Criminal Law and Philosophy

, Volume 2, Issue 1, pp 67–84

Punishing Cruelly: Punishment, Cruelty, and Mercy


    • The Law School and the Committee on Social StudiesHarvard University
Research Paper

DOI: 10.1007/s11572-007-9038-x

Cite this article as:
Barrozo, P.D. Criminal Law, Philosophy (2008) 2: 67. doi:10.1007/s11572-007-9038-x


What is cruelty? How and why does it matter? What do the legal rejection of cruelty and the requirements of mercy entail? This essay asks these questions of Lucius Seneca, who first articulated an agent-based conception of cruelty in the context of punishment. The hypothesis is submitted that the answers to these questions offered in Seneca’s De clementia constitute one of the turning points in the evolution of practical reason in law. I conclude, however, by arguing that even the mainstream punitive practices of contemporary western societies fail to meet the modest imperatives of the rejection of cruelty and the unconditionality of mercy propounded by Seneca.




With its norms, law weaves together experience and values. Under the influence of the humanistic sensibility of the Renaissance1 and later of the moral worldview of the Enlightenment,2 modernity initially experienced an unwavering sequence of penal reforms reflecting the rejection of cruelty and the cultivation of mercy. From one country to another the repudiation of cruelty and the imperative of mercy operated as regulative ideals curbing some of the zealousness of socially organized reaction to behavior perceived as anomic or antisocial. By and large, this reformist élan is now gone.

Absent the élan, we now seem to be preoccupied mostly with “marginal” problems of the repressive and punitive apparatuses of contemporary states. Indeed, the critical and reformist agendas in the rich North Atlantic and other parts of the world tend to focus on problems such as the racialization of prosecution and punishment, the unjustified impact of the criminal law system on vulnerable minorities, the adverse impact of low-income and education on criminal defense, the criminalization of sexual behavior and reproductive choice, the conditions of imprisonment, capital punishment, and the use of criminalization and retribution as electoral currency. These problems are perceived as “marginal” and understood as deviations from the norm, as the will-to-punish gone insane, furious or inequitable. To focus on such problems is of course not only justified but also urgent.

However, one of the effects of the focus on deviation is, unsurprisingly, the overlegitimation of the norm, of the mainstream of our punitive practices. In this context, the assumption seems to be that “mainstream” criminalization, prosecution and punishment are adequate and the real need is for more of it. If only our societies could bring some normalcy to the pathological fringe areas of crime and punishment, this view that I am describing submits, our humane and enlightened aspirations would have been met and our will-to-punish redeemed.

This view is mistaken in two important ways. First, the distinction between norm and deviation fails to adequately capture the ideological and functional integration and complementarity of criminal law in modern societies. In fact the distinction between mainstream and fringe in criminal law serves to veil the systemic unity of state-backed repressive practices. But it is also mistaken to think that the punitive practices wrongly perceived as normal meet the normative standards that practical reason in law became capable of imagining. Perhaps because it never was fully reflective, early modern penal reform movements failed to move accepted punitive practices and attitudes beyond a compromise between an unconditional rejection of cruelty and traditions, costumes and occasional waves of hysterical mass reactions to actual or perceived increase in criminality. Blackstone’s compromise in his restatement of the common law is an early example of a limited reflectivity that, though of course changing over time, continues to mark mainstream criminal law and punishment and to haunt modern efforts to legally constitute a cruelty-free criminal policy. He wrote:

“But the humanity of the Englifh nation has authorized, by a tacit confent, an almoft general mitigation of fuch part of thefe judgments as favour of torture or cruelty: a fledge or hurdle being ufually allowed to fuch traitors as are condemned to be drawn; and there being very few inftances (and thofe accidental or by negligence) of any perfon’s being emboweled or burned, till previoufly deprived of fenfation by ftrangling.”3

Trapped in the predicament shaped by these two mistakes, we ought to seek escape. In seeking a way out I turn back to Lucius Seneca’s De clementia for help.4 In this essay Seneca left us the first comprehensive treatment of the problem of cruelty5 and the demands of mercy6 in the context of punishment. As we look back to Seneca in his struggle to educate and restrain emperor Nero’s will-to-punish, we are reminded of the promises of goodness and decency that we have made ourselves for the last two millennia.

Indeed, to return to Seneca is a way to engage in cultural self-introspection. While his teachings failed to nurture in Nero a merciful sensibility and prevent his cruelties, Seneca’s arguments cast a lingering spell on power.7 Hence, in revisiting these arguments we begin a journey to the center of our legal and political civilization, and therefore to the center of our civilization simpliciter. As is always the case with law, though, we do not just jump there; we have to actually travel all the distance that separates us from our destination. To undertake this journey in the company of ancient ideas about punishment means that we will have to learn, in the multitude of their details and implications, the normative vistas they started to open. This certainly is not the only path to the center of our civilization, but this is the path a legal experience that hopes to be increasingly reflective must follow. And whatever it is that each of us will find at the end of this path, and no matter how strenuous the journey may prove to be and how severe the institutional and attitudinal changes it may demand from us, we would all be ill-advised to ever return from it.

Four Conceptions of Cruelty

The legal and, more broadly, political import of the rejection of cruelty should not surprise us. Even in its archaic philosophical origins in Seneca the repudiation of cruelty was already profoundly interlaced with law. Whereas bestiality—as ferocity not accompanied by a punitive or otherwise legal intention—belonged to the realm of brute nature and instinct, the preoccupation with cruelty was primarily contained within the sphere of criminal law, posited or natural. In fact, the history of the evolution of practical reason in law can be at least in part told as the history of the development of four conceptions of cruelty and the influence of the denouncement of cruelty, suffering and human vulnerability in the formation of legal consciousness and institutions.

In the laws of individual rights, disabilities, family, health care, and animal protection, to give a few examples, in operation seems to be a conception of cruelty that foregrounds the unjustifiable sufferingof victims of neglect, discrimination, abuse, and brutality. In the contexts of national and international criminal law and culpability, the concept of cruelty is sometimes used to identify and evaluate a particular state of mind of agents of criminal behavior in terms of delight or other form of hedonistic gratification in causing suffering. In the context of the prohibition of torture and other cruel punishment—or, less commonly, in the discrete area of executive decisions on whether to grant mercy—the idea of cruelty seeks to mark and check the behavior of the agents of punishment in cases in which it was unreasonably instituted, disproportionately applied, or excessively enforced.

Finally, a fourth conception of cruelty seems to be present, although rather implicitly, in our commitments to inherent and unconditional human dignity, general solidarity, and structural conditions of justice. Bringing together ideals of rights and justice, these commitments provide, in some form or another, the normative foundation for the extensive juridical corpus and policy initiatives regulating and managing the institutional continents of social welfare and solidarity, equality and liberty. In their synergetic interconnections, norms and policies set on this normative foundation constitute, on the one hand, part of the means by which we face structural problems of injustice in modern societies and, on the other, the matrix of fundamental individualand collectiverights.

In these four conceptions, two criteria of identification stand out. The first criterion separates agent-based from victim-based conceptions of cruelty, while the second one distinguishes between objective and subjective conceptions. The first criterion underscores where, in the perpetrator–victim relationship, the definitional focus of cruelty falls. In relation to the second criterion, objective conceptions define cruelty on the basis of specific kinds of behaviors, measured against objective and general standards, or as the victim’s predicament. In contrast, subjective conceptions further require, on the part of the active agents of cruelty, some degree of subjective gratification in causing suffering or, when it comes to the victims, some minimal degree of self-consciousness about one’s own suffering or predicament. In light of these definitional criteria, the four conceptions can then be said to be agent-objective, agent-subjective, victim-subjective and victim-objective/agent-independent.

Although important, these distinctions are however more analytical and heuristic than descriptive of actual legal experience. In modern law the four conceptions of cruelty relate with one another in a semantic and normative reciprocal regime of foregrounding and backgrounding. Consider, for example, the debate about whether the death penalty amounts to cruel punishment. In the American context,8 this debate is played out against the background of questions about the possible structural conditions behind the discriminatory manner in which the death-penalty seems to afflict specific social groups rendered more vulnerable on the basis of race or social station.9 Consider, furthermore, how a discussion of “less brutal” methods of execution interacts with a critique of the legitimacy of state killing in the face of claims about the inherent dignity of each and every person.

Of these four conceptions of cruelty, Seneca introduced the first. In De clementia, for the first time normative reasoning about punishment took the form of a somewhat systematic reflection on the repudiation of cruelty and the corresponding duty of mercy in the context of punishment. It is to the ideas about cruelty and mercy articulated by Seneca that we now turn.

Cruelty and Mercy in Seneca

If we condition the achievement of a canonical status by any literary work to its “strangeness,” a strangeness “that we either never altogether assimilate, or that becomes such a given that we are blinded to its idiosyncrasies,”10 we could explain the canonical status of Seneca’s De clementia by the way its strangeness has so deeply influenced our conceptions of cruelty and mercy. We are now not only no longer capable of freeing ourselves from its intellectual grip but, more fundamentally, the mere formulation of a desire to free our legal cultures from this grip impacts us as unthinkable.11 Our journey back to Seneca is therefore a journey back to the first formulation of what is now very much an integral part of our moral worldview and the corresponding legal meaning matrix upon which we stand. When it comes to cruelty we are, I believe, forever condemned to such a journey back to Seneca. And when we think of the argumentative apparatus developed over the ages to curb and discipline power, Seneca is once again a rich source.

De clementia, which was written by AD. 55 or 56 to the then young emperor Nero, is an essay by a stoic philosopher trying to persuade an ill-inclined pupil, who happened to be the mightiest person in the world of his time, of the advantages of mercy. Stylistically structured as a monologue addressed to Nero, the essay lacks many of the analytical virtues we have come to expect from modern analytical philosophy. Moreover, Seneca’s essay does not conceal the way in which political considerations and analytical arguments are bound together.

Nero (37–68 AD.), the last emperor of the Julio-Claudian dynasty (54–68), had been adopted by Claudius in order to become heir to the Roman Empire. He succeeded, after the suspicious death of Claudius, to the throne in 54, at the age 17. In 68 a coup forced Nero out of power and into hiding. Fearing violent death ordered by the Senate, Nero is said to have commited assisted suicide.

After an initial period of normality, his reign deteriorated into a festival of brutality and cruelty, which was precedeed by Nero’s assassination of Britannicus, Claudius’s biological child and favorite of some to take the throne from Nero, and his own mother Agrippina who, after severed relations with Nero, was preparing the young Britannicus to claim the throne. In a struggle to consolidate power after 62, Nero is said to have cruelly killed real, suspected and imagined opponents and, with equally unremitting violence, quashed rebellions and treated enemies of war.

It is in this context that De clementia offers a conception of cruelty as excess in punishment, as an act of punishment that goes beyond the limits established by natural law. The conception of cruelty to be found in Seneca’s De clementia presupposes a victim who suffers and necessitates action or omission that causes this suffering by the violation of a prescribed standard of behavior. Centered on the actual behavior of the agent of cruelty, this conception is agent-oriented and bound up with the defense of mercy as both a virtue and a duty. This objective, agent-oriented conception has had, since its insertion in legal culture, a long and very distinguished history. In fact, it was for centuries the dominant conception of cruelty alongside the agent-subjective one introduced above and originally formulated by Thomas Aquinas.12

The following passage clearly defines punishment as the appropriate domain of cruelty while setting aside the subjective elements of gratification or compulsion in causing suffering as necessary elements of a notion of cruelty:

“But,’ you say, ‘there are some who do not exact punishment and yet are cruel, such as those who kill the strangers they meet, not for the sake of gain, but for the sake of killing, and, not content with killing, they torture, as the notorious Busiris and Procrustes, and the pirates who lash their captives and commit them to the flames alive.’ This indeed is cruelty; but because it does not result from vengeance—for no injury was suffered—and no sin stirs its wrath—for no crime preceded it—it falls outside of our definition; for by the definition the mental excess was limited to the exaction of punishment. That which finds pleasure in torture we may say is not cruelty, but savagery—we may even call it madness; for there are various kinds of madness, and none is more unmistakable than that which reaches the point of murdering and mutilating men. Those, then, that I shall call cruel are those who have a reason for punishment, but do not have moderation in it, like Phalaris, who, they say, tortured men, even though they were not innocent, in a manner that was inhuman and incredible. Avoiding sophistry we may define cruelty to be the inclination of the mind toward the side of harshness. This quality mercy repels and bids it stand ajar from her; with strictness she is in harmony.”13

Seneca’s argumentative strategy in De clementia is well known to readers of ancient philosophy and history. Weaving together pragmatic enticements and conceptual pressure in his efforts at persuasion, Seneca starts out by cultivating Nero’s vanity, acknowledging just how great and incomparable his power, only controlled by himself, was. He then proceeds to try to persuade his pupil that such power is so evident and incommensurable that no further proof of it is required. In the sequence, in an anti-Machiavellian move avant la lettre, he accepts the challenge of trying to persuade Nero that the greatest assurance for any prince is the love and respect of his subjects and to show just how mercy is a habit indeed appropriate to inspire respect and affection for the prince in his subjects. Cruelty then makes its appearance in the text of De clementia as an excess of the right to punish and contrary to mercy.

Following the ancient Platonic and Aristotelian traditions of legal and political thought, Seneca starts his essay by placing the public good above private interests and devotions. But in his elected strategy for the persuasion of Nero, this hierarchy between the public and the private has a two-pronged objective: firstly to assure Nero of his unmatched position of power and appreciation as the personification of the Roman Empire itself; and then, second, to bring Nero’s attention to the fact that, exactly because of his station and power, his deeds would always be under an incomparably greater, more attentive, and detailed scrutiny than anyone else’s actions. The expected beneficial consequence of this heightened exposure was to be made clear to the young emperor, for wise men, when they place the public good above their private interests, by necessity tend to hold in higher esteem and regard the emperor upon whom the whole polity rests.14

The path to Nero’s glory or ruin would therefore depend, insists Seneca, on the actual exercise by the emperor of his unmatched power—on his behavior. Because he is the center of the state and, consequently, placed under constant scrutiny and evaluation, the emperor could as easily build his glory as his ruin. From this general predicament in which the all-powerful find themselves, Seneca derived the important advice that cruelty and anger are unbecoming to a sovereign, lest he becomes unable to rise above his subjects. Few things would then be more efficient in building and consolidating a good reputation for the emperor than his being merciful in his deeds toward all those who are, by the very nature of his title and means, below him in power and position. In so doing, Nero was bound to come even closer to the nature of gods15 and farther from the nature both of common citizens and of ferocious beasts. Pointing to the imitation of the gods as the path to imperial glory, Nero is reminded that if he were lenient toward those who deserved to lose their lives and positions, it would amount to granting, like the gods, life and prosperity. This godly kindness would then bring all, “the evil as well as the good, [...] forth into the light.”16

In line with this consequentialist, proto-utilitarian approach, Seneca foresees an equally utilitarian counterargument and then provides a traditional utilitarian defense of limited punishment. The counterargument is to suggest that mercy would actually operate against both the particular, addressed to the individual being punished—and the general—addressed to every other member of society, deterrent effects of punishment, and in so doing conduce to higher levels of anomie and, ultimately, disrespect and threat to authority. This state of affairs could potentially deteriorate to the point of threatening social cohesion and the mechanism of a social order on the top of which sits Nero himself. Preemptively addressing this consequentialist counterargument to his prescription of mercifulness, Seneca submits that mercy is also very much in the interest of the innocent and virtuous, for it works for them as an assurance against judicial mistake or unexpected or unannounced change in the moral code of their society.17 In a move with extraordinary reflective implications, Seneca thus forces into the calculus of power the recommendation to factor in restraint and generosity.

Two other traditional tropes of ancient political philosophy are then summoned by Seneca in order to seal the question to his satisfaction. The first one is the analogizing of medical care and political governance, arguing that many and excessive punishments discredit a statesman as much as funerals of his patients discredit a physician.18 The second trope is the recurrent image of the patriarchal household. Here the nature and extent of the authority of the head of state with its accompanying responsibilities is analogized to that of the Roman pater familiae. Following the example of good fathers in disciplining those under their responsibilities, the sovereign is to consider punishment the last resort to correct and amend his subjects, thus living up to the title of “Father of his Country’.”19 In another fundamental move, to restraint and generosity Seneca now adds caring responsibilities to the portfolio of the political mighty.

Thus, as his argumentative strategy unfolds, Seneca insinuates that sovereign power is to be exercised with divine, healing, and fatherly wisdom and restraint. The sovereign is, in this view, someone to bestow life unto, to heal the evils of, and to protect his subjects. Presenting and conducting the argument for mercy in this vein, Seneca succeeded in establishing two of its most important and resilient characteristics. First, the possibility of mercy does not, in itself, create an obligation for the sovereign to be merciful. In line with this archaic prescription, the bestowing of mercy remains to this day a faculty associated with the prestige and power of the highest offices. Second, this understanding of mercy reinforced a highly hierarchical view of the juridical relationship between the state and the people. In the very act of mercy the office of the chief executive in the polity renews its aura of conspicuous and unmatched power and its claims to legal immunity and transcendence. But as it does so, the exercise of power is already trapped in a normative field from which it will not be able to completely dissociate itself.

In supplementing the consequentialist defense of limited, merciful punishment, Seneca introduces yet another type of argument, that of the “self-mercifulness” of the sovereign who is merciful. I will return to this argument in a moment, but first it is important to allude to Seneca’s stoic move in raising the argument that though there is no equality in virtue, sin makes every person a perfect equal to his fellow humans. “Consider,” he invites Nero, “how great would be the loneliness and the desolation of [Rome] if none would be left but those whom a strict judge would acquit.”20 This leveling of the moral field by the ubiquitous occurrence of sin is yet another normative counterpoint to the above mentioned hierarchical nature of mercy, for it instills the question of desert in the considerations of power deployment. As we shall soon see, though, desert is a relevant consideration for mercy only negatively, by bridging with universal fault the gap between the powerful and the weak.

Another contention recruited by Seneca in his effort to soul-craft his pupil and tender his overgrown ego in relation to the advantages of mercy would later become famous through Rousseau’s metaphysics of political representation. Interesting to note here is that Rousseau, as well as many of the medieval and ancient thinkers preceding him, advances the idea that the political body is forged by the unification of each and every citizen through a common purpose and a particular mode of collective deliberation.21 The difference between Rousseau’s reduction ad unun of the polity and Seneca’s is one of perspective and not of final outcome. While Rousseau adopts an ex parte populus approach, Seneca identifies the private person of the sovereign—in the original sense of body politic—with the state itself—“[f]or if you are the soul of the state and the state your body”22—thus adopting an ex parte principis perspective. Pushing this proposition one logical step farther, Seneca submits that when the sovereign is merciful, he is merciful to himself. Moreover, this self-protection argument, by giving further justification to mercy in terms of preemptive self-defense of the prince, connects mercy to the natural law command to act in self-defense. In this view, the “quality of mercy [...] is indeed for all men in accordance with nature.”23 Once again, Seneca links power exercise to a normative framework, in this case natural law.

The next relevant stop in Seneca’s condemnation of cruelty is the establishment of a proportion and correspondence between the unrivaled powers of the sovereign and their potentialities for virtuous deeds. For if it is hardly virtuous to abstain from doing what one could not do anyway, the greater virtue is in doing what is, prima facie, unusually or unexpectedly good or in abstaining from causing harm which is usual and expected. This argument—and we can already recognize in it what is later repeated by Montaigne and Kant—skillfully inverts the direction of the prince’s power. To prove the supremacy of his power, Nero should do what is most difficult, namely, to use his authority and might to avoid embarking upon courses of action usually associated with the exercise of great power,24 for “any one can violate the law to kill, none but I, to save?”25 To control or, even more so, to self-control a great power is the only virtue to equal great power itself. Self-control is, therefore, the virtue par excellence of the powerful; it is the virtue most coherent, by enactment of nature, with the status of an unrivaled sovereign power.

Seneca does not remain in the province of moral norms and recommendations of prudence, though. He is ready to draw their institutional implications. It is at this point that the prominent theme of the forms of government and their corresponding corrupted variations comes onto the argumentative stage in order to supplement the normative perspective that Seneca seeks to articulate. Closer to the concerns of his pupil, and in line with Aristotle, Seneca distinguishes monarchy from tyranny. And as he reaffirms his commitment to the rejection of cruelty, in this distinction between the legitimate and the corrupted mercy plays a paramount role. It is mercy, he explains, that ultimately distinguishes a king from a tyrant.26 Striking the same chord only now from the perspective of the vice corresponding to the virtue of mercy, Seneca postulates that while kings and tyrants kill, the former does it only to protect the state while the latter “take delight in cruelty.”27 But one is not to mistake the issue here, for the difference between them is a matter of actual deeds and not of labels. The conclusion is then whispered into Nero’s ear—“Mercy [...] makes rulers not only more honored, but safer, and is at the same time the glory of sovereign power and its surest protection.”28

After the legion of arguments and considerations deployed to persuade Nero of the naturalness, godlikeness, utility, justice, and even beauty of mercy, Seneca ultimately establishes mercy as a virtue unconditioned by any circumstances, biographical or contextual, affecting its potential beneficiaries. Every man and woman, young or elderly, no matter from which social position, political station, or what awful and atrocious deeds they may have committed, everyone is entitled to mercy. From the proposition that mercy, as a virtue, is so absolutely independent of the desert of its potential beneficiaries, Seneca follows with the claim that it is exclusively dependent on the obligations and rules of prudence controlling the intention and the deeds of the merciful, who is to be prepared to say that with all his power, he has not succumbed to anger and injustice in punishing. The sovereign is under the obligation to be in command of his self and power, as if he were “about to render an account to those laws which [he has] summoned from decay and darkness into the light of day [...]. Today, [he is to say] if the immortal gods should require a reckoning from me, I am ready to give full tale of the human race.”29

Nothing can excuse the sovereign from his duty of mercy. But contrary to the previous utilitarian line or argumentation, mercy is made to shine not as the efficient means to maximize some power utility but as an absolute and unconditional imperative. This has, of course, great import in explaining why, in the agent-oriented conception of cruelty founded by Seneca, the victim’s desert or other circumstances are irrelevant. It also shows how influenced Seneca was by the classical natural law quasi-aesthetic inclination toward proportion and balance—“until he has combined the whole into a regular and well-ordered production,” as Plato phrased it.30

Under the double requirement of unconditionality and ordered behavior, mercy is, however, limited. The principal limit to mercy comes from the necessity of using punishment in its expressive function. Seneca thus recommends punishment as a tool to reaffirm the distinction between the virtuous and the vicious. Accordingly, mercy is to be exercised not to blunt the distinction between good and bad31 but only to adjust or rectify the severity of the punishment. The gap between the sufficient punishment to display social condemnation and the maximum punishment possible is filled, according to him, with cruelty.

After appealing to a whole constellation of classical philosophical allegories of the political, Seneca concludes his rhetorical seduction of Nero, the success of which was to be measured by the test of history,32 by importing yet another of its recurrent images. This time he turns to the analogy with animal communities. This argument attempts to bring together natural law—with its aesthetic-cognitive preference for symmetry and proportion and the correspondent moral commitment to order—and the bulk of the other arguments—of both principled and consequentialist natures. The contention Seneca seeks to vindicate by analogizing the polity to beehives is twofold. First, that the mightiest ought also to be the quietest: the prince’s quietude in refraining from cruelty is essential to the naturally ordered functioning of human society, for his “power need not be harmful if it is adjusted to Nature’s law.”33 Second, that mercy is according to a larger plan of nature for human communities, for it “did not wish [the king] to be cruel.”34 In fact, nature withheld from him the instruments of aggression it bestowed upon other animals. Nature, he argues, has an appreciation for details, wherein it cleverly “bestow[s] the tiniest proofs of great principles.”35

The agent-objective conception of cruelty is made intelligible in its complex set of assumptions and implications and cast in its pristine philosophical light by prefacing it, as was done above, with the host of arguments summoned by Seneca in preparation for its introduction. But more importantly, it was in the connection with the arguments used to introduce it that Seneca’s condemnation of cruelty found its enduring and pervasive reflective influence. Indeed, the evolution of practical reason in law and the places cruelty and mercy came to occupy in its structure own a great debt to the argumentative seeds sown by Seneca and later harvest by Thomas Aquinas, Michel de Montaigne and those involved in the articulation of the penal philosophies of the Enlightenment.

Turning to the more technical aspects of the concept of cruelty demarcated in De clementia, it is important to notice that the essay is divided into three parts, the third of which—dedicated to the psychology of the virtue of mercy—is lost.36 While the first part deals with the question of the remission of punishment, the second is dedicated to the nature of mercy itself. It is in this second part that Seneca, in an oppositional definitional strategy,37 sets forward his concept of cruelty as the counterpart of mercy.

As symmetrically counterposed to mercy, the notion of cruelty is also deeply immersed in the context of punishment. Seneca sees two motivations for the sovereign’s infliction of punishment: the vengeance of harm caused to him or the vengeance of harm caused to another person. Vengeance, on its turn and as a motivation for punishment, serves two distinct and valuable purposes: compensation and “immunity for the future.”38 Vengeance, therefore, motivates punishment for either of two objectives namely, compensation and deterrence. These two purposes together serve the cause of “discipline” in a well-ordered society. However, Seneca proceeds to acknowledge yet another purpose, this time a spurious one, for vengeance to be served through punishment: to hearken to anger.

Of course Seneca is here clearly confusing archaic retributivism—as when he speaks of vengeance –, a very special kind of it—as when he speaks of compensation39 –, and then a more standard utilitarian justification of punishment—as when he speaks of deterrence. He also confuses motivation—the inspiration for doing something—and purpose—the object of action. And to make things worse he brings anger, which is a traditional motivation—but not an object of action—for punishment in archaic retributivism, to join compensation and deterrence in the set of possible purposes for vengeance through punishment. However defective from a rigorous technical perspective Seneca’s theory of punishment may now seem to us, his move in bringing anger in to join compensation and deterrence as purposes to be secured through punishment serves well his conceptual strategy, as we shall see.

The sovereign is so much better-off than anyone else, insists Seneca, that he does not need compensation or risk immunity for himself. These two purposes of punishment are, for the emperor, moot ones. Hence, the sovereign should only punish when the victim of the deed to be punished is someone else. If, nonetheless, he decides to punish deeds against himself, for this vengeance of his there remains only one possible purpose: anger. And anger, because it is a spurious purpose for vengeance, makes the respective punishment an offense to natural law. Thus, when Seneca relocates anger from the agent’s motivation to the purpose of vengeance, he places it under the objective and exterior censuring jurisdiction of natural law and away from the subjective and innermost domain of the conscience. It is now no more a private matter hidden in the inner loci of human minds but a public question par excellence, concerning the good order of society and subject to an arguably universally valid standard concerned with action—the natural laws.

Nevertheless punishment, even as a means for vengeance, is also allowed, according to Seneca, to have its own purposes distinct from but subordinate to, those of vengeance and its instrumental agent. He lists three such purposes for punishment: “to reform the man that is punished, or [...] punishing him to make the rest better, or [...] removing bad men to let the rest live in greater security.”40 Translating these three purposes into contemporary jargon we would have re-education of the offender, affirmation of social values and satisfaction of social expectation,41 and special deterrence. Seneca’s point here is that each of these three objectives is better achieved by limited, if not light, punishment. In fact, he explains, the milder the chastisement the more efficient the disciplining of the criminal becomes,42concluding with a line that anticipates both Durkheim and Foucault by saying that “the sins repeatedly punished are the sins repeatedly committed.”43

The resulting basic principles of Seneca’s theory of punishment can thus be summarized: vengeance is the motivation of punishment; punishment is the means of vengeance; compensation of harm to self or harm to others, immunity for the future, and anger are vengeance’s purposes; reform of the criminal offender, expression of social values or fulfillment of social expectations, and special deterrence are the specific purposes of punishment, which now appears not solely as a means for vengeance but as an independently justifiable instrument of social discipline.

Following closely this line of argument, it can be said that Seneca is trying to enunciate two distinct problems which could affect the deeds of an all-powerful sovereign. While both problems fall under the jurisdiction of natural law, the first occurs when the prince, motivated by vengeance, uses punishment as a means to satiate the ferocity of his anger. The second problem is the use of punishment as an independent power technique or device beyond its three legitimate purposes. Thus, the first problem has to do with the purposes of vengeance while the second addresses the independent purpose of punishment as an institution.

Moreover, it is important not to confound Seneca’s use of utilitarian and retributive reasoning about the purposes of vengeance with his use of the same kind of reasoning when speaking about the purposes of punishment itself. While he is both an archaic retributivist and a more standard utilitarian concerning vengeance, he is stringently utilitarian in justifying punishment.

The prince would thus morally err when (i) he is paying back personal injury in the case of vengeance, when (ii) he steps out of the strict line drawn by the three legitimate purposes of punishment or, finally, when (iii) his judgment concerning the means to achieve those three legitimate purposes for punishment is defective. Each of these three mistakes is, for Seneca, a sufficient and necessary condition for cruelty. The context of cruelty is hence set by Seneca with reference to compensation, re-education, general deterrence, specific deterrence, and means-end evaluations. A context thus set is, unsurprisingly, bound to forge an objective conception of cruelty; a conception in which the delight or gratification with the suffering of others, the very opposite of empathy or compassion, is neither a necessary nor a sufficient requirement for the correct application of the concept of cruelty.

This is not to say, on the other hand, that in Seneca’s consideration of cruelty no reference is made to mens rea or to the humanity of those who relate to it under the terms of a form of gratification or under the pressure of compulsion. As was said above, Seneca’s essay misses. In the good company of numerous other ancient philosophers, his way of approaching and explaining issues as well as the warranties he offers for his arguments often show a higher commitment to literary beauty and sophistic persuasion than to the analytic virtues common in contemporary legal philosophy. One mark of Seneca’s style is the strong colors he uses to portray cruelty in the second part of the essay before turning to the comparatively more analytical task of defining it. So, for instance, Seneca rhetorically prepares his definition of cruelty—in another move copied by Montaigne—by associating it with madness and bestiality—“when cruelty has changed into pleasure and to kill a human being now becomes joy.”44

Madness and bestiality trespass, for Seneca and then Aquinas, the very boundaries of our humanity. “Cruelty is,” claims Seneca, “an evil thing befitting least of all man, and is unworthy of his spirit that is so kindly; for one to take delight in blood and wounds and, throwing off the man, to change into a creature of the woods, is the madness of a wild beast.”45 Speaking with the words of another prominent Roman, cruelty would offend the Ciceronian consensus juris, that is, the system of customary norms and normative beliefs that sustain the moral order and bestow meaning unto a collectively shared world. However, and this is a fundamental element of the agent-objective conception of cruelty, brutality, “because it transgresses first all ordinary, and then all human, bounds,”46 is an aspect of the negative evaluation of cruelty as a worldly phenomenon rather than as a necessary component of its phenomenology. This allusion to the compulsive or hedonistic aspects of cruel agency does not mean, as will be shown, that those features have, after all, made it to the list of the necessary definitional conditions for cruelty.

As his essay approaches the sharpest definition of cruelty, Seneca tries to systemize the virtues and corresponding vices that are at stake in the context of cruelty. Mercy then appears as an intentional mitigation of punishment. But what if the punishment is just? If it is just both according to positive law and to natural law, would it not be a vice, instead of a virtue, to mitigate it? Is it not unjust not to deliver what is just? Seneca himself said that there is cruelty both in “to pardon all as [in] to pardon none.”47 So, how does he reconcile those seemingly contradictory claims? Seneca’s solution is of paramount importance for it will make room for cruelty as the vice corresponding to the virtue of mercy, where mercy “consists in stopping short of what might have been deservedly imposed.”48

Mercy is a virtue to be exercised within the limits of just punishment, where just punishment is a punishment that as a means of vengeance does not serve anger and as an independent power device serves only its three legitimate purposes as mentioned above. That is why cruelty, as the counterpart vice to the virtue of mercy, has necessarily also to do with punishment. As mercy is a virtue to be exercised in face of a right to punish or to avenge, cruelty is a vice committed in punishing or avenging on the basis of a claim to this right. It is not strictness in punishing or legitimate vengeance that opposes mercy either, but cruelty. In Seneca’s essay, mercy and strictness in exacting punishment not only can but must live together, for “with strictness [mercy] is in harmony.”49 The most stringently just punishment is one that stops short of the most prime facie legally permissible severe punishment; beyond this line we are dealing with cruelty, as “it is a fault to punish a fault in full.”50

In search of further conceptual systematization, Seneca also distinguishes mercy from pity, which for him is not at all a virtue. This distinction is another important building block for the understanding of the agent-objective concept of cruelty in Seneca. Pity is seen as a mental defect, as a frailty of the will in the carrying out of one’s responsibilities, as, in short, an exaggeration of mercy.51 While mercy is a rational inclination pity is, for Seneca, just an emotional reflex befitting “old women.”

Another important distinction, the one between mercy and pardon, is then established. Pardon is not a virtue, since it is based on the mental annulment of something wrong and deserving punishment. Mercy, on the other hand, while reaffirming the historical fact of the harm under consideration, makes room for a punishment in more perfect agreement with natural law. Moreover, and for the sake of what is being argued here, pardon understood as the annulment of an injury forces it out of the domain of just punishment.

An additional reason why, when Seneca speaks of cruelty as harshness of the punitive mind he does not mean to count mens rea as a condition for cruelty, is that this harshness already presupposes, so the architecture of his conceptual edifice requires, a legally established right to punish. As a vice opposed to mercy, cruelty does not apply to other cases outside punishment where the agent might find pleasure or gratification in causing pain and suffering. My interpretive claim here is in all akin to the preceding interpretive claim about bestiality. Seneca’s “harshness of mind” idea makes better sense in the whole structure of his argument if understood as the agent’s voluntarism in being cruel, while taking delight in someone’s suffering is more properly savagery or bestiality.


The conception of cruelty to be found in Seneca’s De clementia presupposes a victim who suffers and necessitates commission or omission that causes suffering by the violation of prescribed standards of behavior. Centered on the actual behavior of agents of cruelty and placed on the normative terrain of virtues, vices and natural obligations, this conception is agent-oriented and bounded up with an understanding of mercy as both a virtue and a duty.

Seneca’s arguments in De clementia can be summarized in seven points: (i) He restricts the questions of cruelty and mercy to the domain of just punishment, for any brutal deed beyond the sphere of just punishment is not properly cruelty but savagery; (ii) Within the limits of just punishment two virtues dwell together and in harmony: mercy and strictness; (iii) And within those same limits two vices are constantly tempting those with access to coercive power: cruelty and pity; (iv) The combination of cruelty-related virtues and vices is a continuum of pity ↔ mercy ↔ strictness ↔ cruelty, where pity is a vicious degeneration of the virtue of mercy in punishing and cruelty is a vicious degeneration of the virtue of strictness. Under the guise of strictness we may fall into cruelty as well as under the guise of mercy we may fall into pity; (v) And because a vice is not, in this conception, opposed to another vice neither is virtue to another virtue, pity is the vice opposed to the virtue of strictness and cruelty the vice opposite to the virtue of mercy. Cruelty is then the case when the agent violates objective norms demarcating the outer limits of strictness. In the case of punishment, cruelty is what fills the gap between strict, necessary punishment and possible punishment; (vi) Confusion is also to be avoided in order not to think that when Seneca sometimes refers to cruelty as a “harshness of the mind” he is indicating that the agent’s enjoyment of the victim’s suffering is a necessary element of the concept of cruelty. Harshness of the mind has to do with the agent of cruelty’s voluntary causation of suffering in the victim rather than with his or her delight in that suffering; (vii) However, and this is what makes Seneca’s conception agent-oriented, any concern with the potential suffering of the victim is an unwelcomed interference of passions into the sphere of reason. As he clarifies, “pity is a weakness of the mind that is over-much perturbed by suffering, and if any one requires it from a wise man, that is very much like requiring him to wail and moan at the funeral of strangers.”52 But this is not to say that he dispenses altogether with the requirement of a suffering victim, for, after all, Seneca’s conception is agent-objective and victim-subjective. This acknowledged, suffering for Seneca is not the strongest definitional element in cruelty. This would come to change only during the Renaissance, under the influence of Montaigne.

An important consequence of the distinction between cruelty and pity in De clementia is that once the context of punishment is defined, it is to be completely assessed on the basis of the compliance of retribution, reeducation, deterrence, and means-end efficacy of punishment with the governing principles of natural law and the virtues it prescribed. However, these principles have for us lost credibility along with their metaphysical foundations. Unsurprisingly, in many respects Seneca’s views on punishment and its limits disappoint the contemporary reader. Indeed, from the vantage point of our richer conceptions of cruelty, Seneca’s is clearly inferior: inferior but not to be abandoned or forgotten.

A reflective stage or practical reason paradigm is superior to another if it meets two cumulative criteria. First, it must be capable of providing a more compelling account of the accomplishments and limitations of a given stage or paradigm than itself could. In addition, the new stage or paradigm must also incorporate the contributions and overcome the limitations of the previous one in a way that constructs a cogent narrative as to how and why it was able to do so.

Under the influence of the cumulative persuasive force of the various stages of reflections on cruelty, in our days cruelty refers to severe violations of the respect, consideration, and care that the unconditional and inherent dignity of individuals command. The violation can be attributable to personified agency and identifiable intention or at least partially to the existence and operation of impersonal factors. What distinguish these factors is that they shape the circumstances of the victims of cruelty by rendering them relatively more vulnerable to violations. These impersonal factors are deemed, in light of the cruelty they engender or facilitate, as gravely defective in light of held conceptions about the requirements of just and decent forms of personal interaction and collective life. Thus, from the perspective of this at once more comprehensive and cogent reflective stage, the agent-objective understanding and repudiation of cruelty understandably pales. But unless its tenable insights are fully incorporated, no reflective evolution obtains either.

It is from the latest paradigm of practical reason in law that the temptation comes to focus our critical and reformist energies in criminal law to questions that are perceived as “marginal.” Those questions are considered marginal to the extent that they are perceived to refer not so much to the core of our punitive practices, which are then deemed to be largely justified, as to issues of institutional dysfunction or to broader social questions of inequality and vulnerability or to deeper moral disputes about the demands of unconditional human dignity.

In the Introduction I argued that the mainstream/marginal mode of understanding the problems of our punitive practices is mistaken. An additional difficulty with this understanding arises when it becomes too often forgotten that if the current reflective stage does not succeed in incorporating the salvageable insights from the previous stages it cannot be said to be superior in any independent and meaningful way.

It is because we yield our allegiance to a truncated version of the latest reflective stage in the rejection of cruelty and demand of mercy that in practice we remain captives of the acceptance of “mainstream” punitive practices and attitudes that celebrate the compromise between, on one hand, an unconditional repudiation of cruelty and, on the other, the claims of tradition and of an insane, furious or inequitable will-to-punish.

To live up to the evolution of practical reason in law we must ask whether, after dispensing with the misguided perceptions of “mainstream” and “marginal” criminal law and punishment, our societies conform to the demand that the gap between punishment sufficient to meet its legitimate and necessary ends53 and the maximum punishment possible not be filled. For if it is at all filled, Seneca teaches us, it is necessarily filled with cruelty.

It would be very hard to argue that even in what our societies accept as mainstream and legitimate punishment, the interval between necessary and maximum punishment is not constantly bridged with state coercion and violence. If we are to take seriously the promises of decency and goodness that we make ourselves we must own the cruel nature of our punitive laws, institutions, practices and attitudes. We must let go of the comfort derived from the misconception that while injustices and cruelty certainly happen, they are a marginal problem that the core of our punitive practices and attitudes has already learned how to correct. And we must find the moral and intellectual energy necessary to forge a new critical-reformist élan of the entire system of organized punishment.

It would be rather appropriate to end with one of Seneca’s typical curses: “for added to all the rest, this is still cruelty’s greatest curse—that one must persist in it, and no return to better things is open; for crime must be safeguarded by crime. But what creature is more unhappy than the man who now cannot help being wicked?”54 In his overwhelming mission of soul-crafting the young Nero, Seneca patiently, stoically, constructed a perspective on punishment based on the praise of its associated virtues and the multisided chastisement of its accompanying vices. Once the lights of De clementia were turned off and the time to leave its learned and agonistic pages arrived, our legal and political traditions walked away having seen something that was theretofore secluded from view. They saw for the first time cruelty explained in one of its fundamental forms and ever since we have lived under a normative horizon that rejects it. If only we could more fully meet its requirements, we could continue to progress with our mutual promises of decency and goodness. The rest is taking sides.


See Montaigne’s essay “On Cruelty,” in Book II of his Essais, which had various modified editions until his death in 1592 (Montaigne 1991).


See Cesare Beccaria’s (1985) manifesto—originally published in 1764. See, in this regard, the following articles of the Universal Declaration of the Rights of Man of 1789: 8. The law shall provide for such punishments only as are strictly and obviously necessary, and no one shall suffer punishment except it be legally inflicted in virtue of a law passed and promulgated before the commission of the offense. 9. As all persons are held innocent until they shall have been declared guilty, if arrest shall be deemed indispensable, all harshness not essential to the securing of the prisoner’s person shall be severely repressed by law.


William Blackstone, Commentaries on the Laws of England: 1765–1769, (, “Book IV,” 370.


Lucius Seneca (2003).


Relevant contemporary discussions about cruelty can be found in Daniel Baraz’s (2003) sound intellectual history of the concept of cruelty in his Medieval Cruelty: Changing Perceptions, Late Antiquity to the Early Modern Period, Judith N. Shklar (1984), Richard Rorty (1989), and John Kekes (1996).


Important contemporary discussions about mercy can be found in Ross Harrison (1992); Austin Sarat and Nasser Hussain (2007); Austin Sarat (2005); Claudia Card (1972); Martha C. Nussbaum (1993).


This is the sort of move that Nietzsche sees as the slave morality’s transmutation of values. See Nietzche (2000). We are, of course, also familiar with what Foucault had to say about this spell. See Foucault (1995).


See Carol S. Steiker and Jordan M. Steiker’s (1995). See also Hugo Bedau’s (2004). Bedau’s edited volume (1997) is a helpful source for the different views on capital punishment in the United States. As far as the history of death penalty adjudication is concerned, the two momentous decisions are the Supreme Court’s constitutional invalidation of Georgia’s death penalty statute in Furman v. Georgia 408 U.S. 238 (1972) and the re-constitutionalization of capital punishment by the court in Gregg v. Georgia 428 U.S. 153 (1976). It is in the Furman case that, in Justice Brennan’s opinion, is found the most extensive and deepest effort to date to reflect on the nature of cruel punishment in the context of American constitutionalism. In his opinion Justice Brennan sought to articulate the four principles under which violations of the Eighth Amendment prohibition of “cruel and unusual” punishments were to be determined. The principles, of which the first was the governing one, were the following: (i) “a punishment must not be so severe as to be degrading to the dignity of human beings;” (ii) “that the State most not arbitrarily inflict a severe punishment;” (iii) that the “severe punishment must not be unacceptable to contemporary society;” and, finally, (iv) “that a punishment must not be excessive. A punishment [being] excessive under this principle if it is unnecessary.” The conceptual shortcoming of this effort is clear: how can the degradation of human dignity serve to elucidate what cruel punishment is when what is cruel is to be defined as that which degrades human dignity? However it may be, in this article the reader will be able to see the seminal formulation of the philosophical views on cruelty appearing decanted here in Justice Brennan’s four principles.


See the statistics presented in support of the claim of racial discrimination in death-sentencing in Georgia in McCleskey v. Kemp, 481 U.S. 279 (1987).


Harold Bloom (1994).


See, for manifestation of this, Jeremy Waldron’s (2005) claim that the prohibition of torture is a legal archetype in his “Torture and Positive Law: Jurisprudence for the White House.”


See Thomas Aquinas’ Summa Theologiae published from 1265–1272, in particular Questions 2a2ae 157 and 159, dedicated to De Clementia et Mansuetudine and De Crudelitate, respectively, as part of his writings on the virtue of temperance (Aquinas 1972).


Seneca, op. cit., 435–437.


Ibid., 369.


Ibid., 371.


Ibid., 369.


Ibid., 363.


Ibid., 423.


Ibid., 399.


Ibid., 374.


See the basic clause of Rousseau’s social contract: “These clauses, rightly understood, all come down to just one, namely the total alienation of each associate with all his rights to the whole community: For, in the first place, since each gives himself entirely, the condition is equal for all, and since the condition is equal for all, no one has any interest in making it burdensome to the rest.” (Jean-Jacques Rousseau 1977a).


Seneca, 371.




This is the sort of move that Nietzsche sees as the slave morality’s transmutation of values (Nietzsche 2000).


Ibid., 371–373.


Ibid., 393–395.


Ibid., 391–393.


Ibid., 391.


Ibid., 359.


Plato (2001).


See Seneca, 363–395.


Which it failed for Nero, as Tacitus (1996) confirms with bleeding examples in his Annals on Imperial Rome, while largely succeeding for us.


Ibid., 409.


Ibid., 411.




Were it not lost, it is difficult not to speculate, Seneca’s and Aquinas’ conceptions of cruelty might be even less dissimilar.


Ibid., 365.


“Deterrence,” in contemporary jargon.


Compensation can, depending on the circumstances, be understood either as a retributive device or as a maximized utility. I cannot address this distinction and its subtleness here. But for the sake of the interpretation of Seneca in De Clementia, it shall suffice to say that compensation is better understood as a question of retribution.


Ibid., 419. Compare this with Blackstone’s theory of punishment. After rejecting the idea that the rationale of punishment would be the “atonement or expiation for the crime committed,” he indicates “precaution againft future offences of the fame kind” as the sole motivation for punishment. Blackstone then proceeds to list the three “types” of punishment consistent with this motivation, namely: (i) “the amendment of the offender himfelf,” (ii) “deterring others by the dread of his example from offending in the like way,” and (iii) “depriving the party injuring of the power to do future mifchief.” But Blackstone not only expounds a typology of just punishment similar to Seneca’s. He further adds some of the objective parameters that Seneca so passionately defended: “The method however of inflicting punifhment ought always to be proportioned to the particular purpofe it is means to ferve, and by no means to exceed it : therefore the pains of death, and perpetual difability by exile, ftavery, or imprifonment, ought never to be inflicted, but when the offender appears incorrigible.” Blackstone’s latter caveat fails, however, to be faithful to Seneca’s strictures with the excuse that “in fuch cafes it would be cruelty to the public, to defer the punifhment of fuch criminal, till had an opportunity of repeating perhaps the worft of villanies.” Ibid., “Book IV,” 11–12.


Or as Feinberg has articulated it: “Punishment is a conventional device for the expression of attitudes of resentment and indignation, and of judgments of disapproval and reprobation, either on the part of the punishing authority or those in ‘whose name’ the punishment is inflicted. Punishment, in short has a symbolic significance.” (Feinberg 1970). A locus that has become a classic of the articulation of a defense for the enforcement of social morality through criminal law is Patrick Devlin’s (1965). For an analysis about the more general expressive function of law see Cass Sunstein’s (1997).


He thus preaches to Nero: “You will more easily reform the culprits themselves by the lighter form of punishment; for he will live more guardedly who has something left to lose. No one is sparing of a ruined reputation; it brings a sort of exemption from punishment to have no room left for punishment. The morals of the state, moreover, are better mended by the sparing use of punitive measures; for sin becomes familiar from the multitude of those who sin, and the official stigma is less weighty if its force is weakened by the very number that it condemns, and severity, which provides the best corrective, loses its potency by repeated application. Good morals are established in the state and vice is wiped out if a prince is patient with chastisement. The very mercifulness of the ruler makes men shrink from doing wrong; the punishment which a kindly man decrees seems all the more severe.” Seneca, 419–421.


Ibid., 421.


Ibid., 423–425.


Ibid., 423.


Ibid., 423–425.


Ibid., 365.


Ibid., 435.


Ibid., 437.


Ibid., 449.


“At this point it is pertinent to ask what pity is. For many commend it as a virtue, and call pitiful man good. But this too is a mental defect. We ought to avoid both, closely related as they are to strictness and to mercy.” Ibid., 437.


Seneca, op. cit., 443.


As understood by each punitive community, as the requirements of honesty and consistency would suffice to thoroughly change whatever the current state of their respective punitive practices is. We would then start from there.


Seneca, op. cit., 397.



I am grateful to students in my law and cruelty seminar at Harvard and to Renata Barrozo, Vlad Perju, Aziz Rana, Marie-Eve Sylvestre and Roberto Unger for helpful conversations about themes discussed in this essay.

Copyright information

© Springer Science+Business Media B.V. 2007