You are on page 1of 27


Prepared for the symposium “Emotions that matter”, University of Tennessee March 6-7, 2003



Jon Elster
Columbia University

I. Introduction
Emotions matter in two ways. They matter because of what
psychologists call their “valence”: they can be intrinsically pleasant or painful,
desirable or undesirable. They also matter because of their “action
tendencies”: they can shape behavior. Often, they matter in the second way
because they matter in the first. The six Frenchmen who killed themselves in
June 1997 after being caught in a roundup on consumers of pedophiliac
material did so (I assume) because they could not stand the unbearably
painful emotion of shame. The same motive must have been at work in two
French women, a mother and her daughter, who killed themselves in 1815

after being raped by Prussian soldiers.1 But emotions can also matter without
any immediate consequences for behavior. I personally observed both the
joy of the French when their team won the World Cup in soccer in 1998 and
the despair when it failed to make it through the qualifying rounds in 2002.
The large impact of these events on the “gross national happiness product”
was not mediated by behavior. Conversely, when emotions do matter for
behavior valence need not be part of the causal story. I run away from a bear
because I want to get out of harm’s way, not to alleviate the painful feeling of
being afraid.
In this paper I consider a set of cases in which emotions shape
behavior directly, with valence having a relatively minor role. These are
episodes of transitional justice, briefly defined as trials and purges of
wrongdoers and compensations to victims that occur after the transition from
one political regime to another. In fact, the set of actions I shall consider is
both narrower and broader than those covered by the definition. They are
narrower, in that I shall mainly focus on punitive behavior, and broader in that
I shall include extra-legal as well as legal forms of punishments. In the extra-
legal behavior we see emotions at work in raw form, whereas in the legal
system they often occur transmuted into a demand for justice.
I shall proceed as follows. In Section II I briefly survey the
main cases of transitional justice on which I shall be drawing. In Section III I
sketch the features of emotion that will be most relevant for the empirical
analyses. In Section IV I discuss the main retributive emotions - their

1 H. Houssaye, 1815. La seconde abdication - la terreur blanche, Paris: Perrin 1906, p.494.
Soviet troops in Germany after 1945 displayed the same phenomenon on a large scale
(A.Beevor, The Fall of Berlin, New York: Viking 2002).

cognitive antecedents and action tendencies. In Section V I offer evidence to

show that these emotions as well as others may shape legal and extra-legal
punishment of wrongdoers. I conclude in Section VI by a discussion of the
transmutation of emotion in transitional justice.

II. Transitional justice

Although historians and others have written about individual
episodes of transitional justice, such as the fate of the American loyalists after
the War of Independence or the Nuremberg trials, it is only in the last decade
that the topic has emerged as a generic one. A milestone was the publication
in 1995 of three large volumes of source materials edited by Neil Kritz. 2
Without exception, the recent literature is limited to justice in the transition to
(or return to) democracy. There are also instructive cases, however, of trials,
purges and compensations in the wake of the restoration of monarchy,
notably the English Restoration of 1660 and the two French Restorations of
1814 and 1815. Another limitation in the literature is the exclusive focus on
20th century cases. The very first instances of transitional justice, however,
are as old as democracy itself. At end of each of two brief oligarchic
regimes, in 411 and in 403 B.C., the Athenian democrats came to terms with
the past in ways that often anticipate recent procedures.
This being said, the bulk of cases involve transitions to
democracy after 1945. The majority arrived in five geographical and
chronological clusters.

2 N. Kritz (ed.), Transitional Justice, vols.1-3, Washington DC.: United States Institute of
Peace Press.

• Western Europe. After the end of World War II, processes of

transitional justice took place in Germany, Italy and Japan, as well as in
a number of countries that had been occupied by or collaborated with
Germany during the war (Austria, Belgium, Denmark, France, Holland,
Hungary, Norway).

• Southern Europe. In the mid-1970s, the fall of military regimes in

Southern Europe induced transitional justice in Greece and Portugal,
whereas in Spain a unique consensual decision was taken to abstain
from opening the past.

• Latin America. When many Latin American countries returned to

democracy in the 1980s, limited transitional justice was observed in
Argentina, Bolivia, and Chile. In Uruguay, a referendum confirmed an
amnesty law enacted by parliament.

• Eastern Europe. After the fall of Communism in 1989-1990,

transitional justice took place in Bulgaria, Czechoslovakia and one of
the successor states (the Czech Republic), the former GDR, Hungary,
Poland and (to a much smaller extent) Romania.

• Africa. In the negotiated demise of the oppressive white regimes in

Rhodesia (1979) and South Africa (1994), an important part of the
bargains that were struck was immunity for the outgoing elite. In South
Africa, though, the work of the Truth and Reconciliation Commission
brought wide publicity to the human rights violations under apartheid.
By contrast, following the 1991 transition in Ethiopia several thousands
of individuals have been tried for wrongdoings under the Dergue

To summarize a complex material, transitional justice requires

the new leaders to face eight sets of decisions:

• Whether to have transitional justice at all. In Spain, Uruguay and

Rhodesia decisions were taken to abstain from coming to terms with
the previous regime.

• Whom to punish for wrongdoings. Basic targets are those who

issue, transmit, facilitate and execute orders to commit criminal acts.
Economic collaboration with an occupying power may also be
prosecuted. After 1945 Belgium and France created a new crime of
“national indignity” for unpatriotic conduct.

• How to punish the wrongdoers. Prison sentences and fines are

commonly used. In Western Europe the death penalty was used. In
Denmark, Holland and Norway it had to be reintroduced in the laws. In
Belgium and France, those convicted of national indignity were
punished by the loss of civil and political rights.

• Whom to sanction in the public sector. In German-occupied

countries, officials were sanctioned (among other things) for stating in
public that Germany would win the war, for excessive zeal in
cooperating with the enemy, and for having social relations with the

• How to sanction officials. The range of administrative sanctions

includes dismissal, forced retirement, demotion, blame (which might
block promotion), salary cuts, and pension cuts.

• Whom to identify as victims. Victims may be classified by the nature

of their suffering, which may be material (loss of property), personal
(harm to body and mind) or intangible (loss of opportunities). In
addition to choosing which of these to identify as “primary victims”,
the new regime must decide which relatives to include among the
“secondary victims”.

• How to compensate the victims. Property is often restituted in kind,

subject to many qualifications. Personal and intangible suffering can be
addressed by financial compensation, and sometimes by giving victims
preferential access to scarce goods such as housing.

• How to proceed. In transitional justice governments invariably decide

to deviate from standard legal practice. Deviations include retroactive
legislation, extension of statutes of limitations, the use of special
courts, the selection of judges and jurors by political criteria,

imposition of collective guilt, reversing the burden of proof, lack of

adversarial hearings and lack of appeal mechanisms.

Even though such violations of due process may occur, we may

speak of a minimal form of legal justice as long as the outcome is not directly
dictated by politics and some uncertainty persists about what will happen.
Political justice, by contrast, occurs when government or parliament carries
out transitional justice directly, bypassing normal or even minimal legal
procedures. In the English and French Restorations, the actions to be taken
against the regicides were decided by the parliaments. In 1945, the Soviets
wanted to turn the Nuremberg process into a show trial, but the Western
Allies blocked their way. Some of the Japanese war trials, however, had
strong aspect of make-believe justice. In many transitions, the incoming
regime has not even bothered with show trials. Leaders and agents of the
outgoing regime have simply been executed, whether for the sake of revenge
or incapacitation. At the end of World War II, this was Churchill’s idea about
how to deal with the top Nazi leaders. Thus in 1945 three options were on the
table. The British wanted extralegal or military action, the Soviets wanted
show trials, and the Americans successfully insisted on having a genuine trial
with real uncertainty whether some of the accused might be acquitted (as
three of them eventually were).
Legal justice, however minimal, must also be distinguished from
private justice, which occurs when individuals or groups take justice in their
own hands, because they do not want to wait for legal trials, do not believe
there will be any trials, or do not believe that the legal punishment will be
sufficiently severe. Private justice can take the form of extra-legal killings, as
in France after 1815 or in France and Italy at the end of World War II. It can

also take the form of deliberate and public humiliation, by tarring and
feathering collaborators 3, forcing them to drink castor oil in public until it
takes effect4, or cutting the hair of women who engaged in “horizontal
collaboration” with the enemy. 5 In Argentina, official transitional justice was
real but limited. It was supplemented by the National Commission of the
Disappeared, which documented 9,000 persons who had “been
disappeared”. Although the commission itself did not name perpetrators,
someone inside it leaked 1,351 names to the press. The named wrongdoers
could suffer serious consequences in their everyday life. One navy captain
who was well known for his brutal acts “has suffered dozen of attacks in
recent years by strangers on the street or people who say he tortured them
and their relatives”.6
The common feature of extra-legal justice, whether political or
private, is that it tends to be directly motivated by the wish for revenge and
other emotionally induced desires. Legal justice, too, may be rooted in similar
motivations, but in a more complex way.

III. Emotion and action

3 This was a standard treatment of loyalists during and to some extent after the American war of
Independence (C. van Tyne, The Loyalists in the American Revolution, Safety Harbor Florida
2001, pp.61, 241, 295). It also occurred occasionally in the Liberation of France in 1944.

4 H. Woller, Die Abrechnung mit dem Faschismus in Italien 1943 bis 1948, München:
Oldenburg, p.281.

5This is the topic of the path-breaking and myth-breaking work of F. Virgili, La France virile,
Paris: Payot 2000.

6 New York Times August 12 1997.


An emotion is a complex bodily and mental state, which is

triggered by perception or cognition and has observable physiological and
behavioral consequences. Here, I focus on emotions that have cognitive
antecedents, ignoring those which may be induced more or less automatically
by the perception, say, of a shape on the path that looks like a snake. I shall
also ignore physiological expressions of the emotions (crying, blushing,
flushing etc.) and limit myself to the tendency for emotions to induce specific
kinds of behavior. It is important to note that the emotion causes a tendency
to act rather than action itself, because the tendency may be checked. In fact,
emotion may check emotion, as when I abstain from giving vent to my anger
because I would be ashamed by losing my cool in public. Even if checked,
however, the emotion may still persist and find other outlets, as we shall see.
I shall focus on four features of the emotion: their action
tendencies, their urgency, their impatience, and their transitoriness. Typical
action tendencies of some key emotions are brought out by Table 1:


Anger Cause the object of the

Indignation emotion to suffer

Hatred Cause the object of hatred to cease

to exist

Contempt Ostracism,

"Sink through the floor"; run away;

Shame suicide

Confess; makerepairs;
Guilt hurt oneself

Envy Destroy the envied object or its


Fear Flight; fight

Love To approach and touch the other; to

help the other; to please the other

Pity To console or alleviate the

distress of the other

Table 1

In Section IV below, I shall make some further differentiations,

but these correlations will do for the time being. Although, as I said, the
action tendencies may be checked, often they are not. As we all know,

emotions are capable of bypassing the machinery of prudential action. 7

Among these “bypass mechanisms” the most important are perhaps the
tendency of the emotions to induce urgency and impatience. I define
impatience as a preference for early reward over later reward i.e. a positive
rate of time discounting, and urgency as a preference for early action over
later action. The distinction is illustrated in Table 2.

t1 t2 t3 t4

A 3

B 5
Case 1:
A 3

B 4

Case 2

A 3

B 6

Case 3
Impatience and/or urgency

Table 2

7 I deliberately write “prudential” rather than “rational”. While impatience makes us behave
imprudently, it does not induce irrationality. Urgency, by contrast, does make us behave
irrationally and a fortiori imprudently.

In each case, the agent can take one and only one of two
actions, A or B. In case 1, these options are available at the same time, in
cases 2 and 3 at successive times. In case 2, the rewards (whose magnitude
is indicated by the numbers) occur at the same later time, in cases 1 and 3 at
different later times. Suppose that in an unemotional state, the agent chooses
B in all cases, but that in an emotional state he chooses A. In case 1, the
choice of A is due to emotionally induced impatience. In case 2, it is due to
emotionally induced urgency. In case 3, it could be due to either or to the
interaction of the two.
Urgency is not always counterproductive. If the emotion arises
in a situation where waiting could be disastrous, it is clearly adaptive. In the
face of acute physical danger, it is often prudent to take immediate flight. In
other cases, however, urgent emotions arise even when nothing would be lost
and something could be gained by waiting. The proverb “Marry in haste,
repent at leisure” reflects this possibility. The desire for revenge can be so
strong that the agent exposes himself to needless danger by acting
immediately. An adulterous person who is overwhelmed by guilt may seek
immediate relief by confessing to his or her spouse, without pausing to think
whether it might not be better for all parties to break off the adulterous
relationship without revealing it. Japanese kamikaze pilots, who were
presumably in a highly emotional state, had to be trained to abstain from
hitting the first target that presented itself and to wait instead until they could
do maximum damage. As these examples suggest, urgency undermines
prudence by making the agent invest too little in information-gathering.
Impatience is the inability to defer gratification - to prefer an
early, small reward over a larger delayed reward. Although the tendency for

emotion to induce impatience seems less well-documented than its tendency

to induce urgency, there are certainly many cases in which emotion tends to
blot out thoughts about the distant future. To the extent that actions and
rewards are correlated in time, it may be hard to distinguish impatience from
urgency. When someone chooses the early action and receives the earlier, but
smaller reward, the choice could be due to either tendency. Often, no doubt,
the two bypassing mechanisms interact.
Emotions, by and large, have a short half-life. Moreover, when
we are in their grip we tend to ignore this fact. If the French pedophiles had
known how horribly they would feel if they were to be exposed, they might
have abstained from buying the material. If they had not known that the
horrible shame would wear off, they might not have killed themselves. A
person might delay taking revenge until the time is appropriate, only to find
that his desire for revenge has abated. If he could anticipate that the emotion
would cool down, he might have chosen to act immediately. Although such
cases exist, as we shall see, I believe they are rare. When we find people
seeking revenge instantly rather than biding their time, it is usually because of
the urgency of their emotion rather than because of anticipation of its decay.

IV. The retributive emotions

I now proceed to discuss the emotions that arise in transitional
justice in a more fine-grained way. Let me first classify the agents that are
involved. First, there are the wrongdoers or, synonymously, the perpetrators.
Second, there are the victims who suffered from the wrongdoings. Third,
there are the beneficiaries of wrongdoing. To these we may add the category
of helpers, who tried to alleviate or prevent the wrongdoings while they were
taking place, and that of resisters, who fought or opposed the wrongdoers

while these were still in power. A further category is that of the neutrals, who
were neither wrongdoers, victims, helpers or resisters. After the transition
there emerge agents of transitional justice, or accusers as I shall call them for
brevity. These include the political actors who decide to proceed with
purges, trials and compensation, and the lay elements (jurors, liberation
committees and the like) who may be involved in implementing their
decisions. More broadly, I also include among the accusers those who are
most vocal in demanding justice. Judges and prosecutors, too, if politically
motivated, may fall in this category. Officials in these professions may also
act as wreckers of transitional justice if they try to obstruct or delay the
process. Often, an individual will occupy more than one category: victims
may become resisters or resisters become victims and either may become
accusers. Wrongdoers may become resisters (before the transition) or
wreckers (after the transition). Many but not all wrongdoers benefit from their
Ideally, one ought to consider the emotions felt by individuals in
each of these categories towards individuals in each of the others. I shall
mainly limit myself, however, to emotions that target wrongdoers, with some
but not much attention to how they are modulated by the status of the subject
who is experiencing them. The main emotions are: anger, two varieties of
indignation, hatred, and contempt. 8 Although in actual situations these
emotions may fuse with one another in analytically intractable ways, I am
going to treat them, somewhat artificially, as distinct and separate. The
emotions have different cognitive antecedents and induce different action

8The following draws on J. Elster, Alchemies of the Mind, Cambridge University Press 1999,

tendencies. As we shall see in Section V, each action tendency also

corresponds, at least roughly, to a specific set of legal and institutional
Consider first the cognitive antecedents. A feels anger towards
B when he believes that B has harmed him without a good reason. The anger
may be weaker or stronger depending on whether A believes that B’s
behavior was negligent, reckless, or deliberate. In the last case, that of
deliberate infliction of harm, the strength of the emotion may also depend on
whether A believes that B wanted to harm him or that he simply didn’t care
whether he was harmed as a means to, or a side effect of, B’s pursuit of
other ends. What I shall call “Cartesian indignation” is the emotion A
experiences when he sees B harming C without a good reason. Compared to
anger, this form of indignation is an impersonal emotion and typically a
weaker one. 9 What I call “Aristotelian indignation” is the emotion triggered
in A by the belief that B enjoys an undeserved fortune. Although this concept
overlaps with the Cartesian one, neither is included in the other.
Whereas anger and Cartesian indignation are triggered by A’s
belief that B has done a bad action, hatred and contempt are triggered by the
belief that B has a bad character. The antecedent of hatred is the belief that
A’s character is evil; that of contempt, that it is inferior or base. Followers of
Hitler thought Jews evil but Slavs inferior. The distinction between action and
character may seem problematic, at least if one believes that a claim that
someone has a bad character can be supported only by pointing to his bad

9 Ernst Fehr (unpublished data) shows that in experiments where individuals are allowed to
punish those who behave unjustly, at some cost to the punishers, “third-party punishments” by
external observers are typically weaker than “second-party punishments” by the victims of the
injustice. Yet even third-party punishments are surprisingly strong.

actions. Yet in some cases, such as sadistic torture, we tend to think that a
single action provides sufficient evidence of an evil character. The same
problem arises for contempt and the correlative feeling of shame. As for the
distinction between hatred and contempt, perhaps it can be exemplified by
the different emotions we feel towards someone who positively wants the
destruction of other human beings and towards someone who simply doesn’t
care whether they are destroyed or not. Earlier I used this distinction to
differentiate among action-based emotions, but it can also serve to
distinguish among character-based ones.
To summarize, we react to the chronically bad with hatred, to
the chronically weak with contempt, to the occasionally and intelligibly weak
(about whom we might think “There but for the grace of God go I”) with
anger or with Cartesian indignation, and to the undeservedly fortunate with
Aristotelian indignation. We feel hatred for torturers and denunciators. We
feel contempt for the opportunists who would enter the Nazi or Communist
part to get jobs they could not otherwise obtain. We feel anger towards the
South African lawyers who failed to speak out against apartheid (the vast
majority) or towards the Norwegian sheriffs who joined the National Socialist
party (again a vast majority) because they would lose their job if they didn’t.
We feel Aristotelian indignation towards the beneficiaries of wrongdoing,
such as economic collaborators in German-occupied countries, the white
liberal elite in South Africa under apartheid, or Swiss banks that profited
from the bank accounts of Jews who did not survive the Holocaust. After the
French restorations, many felt the same way towards those who had
purchased émigré properties at bargain-basement prices after they had been
confiscated by the revolutionary authorities.

These emotions have different action tendencies. Aristotle says

that “much may happen to make the angry man pity those who offend him,
but the hater under no circumstances wishes to pity a man whom he once
hated; for the one would have the offenders suffer for what they have done;
the other would have them cease to exist” (Rhetoric 1382a 2-16). This
observation, while accurate, is incomplete. If A is angry at B, he wants B to
suffer through his, A’s, agency. If B is injured in a car accident, this will not
slake A’s desire for vengeance. In hatred by contrast, agency is not essential.
What matters for the anti-Semite is that Jews cease to exist, whether it be
through his or someone else’s agency. Also, hatred may result in the desire to
expel the object of the emotion rather than to destroy him. Thus “ethnic
cleansing” may take the form of mass deportations rather than genocide.
Cartesian indignation is similar both to hatred and to anger. The
indignant person wants the target of the emotion to suffer, but since the agent
is not himself the victim of the target, the agency is immaterial. The person
who is subject to Aristotelian indignation wants the undeserved fortunate to
be confiscated; again agency is immaterial. Although the agent can hardly fail
to perceive that the target will suffer as a result of the confiscation, this effect
will be the goal of action only if the target is himself a wrongdoer rather than
merely the beneficiary of wrongdoing. Contempt - the action tendency of
which is ostracism - is more similar to anger. The contemptuous man refuses
to have any personal dealings with the target of his contempt. Unlike what
occurs in hatred, the fate of the target is not his concern. Expressions of
contempt often occur spontaneously, as when we turn away from someone
who shows immoderate fear in the face of danger. The effect but not the
intention of such behavior is to induce shame in the target.

V. Emotions in transitional justice

There is evidence that transitional justice exhibits both the
urgency and the transitoriness of the emotions. Consider first urgency. In
transitions to democracy, there is often an urgent demand for justice.
Objectively, other matters such as economic reconstruction might suffer
more from delay. Subjectively, punishment of the former oppressors and
collaborators becomes the more urgent task. The number of extra-legal
executions after WW II in France and Italy, around 10,000 in each country, is
one indicator. The French practice of establishing summary martial courts to
prevent people from taking justice into their own hands is another. Maurice
Rolland, the official in charge of the early stages of transitional justice in
France, asserted that “the government should establish justice before
railroads”. The widespread practice in all German-occupied countries of
cutting the hair of women who had had relations with Germans is even more
telling, because they presented no possible risk that could justify immediate
action on prudential grounds. In the first period after Liberation, there was in
several countries increasing popular frustration with the slowness of legal
prosecutions against collaborators.
The Japanese war crimes trials, carried out in Yokohama under
American auspices, offer another example of the impact of urgency. The trial
of General Yamashita notoriously ignored due process and mens rea
requirements, in favor of strict liability. In his dissenting Supreme Court
opinion, Justice Frank Murphy observed that

No military necessity or other emergency demanded the suspension of

the safeguards of due process. Yet [Yamashita] was rushed to trial
under an improper charge, given insufficient time to prepare an
adequate defense, deprived of the benefits of some of the most
elementary rules of evidence and summarily sentenced to be hanged. In

all this needless and unseemly haste there was no serious attempt to
charge or prove that he committed a recognized violation of the laws
of war. 10

“Needless haste” is indeed the essence of urgency, as I have

defined it. Even where there is nothing to be lost and something to be gained
by waiting, the psychic momentum of the emotion may prove too strong.
The short half-life of emotion shows up in two ways. First, the
desire for retribution is blunted if the most serious acts of wrongdoing
occurred long before the transition. In the first French Restoration in 1814,
even the regicides who had voted for the execution of Louis XVI were
exempt from prosecution. In the second Restoration in 1815, those among
the regicides who had joined Napoleon during the Hundred Days were exiled.
In some of the German-occupied countries, the worst crimes were
committed shortly before Liberation. In Italy, Denmark and France, new and
more oppressive occupational regimes were created after 1942-1943. In
1944-45, the Germans often engaged in scorched-earth tactics as they
retreated. In any case, no occupation lasted more than five years. By
contrast, when the East European Communist regimes fell in 1989-1990 they
had been in existence for fifty years and the worst atrocities were in a
relatively remote past. There was not, therefore, the same urgent demand for
retribution. The Spanish transition of 1978 also fits this pattern.
This picture needs some nuances, since memory and emotion
do not always decay with time. Although Louis XVIII was able to contain
the demands of the émigrés for revenge in 1814, they did feel extremely
strongly about what had happened to them 25 years earlier. A common

10 Cited after R. Taylor, A Trial of Generals, South Bend, Indiana.: Icarus Press, p.163.

saying about them was that “ils n’ont rien appris ni rien oublié” (they have
learned nothing and forgotten nothing). Many of them wanted simply to
restore the ancien régime, including feudal dues and the tithe. Others wanted
to get their properties back, and not merely a monetary compensation. Still
others wanted those who had purchased their properties at bargain-basement
prices to fund the indemnification, or perhaps even be punished for their
acts. At the Court the émigrés behaved with unbelievable arrogance, inducing
for instance Maréchal Ney to join Napoleon during the Hundred Days
because of the contempt they displayed for his wife. It seems reasonable to
assume that their memory had been kept artificially alive by their artificial
existence in exile, during which they had little else to do besides thinking and
talking about how they had been mistreated and hoping against hope for
affairs to take a better turn.
Another instance of lingering memory may be cited from the
Liberation of Italy. In one of the first trials after the Law of July 27 1944 that
created the framework for transitional justice, a court in Grosseto condemned
11 fascists to prison for two or three years for their public humiliation of four
antifascists (by forcing them to drink castor oil) more than 20 years earlier.11
In a society that accords great importance to honor, such humiliation would
be resented deeply and remembered strongly. (Strictly speaking, of course,
the sentence was a legal action and not a form of personal revenge, but the
evidence indicates that in such cases the courts largely acted to preempt and
prevent more drastic acts of private reckoning.) More generally, in societies
with strong codes of honor emotions of revenge seem to form an exception
to the rule of a short half-life. Revenge can go on for years and decades until

11 Woller, Die Abrechnung mit dem Faschismus in Italien 1943 bis 1948, p.183.

each and every offender has been killed, because the social norm that an
offended person must take revenge make it impossible for the emotion
simply to fade away.
Second, the retributive emotions decay if there is a long delay
between the transition and the trials. In trials in German-occupied countries
after WW II, sentencing was almost invariably more severe in the initial
stages than after two or three years (keeping the crime constant). Another
explanation is also possible, however. Aristotle, observed that “men become
calm when they have spent their anger on someone else. This happened in the
case of Ergophilus: though the people were more irritated against him than
against Callisthenes, they acquitted him because they had condemned
Callisthenes to death the day before” (Rhetoric 1380b 11-13). When the thirst
of the gods had been slaked by the sacrifice of some of the guilty, the others
got off more lightly. This explanation is consistent with the fact that the
demand for retribution in Chile and Argentina shows no signs of dying out.
Because no trials (Chile) or few trials (Argentina) took place in the immediate
aftermath of transition, there is a pent-up popular demand that never has been
satisfied, On this background, one may well wonder how long the patience of
the South African people will endure.
Could there be, as discussed above, a causal link between the
demand for immediate justice and the tendency for emotions to decay with
time? If the actors anticipate that they will feel less strongly in a few years’
time, would they not have an incentive to impose severe punishments while
their emotions are still at peak level? In Belgium, on the basis of the
experience from WW I, “it was believed that after a while, the popular
willingness to impose severe sentences on the collaborators would give place

to indifference”. 12 Hence some Belgians wanted the trials to proceed as

quickly as possible, before passion was replaced by a more deliberative
attitude (or by “politics as usual”). In France, many felt that the purges had to
be carried out immediately because “it was necessary to act before the voices
of timidity reasserted themselves”. 13 In most cases, however, I believe the
demand for instant action is simply due to the urgency of emotion.
The demands for legal punishment of wrongdoings closely
parallel the spontaneous emotional action tendencies. As we would expect,
the second-party emotion of anger is stronger than the third-party emotions
of Cartesian indignation. Within the Roosevelt administration, the Jewish
members, notably Henry Morgenthau and Harry Dexter White, felt very
strongly about the Holocaust and adopted the attitude of “an eye for an eye”.
When Henry Stimson reacted to Morgenthau’s “Carthaginian” proposal of
turning Germany into an agricultural economy by saying “you might have to
take a lot off people out of Germany”, Morgenthau responded “Well, that is
not nearly as bad as sending them to gas chambers”. 14 When White cited a
reporter saying that the Morgenthau plan was “a violation of every moral
precept”, Morgenthau retorted that “I suppose putting a million or two million
people in gas chambers is a godlike action”. 15 The attitude of other members
of the administration is summarized as “The war first, the Holocaust

12 L. Huyse and S. Dhondt, La répression des collaborations, Bruxelles: CRISP 1993, p.115.

13 P. Novick , The Resistance versus Vichy, London: Chatto and Windus 196., p.39.

14 Morgenthau Diary (Germany), Washington DC. US Government Printing Office 1967, vol.I,

15 Ibid., p.664.

second”, with particular “concern over war crimes committed against

American soldiers”. 16 They reacted to crimes against Americans by second-
order anger, and to the Holocaust with third-party Cartesian indignation.
Aristotle singled out “sycophants”, i.e. professional informers,
as especially deserving of hatred rather than anger. After 1945, denunciators,
together with torturers, were in fact especially likely to receive the death
penalty, which according to Aristotle is what hatred desires for its object.
Many ordinary acts of collaboration could be due to intelligible even if
unjustifiable moments of weakness, but torture and similar acts were seen to
represent “radical evil”, in Carlos Nino’s phrase. 17 The intuition was that
these individuals are unreformable and should be put away forever or, better,
put to death. In France after 1945, François Mauriac argued on religious as
well as political grounds that this intuition was wrong. It is, in fact,
uncomfortably close to the Nazi view that Jews were unreformable and
should be put to death: “it was precisely by refusing to write off any human
being as irredeemable that the republic would establish its own moral
legitimacy and durably distinguish itself from the totalitarian and authoritarian
regimes it fought”.18 The primate of the Norwegian church, Eivind Berggrav,
was also against the death penalty.
The emotion of contempt targets those whose character is
inferior rather than evil. Here again there is a parallel with Nazi thinking:

16G.J. Bass, Stay the Hand of Vengeance, Cambridge, Mass.: Harvard University Press 2000,
pp.173, 177.

17 C. Nino, Radical Evil on Trial, New Haven, Conn.: Yale University Press 1996.

18A. Sa’adah, Germany’s Second Chance, Cambridge, Mass.: Harvard University Press
1998, p.54.

Followers of Hitler thought Jews evil but Slavs inferior. In transitional justice,
the emotion of contempt corresponds closely to the demand for “civil
degradation”, that is, loss of civil rights. If we accept the Kantian view that
retribution is a form of recognition of the offender as a moral agent, non-
prosecution may also be an expression of contempt. 19 The demand for the
dismissal of tainted individuals from public service may reflect anger or
contempt, depending, as I said, on whether they are condemned for their
actions or their character. Those who joined the Nazi or the Communist party
to keep their job might fall in the former category; those who joined to get a
job in the latter.
After the transition, those who remained neutral may be targeted
for their passivity, and be at the receiving end of contemptuous reactions.
More importantly, the guilt they feel for having done nothing may strengthen
their demand for retribution, as if their post-transition aggression towards the
wrongdoers could magically undo their pre-transition passivity. The tendency
for the neutrals, those in the “gray zone” between collaboration and
resistance, to be especially vindictive seems in fact to be a general
phenomenon. In France, a defense lawyer explained the severity of the first
court sentences by “the fact that many jurors were latecomers to the
resistance and were eager to demonstrate a zealousness which they had not
shown earlier. Later, when the deported came back from Germany, one had
much more thoughtful jurors who [...] did not feel the need to prove
themselves”. 20 The “résistants de septembre”, who suddenly emerged from

19 Claus Offe (personal communication) made this observation in the context of the low level of
prosecution of leaders and agents of the former GDR.

20 H. Lottman, L’épuration, Paris: Fayard 1986, p. 272.


their passivity after the liberation of the territory in August 1944, were often
more zealous in the harassment of women who had had sexual relations with
the Germans.21 In these cases, guilt is transmuted into aggression towards the
wrongdoers. In other cases, guilt may turn into aggression towards the
victims. Commenting on the psychology of the purchasers of confiscated
émigré property after the original owners came back in 1814-15, writes that
“There are some forms of remorse which become twisted and turn into
hatred. The history of the Restoration can be summarized in the famous
saying: ‘Whoever has offended cannot forgive’.”22 .

VI. Conclusion: transmutation of emotion in transitional justice

In transitional justice, it can be difficult to distinguish among
three attitudes towards wrongdoers: the emotionally based desire for revenge,
the desire for substantive justice, and the desire to procedural justice. As
noted earlier, the fact that the demand for justice tends to decay with time
suggests that it has a strong emotional underpinning. In addition, there is an
internal tension within the desire for justice itself. In democracies that emerge
from lawless regimes, whether authoritarian or totalitarian, the new leaders
often want to show their adherence to the rule of law and the “Rechtsstaat”.
The post-1945 trials in Western Europe were shaped by this consideration. In

21 Virgili, La France virile, pp.111-15, 208-9.

22 P. de la Gorce, Louis XVIII, Paris: Plon 1926), p.162-63. In his exhaustive treatment of the
subject, A. Gain (La restauration et les biens des émigrés, Nancy: Société d’Impressions
Typographiques 1928, vol.I, p.348) asserts that “it would be more correct to say that the
purchaser, looked down upon and despised by the former owner, envied and ridiculed by his
neighbor, retreated into a defiant isolation vis-à-vis the regime and, until 1830, gladly posed as a
victim”. These motivational nuances are hard to assess.

Norway, the use of summary trials was dismissed as an expression of an

unacceptable Nazi mentality In Belgium, internment practices were severely
criticized for resembling how things were done “on the other side of the
Rhine”. In France, retroactive legislation was condemned as a Vichy practice.
Anonymous denunciations - another Vichy custom - were not accepted.
The desire for legality often goes together with a strong desire
for a large fraction of the collaborators to be convicted. As Peter Novick
remarks about France, “side by side with this passionate longing [for
retribution] was the attachment of résistants to those principles of justice and
equity which distinguished them from the rulers of Nazi Germany and Vichy
France”. 23 In post-1945 Hungary, the Minister of Justice insisted “both on
the need to observe strict legal procedures and on the need to exercise
revolutionary political justice”.24 In many cases, however, there is a conflict
between the desire for procedural justice and the desire for substantive
justice; between the desire to demarcate oneself from the earlier regime and
the desire to punish the regime as severely as it deserves. By acting on the
second desire, one may thwart the first.
New democracies can resolve this trilemma in one of three
ways. First, one can insist on respect for basic legal principles such as a ban
on retroactive legislation or an extension of the statute of limitations. This
has, for instance, consistently been the approach of the Hungarian
Constitutional Court after 1989. Former East German dissidents imputed the

23 Novick, The Resistance versus Vichy, p.141.

24 I. Deák, “Political justice in Austria and Hungary after World War II’, in J. Elster (ed.),
Retribution and Restitution in the Transition to Democracy, forthcoming from Cambridge
University Press.

same strategy to the German Courts after 1990. In a famous phrase by

Bärbel Bohley, “We expected justice, but we got the Rechtsstaat instead”.
Second, one can frankly and openly accept the need to violate these
principles in an unprecedented situation. After 1945, Denmark and Holland
adopted explicit retroactive legislation, a procedure that was probably
facilitated by the fact that neither country has a ban on retroactivity in the
constitution. The same practice was observed in Hungary. Thirdly, and this
is the most common procedure, one can use subterfuge to try and have it
both ways. In 1948, a Belgian commentator wrote that “The Dutch system
[of specifically permitting retroactivity] is more sincere than ours. The
Belgian legislator pretended to adhere to the principle of non-retroactivity in
criminal law. In reality [the Penal Code]... was made increasingly severe by
so-called interpretative laws.”25 In the same year, a Dutch law professor
“criticized the French who - in order to avoid retroactive penalties - called the
often severe sanctions of the new ‘national indignity’ crime ‘losses of rights’
instead of penalties. ‘This seems to me a mere playing with words; a
confiscation of one’s entire property, or even a loss of certain rights, is as
much a ...[penalty] as say a fine or the deprivation of liberty”. 26 In his
discussion of the French magistrates after the Liberation, Alain Bancaud
confirms the impression of “a certain hypocrisy” in the attempt to impose
retroactivity without having the air of doing so. 27 Similar tendencies were
observed in Italy and Norway.

25 H. Mason , The Purge of the Dutch Quislings, The Hague: Martinus Nijhoff 1952, p.130.

26 Ibid.

27 A, Bancaud, Une exception ordinaire: La magistrature en France 1930-1950, Paris:

Gallimard 2002, pp.113-16.

The trials of the border guards in the former GDR arguably

rested on legal subterfuge. To comply with the clause in the unification treaty
that prosecution could only target acts that were crimes under East German
as well as West German law, the Federal Supreme Court of Germany
reconstructed an “ideal” law of the GDR from supra-legal principles of
natural law. Commenting on the decision, Peter Quint writes that

For all their earnestness and complexity, opinions of this sort seem to
be lacking in candor. The court creates an ideal law of the GDR,
through the use of techniques and principles resembling those current
in the Federal Republic, solely for the purpose of saying that this
hypothetical construct was ‘really’ the law of the GDR and therefore
its application today is not retroactive. [...] It would seem much more
direct and honest to say: The law of the GDR as it actually existed was
unacceptable and therefore we are applying a new law to these cases.
Perhaps under prevailing interpretations of the Unification Treaty [...]
that acknowledgment could mean the end of these cases, but these
issues nonetheless deserve a more general consideration. 28

Transitional justice is haunted by hypocrisy and by the

transmutation of base motives into nobler ones. In a democracy based on the
rule of law, legal justice must be kept strictly separate from private justice
and political justice. In fact, apart from informal social ostracism and
impeachment of high officials, there is no room in a constitutional democracy
for these two forms of extra-legal justice. Yet as I have argued, extra-legal or
pre-legal motives may underlie attempts to come to terms with the crimes of
the past through the criminal justice system.

28 P. Quint , The Imperfect Union, Princeton University Press 1997, p.203.