war crimes.221
Sari´ had sought asylum in Denmark. The jury found aggravating circumstances
under the applicable Danish law, but the judgment provides no elaboration.222
The jurisdiction of a Dutch court has been invoked in a compensation case
against the Netherlands brought by relatives of victims of the 1995 Srebrenica
massacre. Dutch soldiers had stood by while Bosnian Serb forces massacred at
least seven thousand Bosnian Muslim men and boys in what had been declared
to be a UN safe area. This kind of civil litigation can help spread responsibility
for atrocity more broadly. Ironically, the Netherlands is the seat of the ICTY,
where judgments regarding individual criminal responsibility for genocide in
Srebrenica have been rendered and continue to be heard.

(iii) world war ii
National military and criminal justice institutions actively prosecuted and pun-
ished perpetrators of Nazi aggression, the Holocaust, and systemic criminality
in the Paci¬c Rim. Thousands of trials took place far away from Nuremberg
and Tokyo in national courts, military commissions, and military courts all over
Europe and the Far East. Although the accused were not as high-pro¬le as those
who appeared before the IMT, many came from senior ranks of Axis armed
forces or were noted for particularly gruesome conduct during the con¬‚ict.
Allied military commissions zealously undertook prosecutions of Japanese
war crimes: over 5,500 individuals were charged, 900 received death sentences,
and 3,500 received prison sentences.223 With regard to the war in Europe, many
prosecutions took place in the British, French, American, and Soviet zones of
occupied Germany and Austria. It is estimated that the Soviets alone tried over
ten thousand cases. Trials of other Nazis occurred in the courts of those countries
where they had committed their crimes, or elsewhere, and included Belgium,
France, Yugoslavia, Italy, Poland, Norway, the Soviet Union, and Czechoslo-
vakia. Although most of the defendants were nationals of enemy countries, in
particular Germany, many courts tried their own nationals as well. For example,
French national courts tried about 100,000 collaboration cases: 65,000 individu-
als were found guilty, although an amnesty law was passed in 1953.224 The most
famous collaborator trials involved Mar´ chal Henri P´ tain and Pierre Laval
e e
(respectively, the Head of State and Prime Minister of France™s wartime Vichy
regime); and also Norway™s Vidkun Quisling. In Italy, attempts to judicialize
atrocity were weak. They were in fact largely superseded by “private revenge,”
which reasonable estimates suggest led to the disappearance or summary exe-
cution of 30,000 Italian fascists.225 Even more so than was the case in Italy, in
France accounts were settled extrajudicially. Carlos Santiago Nino reports that
“[i]n 1944 alone, private citizens killed approximately 40,000 people accused of
collaborating with the Nazis.”226
The judicialization of the Holocaust and extraordinary international crimes
committed by the Nazis has been considerably more pronounced than the
Punishment of International Crimes Nationally and Locally 111

judicialization of extraordinary international crimes committed by the Japanese.
Proceedings involving Japanese defendants tapered off by the end of the 1940s,
whereas those against Nazis and collaborators continued for many decades, albeit
not steadily. Overall, trials for World War II atrocities have proceeded in waves:
(1) military and civilian proceedings in the immediate aftermath of the war in
both Europe and the Far East; (2) civilian proceedings that resurged in a variety
of jurisdictions in the 1960s with regard to Nazi atrocity; and (3) a handful of high-
pro¬le cases, again with regard to Nazi atrocity, in civilian courts in the 1980s
and 1990s. Courts continue to investigate, convict, and sentence perpetrators
over sixty years after the Holocaust, although the number of defendants now has
dwindled to a tiny handful of feeble and frail old men. Coincident with this third
wave of criminal prosecution is acceptance by Germany and Austria, along with
Swiss banks and other entities, of policies of restitution and reparations. German
reuni¬cation also triggered the construction of commemorative memorials, for
example in Berlin. One notable gap in the judicialization of wrongdoing from
World War II is the absence of discussion of Allied conduct, for example, the
¬rebombing of German cities and the dropping of two atom bombs.
These three waves of judicialization represent decreasing levels in the volume
of defendants, although not necessarily in the symbolic value of convictions.
That said, with the possible exception of the Adolf Eichmann trial in Israel
and certain of the concentration camp trials, the expressive, pedagogical, and
didactic value of national proceedings has not approached that of the Nuremberg
A qualitative review of the thick case law emerging from military instrumen-
talities and civilian courts prosecuting World War II atrocity reveals, in a manner
consistent with my ¬ndings from other sites of judicialization, a paucity of discus-
sion with regard to the purposes of punishment, the application of punishment,
or how application may promote purpose. Judges were granted a tremendous
amount of discretion in sentencing. Retribution and expressivism received stray
mentions as purposes of punishment. Initially, sentences by military instrumen-
talities gravitated more toward the death penalty than in the Nuremberg or
Tokyo proceedings. What is more, many of these death sentences quickly were
carried out, especially in cases of former concentration camp of¬cials appearing
before military courts (for whom death sentences were the norm). For example,
¬fty-eight of the sixty-one defendants charged by an American Military Tribunal
in relation to the Mauthausen Concentration Camp (First Mauthausen Trial)
were sentenced to death (nine of these sentences later were commuted to life
imprisonment).227 Soviet military instrumentalities sentenced many individuals,
including Russians and Ukrainian collaborators, to death.
In many cases brought in Germany in the 1960s, defendants “ even those ac-
cused of involvement with concentration camps “ were acquitted.228 In terms of
convicted defendants, sentences of national courts “ especially German courts “
were somewhat lenient. That said, leniency was not evident in all national courts.
After all, the Israeli Supreme Court in 1962 upheld Eichmann™s death sentence.
However, even at the level of military instrumentalities run by the occupying
112 Atrocity, Punishment, and International Law

powers, pressure soon emerged to parole most of those convicted. The onset
of the Cold War dampened U.S. interest in prosecuting Nazi crimes. Amnesty
laws were passed in a number of national jurisdictions. The situation of prose-
cutions in Austria is indicative: in immediate postwar years, 17,500 individuals
were prosecuted in national courts (43 were sentenced to death, of whom 29
eventually were executed); in 1949 the Austrian government enacted an amnesty
law for those only loosely implicated in the Nazi regime; and in 1957 a general
amnesty was enacted for all members of the Nazi Party.229 Looking back from
a perspective sixty years after the end of World War II, it is clear that, although
some perpetrators have faced legal process, many have evaded it.

a. Immediate Aftermath of the War
Many trials were held by military courts and commissions, along with national
courts, in the late 1940s. Thousands of individuals were charged and convicted.
These proceedings involved a medley of international law, municipal law, and
military law. Although the Nuremberg judgments frequently were referenced,
the proceedings largely remained dependent on provisions of ordinary municipal
criminal law.
The case reports show very little, if any, thought given to penological pur-
poses, although some discussion is found regarding the ¬xing of sentence and
mitigating factors. This paucity of discussion occurs even in the handful of cases
where national courts reviewed military commission sentences, such as the case
of General Yamashita before the U.S. Supreme Court.230 One exception is the
judgment of the Netherlands Special Court of Cassation in the matter of Hans
Albin Rauter. Here, the Netherlands Court discussed punishment for extraordi-
nary crime and underscored the expressive value of punishment in this context.
It noted that when a court punishes acts of extraordinary international criminal-
ity it has “the object of giving expression to the sense of justice of the community
of Nations, which sense has been most deeply shocked by such crimes.”231 The
Netherlands Court also underscored the relevance of the gravity of the acts and
the need for punishment to be proportionate thereto.232
Overall, the sentences issued by national courts and military instrumentali-
ties ranged from death (by hanging),233 to long-term imprisonment (at times with
hard labor), to shorter terms of imprisonment. Judges were given tremendous
discretion in the sentencing process. For example, the Polish Law Concerning
Trials of War Criminals provided for death, imprisonment (for life or a term sen-
tence) and, similar to contemporary Rwanda, loss of public and civic rights and
forfeiture of all property of the sentenced person.234 The Netherlands East Indies
Law allowed courts to punish war crimes through the death penalty, life impris-
onment, or imprisonment from one day to twenty years.235 The Chinese Law
Concerning Trials of War Criminals restricted punishment to either death or
life imprisonment in cases of crimes against humanity, crimes against the peace,
and serious war crimes; with regard to other ostensibly less grave war crimes,
punishment could encompass death, life imprisonment, or imprisonment for
Punishment of International Crimes Nationally and Locally 113

ten years and, with regard to other war crimes (ostensibly those of even lesser
gravity), the choice was between life imprisonment and term imprisonment of
not less than seven years.236 Regulation 9 of the British Royal Warrant accorded
a Military Court the ability to sentence a person found guilty to any one or more
of the following punishments: death by shooting or hanging, imprisonment for
life or any less term, con¬scation, a ¬ne, and “ additionally “ restitution.237 The
Canadian law basically was identical.
Interestingly, the Dutch Extraordinary Penal Law Decree increased the penal-
ties for war crimes from those ordinarily available under municipal law for ordi-
nary common crimes, suggesting the increased gravity that may be attached to
these extraordinary international crimes; however, it still left signi¬cant discre-
tion in the hands of the sentencing authority.238 The Norwegian Law Concern-
ing Trials of War Criminals also explicitly increased the sentences for extraor-
dinary international crime over those available under ordinary domestic penal
law. It permitted sentences to be doubled in the most serious cases from what
would be ordinarily available; capital punishment was possible based on aggra-
vating circumstances. The Norwegian government™s reasons for enhancing the


. 4
( 11)