the right to punish (retributivism, communicative theories, self-defence theories)

intro

• Punishment involves the intentional infliction of harm or loss on a person in response to wrongdoing. This raises an immediate moral puzzle — intentionally harming someone is generally impermissible, so what is it about past wrongdoing that changes this? The justification of punishment is one of the most contested questions in philosophy of law

• Four main theories attempt to answer this question: consequentialist theories justify punishment by its beneficial effects, self-defence theories ground it in preservation rights, retributivist theories ground it in desert, and communicative theories ground it in the need to address offenders as rational moral agents

• State your thesis: consequentialist and self-defence theories both fail — consequentialism cannot avoid punishing the innocent and self-defence cannot explain why punishment is permissible after the relevant threat has passed. Retributivism in Moore's form survives as the most defensible pure theory but communicative theories in Duff's form offer a richer account of what punishment should achieve — and the most defensible position combines retributive desert as the ground of punishment with communicative penance as its appropriate form

section 1 - consequentialist theories and their key issue

Consequentialism stated: Consequentialist theories justify punishment by its beneficial effects — deterrence, incapacitation of the dangerous, rehabilitation, and the promotion of social values. Bentham's formulation: the value of punishment must outweigh the profit of the offence. Moore identifies the full range: special deterrence prevents reoffending, general deterrence deters the wider population, incapacitation prevents further crime during the period of punishment, and rehabilitation prevents future crime by reforming the offender. All of these share crime prevention as the good that justifies punishment — making them all instances of the utilitarian theory that punishment is justified if and only if there is a net social gain.

Consequentialism has genuine appeal — it tells a simple plausible story about why we imprison people, it explains why punishment should be proportionate to deterrent needs, and it connects punishment to values we already accept as important. How we should punish is contingent on human psychology and social facts about what deters effectively.

The decisive objections to consequentialism — Moore's reductios:

Scapegoating: if a net social gain justifies punishment then punishing an innocent person is justified whenever doing so produces greater benefit than harm. Moore's DB Cooper case — if punishing an innocent person who resembles the criminal would prevent a wave of similar crimes the utilitarian must endorse it. Most people find this conclusion unacceptable — we should not punish an innocent person even if doing so would benefit society.

Preventive detention: if a psychiatrist knows his patient has dangerous propensities and the patient is on trial but innocent, a utilitarian judge who knows both facts should find him guilty since incarceration prevents predicted future harm. Again the utilitarian must endorse what most people find clearly unjust.

Moore's reductio stated precisely: Premise 1 — punishment is justified iff there is a net social gain. Premise 2 — there is a net social gain from punishing this innocent person. Premise 3 — this person should not be punished. These generate a contradiction — the person both should and should not be punished. The only escape is abandoning Premise 1 — rejecting the utilitarian theory.

The indirect utilitarian response — that punishing the innocent would cause demoralization and that no realistic case would actually produce net social gain — fails because it misunderstands the thought experiment. We only need the case to be conceivable for it to test the theory. If there is any conceivable case where a utilitarian would refuse to punish the innocent they are not a utilitarian.

The unifying worry: consequentialism treats people as having only instrumental value — it uses them as means to social ends, ignoring what they deserve and violating their autonomy and self-determination. This is the deepest objection.

section 1 - self-defence theory and their key issue

Self-defence theory stated:

Self-defence theory starts from the intuition that we have natural rights to self-preservation — rights we possess simply in virtue of being human, not because any state has granted them. The most plausible examples: the right to self-defence, other-defence, and the protection of certain property. Preservation rights make permissible what would otherwise be impermissible — using violence to stop an aggressor — subject to two moral conditions: necessity, meaning the response must be necessary to avert the threat, and proportionality, meaning the response must be proportionate to the threat.

The attraction of the self-defence theory is that it recognises the good consequences of punishing — especially deterrence and incapacitation — while not being a consequentialist theory. It grounds punishment in rights rather than consequences, thereby avoiding the scapegoating and preventive detention problems. You cannot punish an innocent person in self-defence because they pose no threat. The state derives its right to punish from individual rights to protect ourselves, transferred to the state when we subscribe to political society.

Self-defence theory also explains why punishment should be limited — you can only respond to genuine threats with necessary and proportionate force. This limits what the state can criminalise and punish in a principled way.

The decisive objections to self-defence theory:

The timing problem — the fundamental objection: in self-defence we take action to prevent harm. Punishment begins when self-defence has already failed — it is imposed after the harmful act has occurred. The person being punished is no longer a threat in the immediate sense that justifies self-defensive force. Beating up a bully five years after the fact is not self-defence. So punishment cannot be justified as self-defence.

The automation response: self-defence theorists might respond that the criminal law is an automated system for carrying out threats — it is acceptable to threaten harm to prevent harm, and punishment is the automated carrying-out of that threat when it fails to deter. The threat is announced in advance through the criminal law and punishment carries it out when the threat fails.

Why automation fails: even granting that the initial threat was justified as self-defence, why are we permitted to carry out the threat once the harm has already occurred? The justification for the original threat was preventing harm — once the harm has happened the preventive justification is exhausted. Carrying out the threat at that point looks like retaliation not self-defence.

The scope problem: self-defence theory has difficulty explaining the full range of what criminal law punishes. Economic offences like tax evasion do not fit the self-defence model — the state is not defending anyone against imminent physical harm. Regulatory offences like selling alcohol to minors similarly resist the self-defence framing. If self-defence theory cannot account for these it is at best a partial theory of punishment's justification.

The comparison with consequentialism: self-defence theory is superior to consequentialism because it cannot justify punishing the innocent — an innocent person poses no threat — and it grounds punishment in rights rather than aggregate welfare. But it fails on its own terms because punishment is not actually self-defence — it comes after the harm has occurred and the threat has passed.

section 2 - retributivism positive case

Moore's closet retributivism — the positive case: Retributivism is the view that punishment is justified if and only if the offender deserves it. The good achieved by punishment is intrinsic not instrumental — it is not that punishment prevents future crimes or achieves social cohesion, it is that someone who deserves to suffer does suffer. Moral desert is both necessary and sufficient for justified punishment.

Moore's most important contribution is the argument from closet retributivism. Most people who have not worked through theories of punishment are already retributivists in the judgments they make — they believe criminals deserve punishment independently of whether it deters or rehabilitates. His thought experiments reveal this. If a murderer has truly found Christ and is already reformed, if the crime went undetected so general deterrence does not demand punishment, and if we could pretend to punish and serve all deterrent ends — should we still punish? Most people say yes. This reveals that our genuine belief is retributivist — punishment is deserved regardless of consequences.

The Chaney case: a defendant convicted of forcible rape receives a minimal sentence. After sentencing but before imprisonment he has an accident that removes his dangerous propensities and inherits money removing his motive for theft. We could pretend to punish him and serve all deterrent ends. Should he be punished? Most people say yes — which shows they are retributivists because no consequentialist reason remains.

Moore's seven confusions — what retributivism is not: Retributivism is not lex talionis — it requires proportionality not that crimes be met with like acts. You can be retributivist and oppose the death penalty. It is not Quinton's view that only the guilty should be punished — Quinton makes desert a necessary condition, retributivism makes it sufficient. It is not satisfying victims' desires for vengeance — the good is objective not psychological. It is not preference utilitarianism about citizens' revenge desires. It is not the view that punishment prevents vigilante justice — a consequentialist justification. It is not denunciation theory — which is utilitarian. It is not formal justice about treating like cases alike — which says nothing about why anyone should be punished.

The reductio against the mixed theory: The mixed theory says punishment is justified iff there is both a net social gain AND the offender deserves it. This avoids the scapegoating and preventive detention problems since desert is a necessary condition. But it faces its own reductio. In cases where an offender clearly deserves punishment but no net social gain would be achieved — the reformed murderer, the Chaney rapist — the mixed theory says no punishment is warranted. But most people believe punishment is still warranted. Therefore the mixed theory must be abandoned. The only remaining theory is retributivism.

The duty to punish: Moore argues retributivism is not merely a right to punish but a duty. We have an obligation to set up institutions that achieve retribution. This is a strong claim that distinguishes retributivism from permissive theories that merely allow punishment.

section 2 - objections to retributivism

The metaphysical problem — what is desert?: Retributivism requires a clear account of what makes someone deserve punishment and why deserving suffering is an intrinsic good. This is philosophically controversial. What is the connection between wrongdoing and deserving to suffer? Why should the universe contain more suffering just because someone has done wrong? This is the objection that suffering is never intrinsically good and retributivism has no satisfying answer.

The emotions objection: retributive intuitions may be grounded in emotions like anger and the desire for revenge. If they are emotionally grounded are they reliable guides to what we ought to do? Moore acknowledges emotions can be good guides — guilt, for instance, is a reliable guide to wrongdoing. But critics argue that retributive anger is not a reliable moral guide and that basing punishment on it is morally dangerous.

The proportionality problem: retributivism requires punishment proportionate to desert — but how do we measure moral desert and translate it into specific punishments? Eye for an eye is not required but what is? This remains an unresolved practical problem.

Communicative theories as the strongest objection: even granting that desert justifies punishment retributivism says nothing about what punishment should look like or achieve. It simply says offenders should suffer in proportion to their wrongdoing. Duff argues this is morally impoverished — it ignores the offender as a rational moral agent capable of repentance and reconciliation. This is the transition to Section 3.

Interim judgment on Section 2: Moore's reductios decisively defeat consequentialism and the mixed theory. Retributivism survives as the only remaining pure theory that cannot be refuted by thought experiment. The duty to punish is a strong and philosophically interesting claim. But retributivism remains morally incomplete — it tells us that and how much to punish but not what punishment should achieve or why suffering is the appropriate response to wrongdoing. This motivates Duff's communicative account.

section 3 - communicative theories positive case

Duff's secular penance account stated: Communicative theories hold that punishment is justified as a form of moral communication between the offender and the community. The offender is treated not as a dangerous object to be incapacitated or a rational calculator to be deterred, but as a responsible moral agent capable of recognising wrongdoing and making reparation.

Duff's specific account develops punishment as secular penance. In a religious context penance serves multiple purposes — it induces or deepens repentant understanding of the sin, it communicates the sinner's repentance to those wronged, and it helps restore the relationships damaged by the sin. It is both backward-looking — addressing the wrong done — and forward-looking — aiming at reconciliation and restored relationships.

Duff argues criminal punishment should be understood analogously as secular penance. The crime violates the values of the political community, damaging the offender's normative relationships with fellow citizens. Punishment communicates the community's censure of the wrong, aims to bring the offender to recognise and repent it, and constitutes a form of moral reparation through which reconciliation with the community is made possible. Community service orders are Duff's paradigm — they focus the offender's attention on the crime, express apology to those wronged, and constitute a burdensome penance appropriate to the seriousness of the wrong.

Two key communicative components: the state communicates censure to the offender, and the offender expresses something by undergoing the punishment. This is why mere words are insufficient — words cannot convey the full weight of blame or the sincerity of apology. Punishment provides the material expression that gives proper weight to the moral seriousness of what was done

The liberal objection and Duff's three-part response:

The liberal objection: punishment as secular penance intrudes into the offender's inner moral life — their motives, attitudes and feelings — in a way that violates individual autonomy and privacy. A liberal state has no proper interest in the state of citizens' souls. Penance seeks to induce repentance which is a matter of private conscience not public law. Unlike voluntary religious community membership political community is not voluntary — citizens cannot simply leave a state they find uncongenial.

Duff's three-part response:

First — penitential punishment addresses offenders as moral agents and aims to persuade rather than coerce their moral understanding. It is coercive in that it is imposed — but it aims to communicate through rational moral force, leaving the offender free to remain unpersuaded and defiant. We try to get them to hear the message but cannot force them to accept it.

Second — criminal punishment need not be as ambitious as religious penance. It does not need to engage the offender's deepest spiritual concerns or whole moral character. It focuses on the wrongfulness of the specific criminal act and the attitudes directly manifested in it — not on the offender's entire moral being. This keeps secular penance within appropriate liberal limits.

Third — a state that respects citizens as responsible moral agents must address them in moral language. Simply prohibiting conduct on pain of sanctions treats people as subjects not citizens. A communicative system of punishment — one that censures public wrongdoing and tries to persuade offenders to recognise and repent it — is what we are owed as moral agents. Simply deterring people treats them as rational calculators not as persons capable of moral response.

The limits of community — Duff's honest concession: Duff raises without fully answering whether some crimes are so destructive of community that communicative punishment is no longer possible. His three examples: single horrific crimes — his answer is no single deed should put someone beyond civic redemption. Persistent dangerous offenders — he is less confident, acknowledging that permanent preventive detention might be warranted in extreme cases. Terrorism — the hardest case, raising questions about whether terrorists should be treated as criminals or enemy combatants. Duff is genuinely uncertain here which is philosophically honest.

section 3 - objections to communicative theories

Why does communication require punishment?: The central challenge. If the goal is moral communication — censure, repentance, reconciliation — why does this require the intentional infliction of suffering? Cannot the community simply express blame and invite apology without imprisonment or other punitive measures? Duff's answer is that words alone are insufficient — some wrongs require a burdensome material expression to give proper weight to the moral seriousness of what was done. But critics argue this anthropomorphises punishment in a way that smuggles in retributivist intuitions — we think punishment is appropriate because wrongdoers deserve to suffer, and then construct communicative rationalisations around this intuition.

The community bonds objection: communicative punishment presupposes the kind of community bonds — shared values, mutual recognition, relationships of citizenship — that underpin the communicative enterprise. In societies marked by deep inequality and marginalisation these bonds may not exist or may be too damaged to support penance as reconciliation. Duff acknowledges this seriously — he says punishing people who have not been treated as full members of the normative community lacks legitimacy. This is an honest concession that is also a potentially devastating critique of actually existing punishment systems.

The offender's perspective: communicative theories presuppose that offenders share the values of the community whose law they have violated. But many offenders — particularly those from marginalised communities — may have good reason to reject the community's values. Treating their punishment as secular penance for violating shared values they do not share is potentially oppressive rather than respectful.

Interim judgment on Section 3: Duff's communicative theory is philosophically richer than pure retributivism — it explains not just that punishment is deserved but what it should achieve and why. The three-part response to the liberal objection is largely persuasive. But the theory faces genuine challenges: the necessity of suffering for communication is not fully established, and the community bonds the theory requires may not exist in actually existing societies. Most importantly communicative theory and retributivism are not strict rivals — the most defensible position may combine them.

section 4 - synthesis of retributivism plus communication

Why the two theories complement rather than compete: Moore's retributivism answers the question of why we punish — because desert is sufficient justification, and we have a duty to achieve retribution. Duff's communicative theory answers the question of what punishment should look like and aim at — moral communication, censure, repentance and reconciliation. These are different questions and both need answering.

A pure retributivism that simply inflicts suffering proportionate to desert with no communicative purpose treats the offender as a moral patient rather than a moral agent — it imposes suffering without engaging their rational agency. This is morally incomplete as Duff correctly argues.

A pure communicative theory that abandons desert risks becoming a form of moral therapy that ignores what offenders actually deserve and treats wrongdoing as a problem to be addressed rather than a wrong to be answered for.

The synthesis: we punish because offenders deserve it — Moore's retributivism establishes this. But the form punishment should take is communicative penance — Duff's account establishes this. Desert grounds the justification, communication shapes the content.

What this means in practice: Punishment should express proportionate censure corresponding to the seriousness of the wrong — Moore's proportionality requirement. But it should do so in a way that engages the offender as a rational moral agent, communicates the community's judgment, and aims at repentance and reconciliation — Duff's communicative requirement. Community service orders that are calibrated to the nature of the offence and give the offender the opportunity to make moral reparation exemplify this synthesis. Pure imprisonment without communicative purpose — simply warehousing offenders — fails on Duff's account even if it satisfies Moore's desert requirement.

conc

• Consequentialism is decisively defeated by Moore's reductios — it cannot avoid punishing the innocent in conceivable cases and treats people as mere means

• Self-defence theory avoids these problems but cannot explain why punishment is permissible after the relevant threat has passed — punishment begins when self-defence has failed

• Retributivism survives Moore's thought experiments as the only pure theory that cannot be refuted — the reductio against the mixed theory leaves retributivism as the only remaining option and the duty to punish is a philosophically serious claim

• But retributivism is morally incomplete — it tells us why and how much to punish but not what punishment should achieve or why suffering is the only appropriate response

• Duff's communicative theory completes the picture — punishment should take the form of secular penance, engaging offenders as rational moral agents, communicating censure, and aiming at repentance and reconciliation

• The most defensible position combines both: desert grounds the justification of punishment, communication shapes its appropriate form. Any system of punishment that satisfies desert requirements without engaging offenders as rational moral agents is getting something morally important wrong — and any communicative system that abandons desert as the ground of punishment risks treating wrongdoing as a therapeutic problem rather than a moral failing that demands response