Punishment : Didier Fassin

Sonja Bäumel & Erich Schopf / Cartography of the Human Body
Sonja Bäumel & Erich Schopf / Cartography of the Human Body

Punishment : Didier Fassin

Is punishment a political concept?1 Indeed, as a concept, it has been mostly discussed by moral philosophers and legal theorists, who have defended the importance of distinguishing between definition (supposedly value-neutral) and justification (definitely value-laden), between the institution in general (viewed as justified) and the act in particular (possibly subject to criticism), between backward-oriented considerations (deontological or retributivist) and forward-oriented arguments (consequentialist or utilitarian).

In fact, there are two sorts of reasons why the concept should be politicized—rather than being contemplated only from moral and legal perspectives. The first reason is structural: punishment largely has to do with a particular arrangement of power, sovereignty, authority, repression, violence, and the state. Of course, it can also be inflicted by a father on his disobedient child or by a teacher on an unruly student, but the remarkable fact in that regard has been the contrast between the increase in the intolerance toward physical and even moral retribution in the family and at school and the acceptance of ever-tougher public policies against offenders, a contrast that attests to the expansion of the state monopoly over the legitimate use of punishment. Significantly, on the 1st of July 2016, the very day when France established a new record for peacetime regarding the demographics of its imprisoned population, its Parliament passed a law prohibiting spanking. The second reason is historical: in recent decades, most countries in the world have gone through a unique punitive moment. I use the term “moment” in reference to a period during which a spectacular extension of the domain of punishment has occurred independently of a rise in crime rate, as well as to a dynamic, as suggested by the etymology of the word that has been maintained with the term “momentum.” What characterizes the punitive moment is this singular configuration in which punishment, the putative solution to the problem of crime, has increasingly become the problem itself due to its consequence in terms of social exclusion, community destructuring, soaring incarceration costs, and ultimately increased crime recidivism. With almost 7 million people under the supervision of the correctional system, more than 2 million of those being imprisoned and this number having been multiplied by eight in 30 years, the United States offers the most extreme illustration of this phenomenon; but statistics collected since the 1990s at a global level indicate that the trend is worldwide (notably in France, where for 10 years I have conducted ethnographic research on the police, the justice system, and the correctional institution) with only few exceptions, such as most Scandinavian countries. In light of these two reasons—structural and historical—the concept of punishment can be deemed intrinsically and circumstantially political.

* * *

In “Violence, Ideality and Cruelty,” a lecture delivered in a seminar at the Collège de France in January 1995, Étienne Balibar notes that one German word, Gewalt, ambivalently corresponds to both power and violence, this conflagration of meanings being revealed by the translation in other languages.2 In the same way, I suggest that the English term “punishment” equivocally refers to what the French would express with three notions: punition, which is the most mundane and generic form; châtiment, which has a religious or literary connotation; and peine, which belongs to the juridical and technical language. Here again, the translation unveils the polysemy of the term. This polysemy should not be seen, however, as an obstacle to the comprehension of punishment but on the contrary as a condition to it. My critique of the prevalent liberal theories of punishment, whether by philosophers or jurists, precisely resides in the fact that this foundational ambiguity is rarely acknowledged. This oversight allows most authors to implicitly limit punishment to its legal sense of the French peine, and thus adopt a normative approach at the expense of a critical one. Reciprocally, it is their tacit use of this restrictive sense that generates their misunderstanding of my alternative reading of punishment, which they consider too inclusive.

The most widely accepted definition of punishment, derived from H.L.A. Hart’s famous 1959 “Prolegomenon to the Principles of Punishment,” provides five decisive criteria: (1) the infliction of a pain or of an unpleasant equivalent (2) to an actual or supposed offender (3) in response to an offense against legal rules (4) that is intentionally imposed by a legal authority and (5) administered by human beings with appropriate roles.3 Although the definition is said to be independent from any justification, it assumes that punishment is both legitimate since it sanctions an offender for the offense he has committed and legal since it is applied under the law for a violation of the law. Actually, Hart himself had called attention to the risk of what he described as a “definitional stop” that would restrain our questioning of punishment, but his important caveat has not been really taken into account by philosophers and jurists, who have mostly focused on the justification while simply using this definition as their starting point. Even social scientists have often adopted it for methodological reasons perfectly grounded but problematically exclusive.

For my part, I want to challenge this definition on three counts: first, it purposely ignores non-legitimate and extra-legal expressions of punishment; second, it questionably takes for granted the centrality of suffering in the administration of punishment; and third, it wrongly presupposes a just and fair distribution of punishment. These three criticisms are crucial for the apprehension of punishment as a political concept. But this discussion of the commonly accepted definition stems from a more general epistemological position regarding the relation of the social sciences to their objects. Indeed, when studying a social fact, one option consists in circumscribing it precisely from the outset through a definition so as to try to recognize it in the real world, whereas the other conversely privileges its progressive emergence from fieldwork through conversations with actors and observation of action. In the first case, the social fact is thus constructed a priori from the author’s viewpoint, while in the second it is established a posteriori from the agents’ perspective. One may call the former approach Durkheimian and the latter Weberian. The choice of one method rather than the other may have significant political implications. In my view, a critique of punishment definitely implies a comprehensive perspective based on empirical work allowing for an understanding of punishment as it is rather than as it should be.

* * *

So, what is in a definition? The typical situation that moral philosophers and legal scholars have in mind when they use their definition of punishment is that of a judge deciding in court the sentence to be given to an offender as a retribution for his offense, which is thus established and punished under the law. It corresponds to the official version of punishment as legally grounded, legitimately imposed, and fairly distributed. This is certainly a common configuration, but far from the most frequent. Let us consider an institution that does not have as its mission to punish but does it routinely and extensively: the police.4 In quantitative terms, their punitive actions are decidedly more numerous than those of the justice system: as an example, in France, there are 135,000 arrests for use of cannabis but 9,000 prison sentences; and to use an even more telling illustration, in the United States, the number of individuals killed by law enforcement agents is 40 times higher than the number of persons executed as a result of a legal process.

Actually, as I realized while carrying out an ethnography of anticrime brigades on the outskirts of Paris for 15 months, policing has become the most ordinary mode of retribution experienced by many individuals, especially those from ethno-racial minorities belonging to the working class and living in disadvantaged neighborhoods. Punishment by law enforcement officers takes multiple forms: harassment through stop-and-frisk; bullying via insults and threats; humiliation by offensive comments and public handcuffing; gratuitous arrests and taking into custody; physical brutality and moral violence. It involves presumed offenders as well as mere bystanders, in large part on the basis of ethno-racial profiling. In other words, it is not necessary that the persons punished be guilty of any specific violation of the law: their culpability is assumed on generic features such as appearance and residence. Two telling illustrations of the disconnection between criminal activity and retributive action are punitive expeditions and random punishment, especially frequent in response to an offense committed against the police. In the first case, it is the whole group, for instance the inhabitants of an apartment block in a housing project who are the victims of the excessive use of force, being pushed, thumped, or pinned to the ground, and when a house search is conducted, have their doors broken and furniture overturned. In the second case, it is anyone in the group, usually the slowest runner or the least lucky passer-by who gets beaten up, arrested, and indicted in place of the culprit.

One could dispute the fact that these acts are considered punishment. Are they not plain abuses of power, sheer domination, pure repression? When it is in response to an offense against the police, is it not simply reprisal? Some have tried to differentiate these terms, often with good theoretical or moral reasons. Thus, Robert Nozick establishes a clear distinction between retribution and revenge.5 The distinction is analytically valuable, with five proposed criteria to differentiate the two, while also ideologically oriented, as it separates us (the civilized who punish) from them (the barbarians who avenge). But empirically the distinction is often difficult to establish and ultimately serves to legitimize or de-legitimize the infliction of pain. How to decide, then, when a given practice can be regarded as punishment? Two kinds of arguments may be used: one, subjective, refers to the meaning given by the agents to their action; the other, objective, concerns the interpretation the analyst can make of the course of action. In the case of the aggressive behavior of the police toward given publics, both arguments speak strongly in favor of punishment.

First, the agents consider it as such. This is true in particular of the officers themselves. It is not only that they would try to mask reprehensible practices of vengeance by fear of consequences from their hierarchy or from the justice system: they do find rational justifications for their acts. On the one hand, they indiscriminately regard those they handle in this way as actual enemies and potential criminals: mistreatment is therefore just desert. On the other hand, they systematically view judges as being too lenient toward the suspects that they arrest: justice in the street becomes a substitute for the magistrates’ supposed indulgence. However, neither justification is factually grounded, since opinion polls indicate that a majority of people, including those in housing projects, have favorable views of the police and even try to collaborate with them, while penal statistics show that judges are increasingly severe in their sentencing, especially with respect to the petty crime typically committed by lower classes. Notwithstanding these discrepancies, for law enforcement agents there is a moral justification to what they deem just punishment. Although most people would describe it as violence, this justification serves to cover-up racist prejudices and discriminatory practices.

Second, the institution also participates in this interpretation. Especially revealing is the fact that over the past two decades, the number of offenses of insulting an officer and resisting arrest has skyrocketed. As the agents themselves and their superiors admitted to me, this incrimination corresponds to situations of police misconduct, either through verbal provocation or use of excessive force. In particular, when the individual is injured, the accusation of insulting an officer and resisting arrest becomes a legal instrument for a preemptive counter-attack in case of complaint for police violence. During the trial, it is the word of the sworn-in agent against the word of the alleged suspect. There is therefore a double punishment, first physically, on the spot, and then legally in court. The institution forcefully backs this practice as the Ministry of the Interior encourages officers to file complaints and request financial compensation, and as it even pays the lawyers. This is part of a broader policy that uses law enforcement as an instrument to enforce a social order in disadvantaged neighborhoods and among low-income populations in the name of the war on crime. Such illegal, illegitimate, and unfair practices serve to inculcate the concerned individuals into their place in society. They simultaneously produce political subjection and subjectivation. The state of emergency declared after the 2015 Paris attacks has extended the discretionary power of the police for these targeted punitive practices.

If we accept these premises, such police interventions involve an institution that has no legal authority to punish and yet does punish outside the law using ethno-racial profiling for stop-and-frisk or unjustified and disproportionate physical force for arrests. Moreover, they affect individuals who are at most suspects, and frequently not even so. They simply belong to stigmatized and dominated social groups, and they often do not have the pretext of an offense having been committed, since it is merely routine behavior. In other words, most of the criteria defining punishment in the legal tradition are absent from the actual practices of punishment by law enforcement agents. Of course such practices are specific neither to the French police nor to the police in general. A more systematic examination of the multiple expressions of state violence exerted against certain individuals or groups in various countries with retributive intentions would show that punishment often considerably exceeds its classic definition. In fact, only one criterion of the latter seems to resist the empirical test: the infliction of pain as a retribution for something that does not have to be an offense but can simply be what those targeted represent, that is, the form of otherness that they embody and the suspicion associated with it.

* * *

That punishment would, in the final instance, be reduced to the imposition of suffering—be it through flogging, imprisoning, or shaming—is probably the least questioned fact about punishment among theorists as well as laypersons. The avoidance of this questioning comes down to a denial of two crucial facts: far from being universal, this reductio ad dolori is temporally and spatially circumscribed; and far from being accidental, it is telling of profound truths rarely acknowledged.

First, punishment has not always and everywhere corresponded to the infliction of pain. In most societies, until a recent period, the commission of a reprehensible action called for a collective response in terms of compensation, generally material. Etymologically, as has been shown by Émile Benveniste, the Greek poinē and the Classical Latin poena, from which the word “punishment” stems, referred to the debt that had to be paid to atone for a crime; only in Late Latin did the connotation of torment appear; it remains explicitly present in the word “pain.”6 Similarly, Latin retribuere, which later gave the noun “retribution,” meant to give in exchange and to restitute what is owed, whether it implied recompense or penalty; interestingly, at the time of the Renaissance, under a differentiated religious influence, the English “retribution” came to have the negative sense of punishment for a wrong, whereas French rétribution took the positive meaning of fair salary rewarding work. Beyond this philological discussion, it is notable that historical findings from ancient worlds as well as ethnological accounts from pre-colonial societies attest to the generalized practice of reparation. Thus, Georg Simmel discusses wergild, which was the sum required to repay a crime, in Anglo-Saxon England.7 Similarly, Claude Lévi-Strauss mentions practices in the form of reparation among contemporary Plains Indians.8 For the Western world, Christianity at the end of the Middle Ages, and for traditional societies, colonization in the Modern Era, radically transformed the moral economy of punishment from a dominant logic of exchange into a dominant logic of suffering. In response to the crime committed, the pain inflicted on the offender replaced the debt owed to the offended or his relatives. Most importantly, whereas the payment of the debt previously involved a collective obligation for the clan or the family, the infliction of the pain henceforth signified the individual liability of the culprit.

Second, the fact that punishment amounts to the imposition of suffering suggests more troubling truths. As noted by Émile Durkheim, the act of punishing often seems to include “an additional torment that serves no purpose,” whose “motivating force is entirely emotional.”9 But why would that be? Why should the judge humiliate the accused with hurtful remarks and embarrassing lectures before sending them to jail, to the point that when I would meet one of them later in his prison cell, he would tell me that in court he could not wait for the end of the trial even when he knew this meant his incarceration? Why would police officers purposely put handcuffs on a suspect they have arrested incorrectly so as to painfully twist his arms and make fun of his complaint while taking him into the precinct for questioning? Why would a corrections officer refuse a shower to an inmate and ostensibly let another take one as they return from the yard where they have played sports? I intentionally use mundane illustrations here, avoiding much more dramatic or tragic cases that result in the death of an individual in the hands of police or corrections officers. This excess in the practice of punishment has been interpreted by Nietzsche in the most explicit and powerful manner, as “the voluptuous pleasure de faire le mal pour le plaisir de le faire.”10 In the assimilation of punishment with suffering and, even more, in the unnecessary torment that is added to it, one cannot not recognize the manifestation of bare cruelty. Such ruthlessness is not, as is often argued, the fact of outliers, of cynical judges or sadistic officers. It is not the exception: it is part and parcel of retribution. In fact, society delegates to certain institutions and certain professions the power to mete out punishment so as to be in excess of what it is supposed to be. Typically, a prison sentence is never only a deprivation of liberty: it is also a deprivation of autonomy, of privacy, of affective and sexual life, of the right to decide on one’s most personal facts and ordinary events of life, and ultimately, in most cases, for lack of work, education, and rehabilitation, it is a deprivation of the very meaning of the sentence.11 The culmination of cruelty is solitary confinement, which is a punishment within the punishment, an unbounded torment added to the normal infliction of pain, which concerns 80,000 inmates on any given day in the United States.

* * *

To understand how punishment can escape its official definition and exceed its determined limits, how extra-legal and non-legitimate retribution is normalized, how suffering is imposed with little restraint, one has to take into account a dimension that has not received the attention it merits in classical theory: the distribution of punishment. It is because certain individuals and certain groups are more easily targeted, that for them such exceptions may become the rule, that policing may be an ordinary mode of punishing them, and that the infliction of pain may seem justified in their case as a form of retribution.

The distribution of punishment across society is, however, obscured by the fact that the differentiation of offenders appears to be a mere differentiation of offenses. It is not that African Americans are more harshly punished, the champions of law and order policies contend, it is that they commit more crimes deemed punishable. Indeed, the question “who is punished?” is strictly linked to another: what gets punished? As shown by Michel Foucault, the distribution of punishment is the consequence of the differentiation of illegalisms.12 Not all offenses are treated in the same way. Today, in France, tax evasion is better tolerated socially than shoplifting, and in the United States, financial crime is settled through negotiations leading to fines for the company, whereas petty crime is settled through plea-bargaining leading to prison sentences for the accused. The explicit differentiation of illegalisms is itself in large part conditioned by an implicit differentiation of populations in terms of their punishability, that is, the social inclination to punish their deviant practices. But whereas Foucault tends to link the selective severity in the eighteenth and nineteenth century with the economic impact of illegalisms, robbery being apparently more damaging to capitalism than fraud, I would argue that in contemporary society it is rather the moral evaluation of perpetrators that is at stake, the undesirability of ethno-racial minorities, immigrants, and the poor, rendering them definitely more punishable a priori. More than the quality of the offense it is the quality of the offender that explains the intolerance of society towards certain illegalisms rather than others.

Dominated categories are, thus, those deemed punishable. They are also those for whom punishment can be something different from what it is supposed to be and imprisonment more than the deprivation of freedom. In the French short-stay prison where I conducted an ethnography, black and Arab men represented two-thirds of the inmates, half of whom declared themselves unemployed. During the four years of my research, the number of prisoners increased by one-fifth as the facility became more and more overcrowded. In contradiction with the centennial legal principle of single-cell housing, cells were occupied by two inmates, although their nine square meters were designed for one person. The indignity of the prison system reflects the assessment of their occupiers worthiness. Translated into a hierarchy of illegalisms, the hierarchy of populations ultimately produces an unequal distribution of punishment not only in quantitative terms (the over-representation of ethno-racial minorities and socioeconomically disadvantaged) but also in qualitative terms (the carceral condition being an indication of it).

* * *

Punishment is not what it is said to be. Both common sense and theoretical discourses tend to represent it in a way that not only justifies it on utilitarian or retributive ground, but also legitimizes it as an institution that is necessary for the common good and the social fabric. The empirical examination of punishment does not confirm this representation. Without adopting a normative stance regarding the justification or legitimacy of punishment, my argument is that what is assumed to justify and legitimize it is simply unfounded. The critical approach I have tried to develop has therefore proceeded in two steps. First, the confrontation of the norm with its implementation revealed the discrepancy between the two: punishment exceeds its definition; its illegal use is trivialized; its social distribution is unfair; its equivalence with the infliction of pain is problematic. Second, the interpretation of this discrepancy unveils its logics: the determination of punishable populations has for counterpart the impunity of others, both processes being made opaque via an apparently neutral hierarchy of crime; punishable populations correspond to undesirable social categories both economically and ethno-racially, whose alleged unworthiness renders possible their mistreatment by the repressive insitutions; punishment is therefore better understood not in terms of justice but in terms of unequal allocation of power relations and social resources.


Didier Fassin is the James D. Wolfensohn Professor of Social Science at the Institute for Advanced Study in Princeton and holds a Direction of Studies in Political and Moral Anthropology at the École des Hautes Études en Sciences Sociales in Paris.


1. See Didier Fassin, The Will to Punish (Oxford: Oxford University Press, 2018 [2017]).

2. Étienne Balibar, “Violence, Ideality and Cruelty,” in Politics and the Other Scene, (London: Verso, 2002 [1997]).

3. H.L.A. Hart, “The Presidential Address: Prolegomenon to the Principles of Punishment,” Proceedings of the Aristotelian Society 60 (1959–1960), pp. 1–26.

4. Didier Fassin, Enforcing Order: An Ethnography of Urban Policing (Cambridge: Polity, 2013 [2017]).

5. Robert Nozick, “Retributive Punishment,” in Philosophical Explanations (Cambridge: Harvard University Press, 1981), pp. 363–397.

6. Émile Benveniste, Dictionary of Indo-European Concepts and Society (Chicago: HAU Books, 2016 [1969]).

7. Georg Simmel, The Philosophy of Money (London: Routledge & Kegan Paul, 1978 [1907]).

8. Claude Lévi-Strauss, Tristes tropiques (Paris: Plon, 1955).

9. Émile Durkheim, The Division of Labor in Society (New York: The Free Press, 1984 [1893]).

10. Friedrich Nietzsche, On the Genealogy of Morals (New York: Vintage Books, 1989 [1887]).

11. Didier Fassin, Prison Worlds: An Ethnography of the Carceral Condition (Cambridge: Polity, 2016 [2015]).

12. Michel Foucault, Discipline and Punish. The Birth of the Prison (New York: Pantheon Books, 1979 [1975]).