Skip to content

Oxford Uehiro Prize in Practical Ethics: “The Justice of Punitive Wars” written by Benjamin Koons

  • by

This essay received an Honourable Mention in the graduate category of the Oxford Uehiro Prize in Practical Ethics.

Written by University of Oxford, Oriel College student Benjamin Koons

  1. Introduction

Contemporary just war theory has largely abandoned punishment as one of the just causes for war, but I intend to show that if one accepts the justice of defensive wars then punitive wars are plausibly justified. I defend this thesis:

Punishment as Just Cause (PJC): It is a just cause for international treaty organization X to initiate a war with member-state Y so as to punish Y for an injustice against state Z.

I have limited my thesis so that only international treaty organizations (e.g. the United Nations) can punish injustices by member-states. This thesis says nothing about the form of punishment and when it should occur. David Luban wrongly equates PJC with a different thesis about jus in bello[1]:

War as Punishment: International treaty organization X may justly punish member-state Y with the destruction of war.

Another possibility is that the punishment is supposed to happen after the war’s conclusion by imposing punitive terms in a peace treaty:

Punitive Terms: International treaty organization X may justly impose punitive terms on member-state Y for crimes committed before and during the war.

An example of this sort of punishment is the Treaty of Versailles in which the victorious Allied Powers punished Germany through reparations payments and territorial concessions. In this essay defending PJC, I will speak solely about Punitive Terms and not War as Punishment.

Group Agency and Responsibility

Imagine the following case: the country Freedonia consists of three ethnic groups: Blue, Violet, and Green, which are each perfectly represented by an ephor, while a king presents the resolutions to be voted on for any judgment. Its founders framed the Freedonian constitution with the following rules for judicial proceedings:

  1. A simple majority of the ephors is sufficient for any resolution to pass.
  2. Each resolution consists of a single clause (to avoid riders).
  3. The executive must enact the ephors’ resolutions without inquiry.[2]

A case arises where a Sylvanian settlement in Freedonia (Sylvania is a neighboring country) is accused of flying a Sylvanian flag. This causes a great stir, and the king proposes the following three resolutions:

  1. The Sylvanians flew a Sylvanian flag.
  2. Flying a foreign flag is a treasonous act.
  3. Treasonous acts are to be punished by execution.

The following resolution, which represents A&B&C, won’t be voted on though:

  1. The Sylvanians are to be executed.

The three ephors vote in the following way on these three issues, and I also show how they would vote on D (A&B&C) if they could:


  A. Flew Flag? B. Treason? C. Execution? (D. Execute Sylvanians?)
Blue Yes Yes No No
Violet Yes No Yes No
Green No Yes Yes No


Each of A, B, and C pass by a two-thirds majority, and so the executive enacts the ephors’ resolutions by executing the Sylvanians. None of the ephors or their respective ethnic groups is individually responsible for this act, since none of them supported D. The ephors’ votes only cause the execution given the constitution, but the constitution itself isn’t intrinsically wicked (even if it’s simplistic). So when the Sylvanians are wrongly executed, Freedonia has done an injustice, but no individual Freedonian is clearly responsible. So if the Sylvanians seek to punish the wrong, whom will they punish? Neither each ephor and his ethnic group nor the king nor the executive nor the constitution’s framers are individually responsible, and so it seems that only one entity is liable to punishment: the whole country of Freedonia.

Group agency and responsibility are regaining currency in political and social theory, especially thanks to Philip Pettit’s work. It’s a pervasive belief that groups may act and be responsible for their actions, and our linguistic and social practice presumes the existence of group entities. Just read in a newspaper what China is up to or what the Catholic Church believes. When a teacher is away from her class and receives a substitute’s negative report about their behavior, she may punish the whole class by revoking its recess privileges. This is fair even if no individual student’s actions on their own constitute a violation of the rules, since each student contributes to a collective rowdiness.

Before delving into the social ontology, let’s be clear how robust a notion of group agency and responsibility is needed to defend PJC, which involved one state’s punishment by an international organization. PJC is distinct from the less controversial claim that states can pursue individual war criminals in other countries. PJC relies on states’ being the sorts of things that can act (unjustly) and can be held responsible for their actions. So PJC relies on a realist account of group entities, and indeed it requires the stronger claim that these real group entities can act and be held responsible for their actions. Yet PJC doesn’t require a non-reductive account of group entities. Realism about groups doesn’t commit one to believing that groups are mysterious entities that emerge from individuals and have a life of their own. Instead, one could accept a reductive account of group entities and still maintain that they are real agents responsible for their actions. This position in social ontology is analogous to reductive materialism in philosophy of mind, since in either case the reducibility of the higher level to a lower level doesn’t make the higher level any less a bearer of intentionality, agency, and responsibility. Although not all groups are agents, some groups do exhibit the sort of goals, deliberation, and intervention in their surroundings that constitute intentional agency. One can speak about the local Boy Scout troop’s goals, how it deliberates on which summer camps to attend and how to earn its merit badges, and then how it goes about camping and canoeing.

It’s one thing for a group to count as an intentional agent and another for it to be capable of being held responsible for its actions. Pettit puts forward three necessary and sufficient conditions for holding an agent responsible. First, ‘value relevance’ means that the choice faced by the agent has moral significance, and this condition relies on the agent’s being autonomous.[3] Some group agents are autonomous agents, since the judgments (and desires) of these group agents don’t supervene proposition-wise on their individual members’ judgments. I cannot go into the full argument for this underivability,[4] but if one assumes that groups may have members with any number of sets of consistent beliefs, that the group’s beliefs and other attitudes are themselves consistent and complete, and that no individual’s attitudes have a disproportionate impact on the group attitude, then it follows that no single aggregation function for any given proposition takes one from the set of individual attitudes to the group attitude towards that proposition. Freedonia can only achieve group rationality by limiting voting to the propositions A, B, and C because if the ephors could also vote on D, then the group would have an inconsistent belief: that the Sylvanians should and should not be executed. The way that the Freedonian constitution prevents this group irrationality is by privileging the individual attitudes on certain propositions (those chosen by the king) over other propositions. There is no function from the individual attitudes on proposition D through which one can derive the group attitude towards proposition D. The complicated way in which group attitudes supervene on individual ones grants the group agent some autonomy from its individual members.

The second condition on being held responsible is ‘value judgment,’ which means that the agent can make judgments about the values at play in its actions.[5] This condition is met by group agents through individuals’ providing evidence to the group and then the group’s coming to some decision about this evidence. The third condition is ‘value sensitivity,’ which consists in the agent’s having control for choosing between alternative options.[6] One concern for groups’ meeting this third condition is that it would entail that if groups and their individual members have control over the same actions, then one action is controlled by two different agents. Yet a similar problem arises in reductive materialism, since one neural event may have two complete causes (a mental and a physical one) that control it. So whatever solution one opts for to explain individual control can be used in the case of group control.

Once we see that some groups can possess agency and be held responsible for their actions, there are at least four reasons for holding groups responsible rather than just apportioning blame to individuals. First, there is the problem we saw in the Freedonia case where it may be that there is no individual (at least, qua individual) responsible for the injustice and that only the group may be held responsible. Second, if one understands group actions as consisting in the attitudes and actions of individuals bearing qua-relations to various social roles (e.g. Texas’ executing a murderer consists in the performance of a lethal injection by George Smith qua state executioner), then it makes sense to punish individuals qua member of the group. So instead of punishing individual citizens of a rogue state qua individuals, one might punish these same individuals qua citizens by fining their public treasury, seizing public goods, and dismantling public infrastructure.[7] Third, there is the practical difficulty of apportioning blame to individuals in a complicated act of group malfeasance. For example, Chicago’s police department plainly covered up damning details in a police brutality case, but it’s nearly impossible to determine who knew what when. Fourth, punishing whole groups for their actions encourages them to take responsibility for their future actions and for individuals to oppose group wrongdoing proactively.[8] So the schoolteacher punishes her whole class for rowdiness even though some students didn’t participate because she wishes to encourage these better-behaved students to stand up to their rowdy peers.

  • Groups and Justifications of Punishment

There is no problem applying any of the traditional justifications of civil punishment (retribution, deterrence, and reform/rehabilitation) to the international setting. The analogy is clearest for Punitive Terms, which take the same form as any judge’s sentence. Retribution is the easiest to apply, since it consists in giving an offender his ‘just deserts,’ which a peace treaty accomplishes by penalizing the offending country. Deterrence is similarly easy, since one state can punish another to deter other states from similar offenses. Reform may also be accomplished in a punitive peace treaty, and this comes about in two ways. The first way is completely analogous to punishing an individual: undergoing punishment can constitute an act of penance for the offender if he should accept his action’s wrongness and the punishment’s fittingness. The second way is unique to states, since a peace treaty can actually reform the offending state by deposing its leaders and imposing constitutional changes.

  •  Conclusion

This paper’s main point has been to show that if one country wrongs another and refuses to accept an international authority’s punishments, then that authority may bring the offending country to accept punitive terms through military force. My argument relied on a moderate claim about social ontology: that there are group agents even if these are reducible to individual agents and that these group agents may be held responsible for their actions. I also dissociated the notion of punitive wars from any specific account of the justifying aims of punishment indicating that it’s compatible with retribution, distribution, and reform as justifications.

[1] David Luban, ‘War as Punishment,’ Philosophy & Public Affairs 39 (2011): p. 299.

[2] For this example, it doesn’t matter whether the executive is a free agent or a robot.

[3] Philip Pettit, ‘Responsibility Incorporated,’ Ethics 117 (2007): p. 175.

[4] Christian List and Philip Pettit, ‘Aggregating Sets of Judgments: An Impossibility Result’, Economics and Philosophy 18 (2002): pp. 89-110.

[5] Pettit 2007 op. cit.

[6] Ibid.

[7] Group punishments may incidentally harm individuals qua individuals too, but many legitimate punishments have significant adverse effects on people who aren’t being punished (e.g. taxpayers must pay to maintain prisons, and the prisoners’ families lose their income).

[8] Pettit 2007 op. cit.

Share on