Proportionality and Necessity in Israel’s Invasion of Gaza, 2023–2024

1 Introduction

The aim of this essay is to demonstrate that Israel’s continuing war in Gaza following the murders by Hamas of more than 1,000 of its civilian citizens fails two of the three important conditions of a just war. It is, therefore, an unjust war.

I should, however, state at the outset what is, in any case, obvious: that Hamas’s war against Israel is also an unjust war, so that my larger claim is that the war in Gaza is an unjust war on both sides.

While I believe that the Palestinians have claims of justice against Israel that in radically different circumstances could together constitute a just cause for war, other conditions of a just war by the Palestinians have always been absent. One reason why this is so is that any discriminate use of military force by Palestinians against the IDF, rather than against Israeli civilians, would have had no chance of military success. It would have caused harm but would have had only a negligible probability of having sufficient good effects to have morally outweighed the bad. The forms of violent action of which Palestinians have been capable, such as the murders and kidnappings of civilians on October 7th, have been impermissible both because they have intentionally harmed people who were not morally liable to those harms and because they have been ineffective and indeed counterproductive for achieving the Palestinians’ just aims.

That acts of violence by Hamas have been counterproductive means that they have violated the just war requirement of necessity, which holds that acts of violence can be permissible only if they are the morally best means of achieving a just goal. While the Palestinians have all along had just goals, the means of achieving them that would have been – and still would be – morally best, because most effective and least harmful to both Israelis and Palestinians, is systematic nonviolent resistance. Israel has been the one country in the world against which tactics of nonviolent resistance would have been most likely to be successful. This is in part because of the collective history of so many of the Jews who sought refuge and safety in Israel following the Nazi Holocaust. This history has disposed them to be unusually sensitive to the claims of oppressed peoples, provided that those claims are pursued nonviolently.

But Israel’s vulnerability to nonviolent action is also explained in part by its having been, and to some extent continuing to be, dependent on financial and military support from the US. Because the PLO and Hamas have repeatedly engaged in acts of terrorism, Americans have tended to think of Palestinians as terrorists rather than victims of injustice. Well-publicized acts of collective nonviolent resistance would almost certainly have elicited the sympathies of Americans, rather than alienating them, as Palestinian acts of terrorism have done. I believe it highly likely that, if the Palestinians had been led by a Palestinian equivalent of Mohandas Gandhi rather than by Yasser Arafat, they would have had their own state many decades ago.

2 The Requirements of Jus ad Bellum

According to the traditional theory of the just war, there are six conditions of jus ad bellum – that is, six requirements that must be satisfied for a war as a whole to be morally justified. These conditions are traditionally referred to as just cause, necessity, proportionality, reasonable hope of success, right intention, and legitimate authority.

I will not consider the last three of these in their application to Israel’s Gaza war. That a war must have a reasonable probability of success is, I believe, subsumed by proportionality, which must take all relevant probabilities into account.

Right intention – the requirement that a war must be intended to achieve the just cause – and, perhaps, not be intended to achieve other aims – is problematic in various ways. One problem is that it presupposes some or other conception of a collective intention, which must be a compound of normally varying and conflicting individual intentions and therefore cannot have the same significance for the permissibility of action that many believe individual intentions can have. Another problem is that right intention may not actually be a moral requirement. If a war would actually be a necessary and proportionate means of achieving a just cause, it would be permissible for soldiers to fight in it even if the political leaders had only self-interested intentions in ordering them to fight.

Finally, legitimate authority, which requires that war be initiated only by those who have the appropriate moral or legal authority to do so, may also not be an actual moral requirement; and in any case it is clearly satisfied in Israel’s Gaza war.

That a war must have a just cause seems obviously correct, though exactly what constitutes a just cause is a matter of dispute. I think there is a clear sense in which any just or good aim that could be achieved only by means of a war, the good effects of which would morally outweigh the bad, can be a just cause for war. But, while this may be a sufficient condition of a just cause, it is not a necessary condition. The concept of a just cause allows that there can be a just cause that it would be either unnecessary or disproportionate to pursue by means of war.

One common view is that only defense – self-defense by a state or people, or defense by a state of another state or people – can be a just cause for war. Although I think this view is overly restrictive, I will not challenge it here; for it is clear that Israel has a right of defense against the murder, maiming, and kidnapping of its citizens by Hamas, and thus in principle has a just cause for war against Hamas. What I will dispute is the claim that Israel’s Gaza war (to which I will sometimes refer as ‘the Gaza war’) satisfies the requirements of proportionality and necessity. In developing my arguments, I will address some theoretical matters concerning proportionality and necessity that have been insufficiently discussed in the philosophical literature on the ethics of war.

3 The Requirement of Proportionality

Proportionality in the resort to or continuation of war is the requirement that the bad effects of the war not be excessive in relation to the war’s good effects. There is a parallel principle governing the conduct of war (jus in bello) that requires that the bad effects of each individual act of war not be excessive in relation to its good effects. These simple claims obscure a range of complicated matters inherent in the notion of proportionality, such as whether some good or bad effects are excluded, whether some effects that do count are nevertheless discounted in certain conditions, how the assessment of proportionality takes account of empirical and moral uncertainty, whether the weight that an effect has in the assessment of proportionality is affected by whether it comes about through an act or an omission, or whether the effect is intended or merely foreseen, and so on.

Just as I am skeptical about whether there really is a requirement of right intention, some just war theorists are skeptical about whether there is a requirement of proportionality. In an article about the Gaza war, Michael Walzer argues that “proportionality is a fool’s game, more suited to propaganda than to reasoned judgement” (Walzer 2023). “Proportionality arguments,” he contends, are now “being used to condemn whatever killing the IDF is doing in the Gaza war. In these condemnations, it seems clear that ‘disproportionate’ just means any number that horrifies me.”

Walzer has two objections to proportionality. One is epistemic: “The numbers are always disputed, and so is the balance of fighters to civilians.” The other is that there is no one answer to “the critical question: proportional to what?”

To the military value of this Hamas communication center or that rocket-launching site or weapons cache? To the contribution destroying such a target would make to an overall victory? To securing justice after the pogrom of October 7th? To the safety of the citizens of Israel? To the deterrence of future pogroms? If I were speaking for the Israeli government, I think I could argue, whatever the numbers, that they are not disproportionate to one or another of these ends. But I don’t believe that arguments of this sort are morally serious. They are too easy to make. (Walzer 2023)

Given the context, the phrase ‘whatever the numbers’ must refer to the number of Palestinian civilians killed or injured. Unless Walzer is simply saying that there is no number of Palestinian casualties that horrifies him, he is claiming that there is no number that it would be disproportionate for Israel to cause as a side effect of achieving the aims he cites. Although he says that the argument for this is easy to make, I will argue that it is impossible to make.

The requirement that Walzer believes is morally serious is that Israel must not ‘intend to kill civilians. But morality requires more than that: Israel must intend not to kill civilians. It must take positive measures to minimize the deaths it causes.’ Provided that, as Walzer believes, Israel does not intend to harm civilians and makes reasonable efforts to avoid harming them, it is doing all that it is morally required to do as it pursues its just cause.

This is false. Consider what Walzer would call a ‘domestic analogy.’ Suppose that one innocent person’s only means of preventing another innocent person from being murdered will almost certainly kill 100 (or 500, or 1,000…) innocent bystanders as a side effect. It might be, for example, that the murderer happens to be in the middle of a large group of people when he intends to shoot his victim, and that the only weapon the third party can use to prevent the murder is a powerful grenade. Suppose further that there was only one precaution the third party could take, at some personal risk, to try to reduce the number of innocent bystanders her defensive act would kill as a side effect, and that she took this precaution. Still, despite her best efforts, her use of the grenade to save the one person kills all of the 100 bystanders.

This person’s defensive act was impermissible. But she did not intend the deaths of the bystanders and did all she could to avoid killing them. The only explanation of why her act was impermissible is that it was disproportionate. This is true, I believe, even if the murderer had embedded himself in the crowd of innocent people with the intention of using them as innocent shields.

Walzer thinks that Hamas’s embedding its members among the civilian population shows that it uses Gazan civilians as innocent shields, and this, he thinks, shifts the responsibility for their deaths from Israel to Hamas. Hamas may well try to use civilians as shields. If so, this strategy does not work. Also, since Hamas members cannot leave Gaza, which is densely populated with civilians, one wonders where else can they embed themselves. Finally, even if they do use civilians as shields, this cannot relieve Israel of responsibility for killing the shields. If a man who is intent on murdering me holds an innocent person in front of him and I shoot through that person to kill the murderer, that is no more justifiable, and may indeed be even less justifiable, than killing an innocent bystander as a side effect of defending myself would be.

We can imagine a similar example in the Gaza war. Suppose that a neutral third-party intelligence officer stationed on the border between Israel and Egypt learns that an IDF unit in Israel intends to launch a weapon that would strike a hospital in Gaza in order to kill one Hamas militant who is sheltering there. But the strike would also predictably kill perhaps as many as 40 patients and doctors as a side effect, which the officer believes would be disproportionate. He has a weapon that he intends to use to destroy the IDF weapon preemptively. He does not realize, however, that his doing so would also kill well over 500 Israeli civilians. A neutral observer can prevent the officer from using his weapon only by killing him. I suspect that in this case Walzer would happily play the fool’s game by acknowledging the permissibility of killing the officer to prevent him from killing a disproportionate number of Israelis.

One might expect Walzer to argue that, because Israel’s war is just, the intervening officer would, by using his weapon, become a combatant in an unjust war, and that would explain why his act should be prevented. But Walzer has denied himself this option because he argues that combatants fighting in an unjust war do nothing wrong provided that they obey the moral rules governing the conduct of war. If the conduct of war is not governed by a requirement of proportionality, Walzer’s principles seem to provide no basis for condemning the officer’s act.

4 Civilian Liability

My concern in this essay, as is usual in discussions of proportionality in war, is only with harms inflicted on people who are innocent in the relevant sense – that is, people who have not through their own action made themselves either deserving of harm or morally liable to be harmed. Proportionality in the infliction of harm on a person who is not liable to that harm is often referred to as ‘wide proportionality.’ It contrasts with ‘narrow proportionality,’ which is proportionality in the infliction of harms on people who are morally liable to some degree of harm.

I will also generally assume here that all civilians are innocent in this sense. I actually believe that this is false and that some civilians may be liable to small harms and that a small number of others may be liable to serious harms. But my assumption will be that there is a general and quite strong constraint against harming civilians in war.

I have, however, heard a number of defenders of the Gaza war make claims similar to that made by an Israeli academic with whom I had an exchange in a podcast organized by the Jerusalem Post. During that discussion, she remarked, “I’m not so sure that we can call most of them [Gazans] innocent people.”^[1] The view that most Gazan civilians are not innocent people and are thus liable to attack is usually based on the claim that many people currently in Gaza voted for Hamas in 2006 and those who did not nevertheless continue to accept Hamas’s leadership.

As a general view, however, this is implausible. It implies that civilians of all nations, including Israel, who might have some influence over their leaders are morally liable to be subjected to attack, at least as side effect, if their nation fights an unjust war. A less extreme view is that, even though civilians of a nation that is fighting an unjust war are not liable to military attack, they are liable to lesser harms because they share responsibility for their leaders’ action. On this view, most Gazan civilians are liable to some proportion of the harm they might suffer as a side effect, thus making it more likely that the total harm they might suffer is justified, since the remainder that must be justified in some other way is a lesser harm. In short, the harms that Israel has inflicted on Gazan civilians as a side effect are more likely to be judged proportionate by a form of justification that takes account of both civilian liability and certain other considerations that I will discuss in the next section.^[2]

The appeal to civilian responsibility and liability is, however, unlikely to add significantly to any overall justification for the harms that Israel has inflicted on Gazan civilians. One reason for this is that 47.3 percent of Gazan civilians are children.^[3] Another is that ordinary Gazans have far less influence over what Hamas does than many other populations have over their governments. It is, of course, true that some adult Gazans voted for Hamas in 2006, many because of their disenchantment with the corruption of the Palestinian Authority. But there have been no elections since then. Moreover, given the conditions that Israel has imposed on Gaza through the blockade, it was virtually inevitable that the most brutal members of Hamas would achieve the greatest power. Thus, any Gazan man, and certainly any Gazan woman, who might challenge or defy Hamas would do so at his or her own peril.

The Israeli invasions of Gaza in 2008, 2012, 2014, and 2021 had taught Gazan civilians that if Hamas kills Israelis, Israel will then kill far more of them and their children in retaliation. It is hard to believe that many Gazans have welcomed these recurring conflicts, which have achieved nothing for them other than destruction. But, while it has been in their interest at each point to prevent Hamas from provoking yet another series of bombardments, they have obviously been unable to do that.

This reasoning is reinforced by polls taken by Arab Barometer in Gaza between September 28 and October 6, 2023. These revealed that, when asked how much “trust they had in the Hamas authorities, a plurality of respondents (44 percent) said they had no trust at all; ‘not a lot of trust’ was the second most common at 23 percent.” That is, just before the war began, more than two-thirds of Gazans said they had little or no trust in Hamas. Furthermore, “when asked what is the most effective way for ordinary people to influence the government, a plurality said ‘nothing is effective’.” Finally, “the majority of survey respondents favored a two-state solution with an independent Palestine and Israel existing side by side” (Jamal and Robbins 2023). The percentages have of course changed since the Israeli invasion, but that only adds to the evidence that the war has been counterproductive for Israel.

These percentages can be instructively compared with one result of a poll taken in Israel in February of 2024: namely, that “68 percent of Israeli Jews oppose[d] allowing food and medicine into Gaza,” knowing that those necessities could be provided to Gazan civilians only through Israel (Kristof 2024). Given (1) that Israeli citizens have far more control over their government than Gazans have over Hamas, (2) that the great majority of them have supported the war, and (3) that the Israeli government has indeed intentionally impeded the provision of food, medicine, and other necessities to the population of Gaza, those who hold Gazan civilians liable to harms as side effects of the IDF’s defense of Israelis will find it difficult to avoid the conclusion that Israeli civilians would be morally liable to even greater harms as side effects of action necessary to prevent Gazan civilians from being starved or killed through the deprivation of food or life-saving medicines – or indeed from being killed or wounded by disproportionate acts of war by the IDF.

5 The Nature of Proportionality in War

There are various facts about wide proportionality that I should make explicit before addressing questions about individual Israeli acts of war and about Israel’s Gaza war as a whole. One of these is that proportionality in war is primarily a constraint on the harms that combatants cause. Only rarely in war is there a proportionality restriction on harms that combatants allow to occur. This is because there is a constraint against harming innocent people but no corresponding general constraint against allowing people to suffer harm. This is in part an implication of the common view that the reason not to cause a certain harm to an innocent victim is stronger, other things being equal, than the reason to prevent an equivalent harm for which one would not be responsible.

Second, proportionality in war is prospective – that is, it weighs the harms that an act may cause against the harms it may prevent, or against benefits it may produce, taking probabilities into account. If a war or act of war were purely retributive, or were a matter only of reprisal, then proportionality would be retrospective – that is, it would weigh the harms a belligerent would cause against harms it had suffered in the past. But I will assume, as I believe to be true, that neither war nor any individual act of war can be justified as either retribution or a reprisal. (To say that proportionality is prospective is not to deny that what wrongdoers have done in the past provides evidence of what they may do in the future if not prevented.)

Third, wide proportionality in war, and in defensive harming generally, sensitive to intention. This is one point on which I agree in substance with Walzer. He believes that harms to innocent people that are intended as a means are more seriously morally objectionable than equivalent harms that are unintended but foreseen, or foreseeable, side effects. I would articulate this point by saying that the former harms have greater weight in the assessment of wide proportionality than the latter.

Fourth, because wide proportionality is concerned only with harms to people who are innocent in the relevant sense, the moral justification for harming them must be entirely circumstantial, a matter of consequences alone. The standard form of moral justification for the infliction of harm on an innocent person is a lesser-evil justification, according to which the harm is justified if it is necessary to prevent other innocent people from suffering substantially greater harm. It is because there is a constraint against harming that the harm prevented must be significantly greater than that caused. This is the form of justification that applies in the familiar Trolley case. The reason one is justified in diverting a runaway trolley from a track on which it will kill five innocent people to a track on which it will kill only one is that killing the one as a side effect is the lesser evil in the circumstances.

There is also what can be called a greater-good justification, according to which harming an innocent person can be justified if it is necessary for conferring very substantially greater benefits on other innocent people. This form of justification is more demanding than a lesser-evil justification because, again, the reason not to inflict a harm on an innocent person is stronger than the reason to confer an equivalent benefit on an innocent person.

It may be that greater-good justifications are more important to the assessment of proportionality in war than has hitherto been supposed. This is because it has been assumed that the prevention of a killing is the prevention of a harm. But, apart from any suffering that the victim might experience, the death of the victim of a killing is not intrinsically bad for that person. According to the most widely-accepted account of the badness of death – the ‘Deprivation Account’ – death is bad for a person, or at least bad primarily, because it prevents that person from having the benefits of continued life. Preventing a person from being painlessly killed does not, therefore, prevent her from suffering an intrinsic harm.

There are, of course, no precise criteria for distinguishing between intrinsic harms and intrinsic benefits. And ordinary language allows that one can ‘harm’ a person either by causing something intrinsically bad to happen to her or by preventing her from having something that would be intrinsically good for her. What is important, however, is not how the word is used but that the moral reason not to cause an innocent person to suffer an intrinsic harm is stronger than the reason not to prevent a person from having an intrinsic benefit of equivalent magnitude. This must be true if the widely accepted moral asymmetry between harming and benefiting is not reducible to the moral asymmetry between doing and allowing.

But if to save a person’s life is to confer an intrinsic benefit rather than to prevent an intrinsic harm, then the justification for causing innocent people to suffer intrinsic harms as a side effect of saving other innocent people’s lives must be a greater-good justification, which, as I noted, is even more demanding than a lesser-evil justification.

If correct, this is a surprising and important point; but this is not the place to explore it in detail. We can largely avoid it here by comparing only the number of innocent people who are likely to be saved by some act or sequence of acts and the number of innocent people who are likely to be killed as a side effect. We thus compare the benefits to those saved with the benefits of which those killed are deprived. The form of greater-good justification will therefore be different. Depriving innocent people of benefits by killing them must be justified by the conferral of greater benefits on other innocent people by saving them. We can then suppose – not implausibly, but again for the purpose of argument – that the necessary ratio of benefits conferred to benefits prevented is the same as the ratio of harms prevented to harms inflicted that is necessary for a lesser-evil justification. Indeed, since this assumption seems plausible, and also because virtually everyone – whether correctly or not – regards even painless killing as an especially serious harm, I will proceed as if the primary justification for killing innocent people as a side effect of saving a greater number of innocent people is a straightforward lesser-evil justification.

6 Permissible Partiality

Given this assumption, the four facts about proportionality I have cited together suggest that wide proportionality in war is simply the limit to the harm to innocent people that can be justified as the lesser evil when it is caused as a side effect of the pursuit of a just cause by means of war or an act of war. But what is the ratio of harm prevented to harm caused that is necessary for there to be a lesser-evil justification? One piece of intuitive evidence is found in the Trolley case. The philosophers who devised and have discussed this case presumably chose a ratio of five people killed to one person saved because, acknowledging the constraint against harming, they wanted an example in which killing as a side effect would be just barely proportionate and therefore just barely permissible. If so, they must have thought that a ratio of four to one would be intuitively disproportionate, while a ratio of six to one would be fairly clearly proportionate.

But is five saved to one killed the correct ratio for justifying the killing of innocent people as a side effect in war? Is it, for example, proportionate and therefore permissible for Israel to kill one innocent Gazan civilian as side effect of an act of war only if that act can be expected to save the lives of at least five Israeli civilians? Neither the members of the Israeli government nor the senior officers of the IDF seem to believe any such thing. They seem to believe, on the contrary, that preventing the killing of a single Israeli civilian can justify killing a substantial number of Gazan civilians as a side effect.

According to two Israeli intelligence officers, the IDF “decided during the first weeks of the war that, for every junior Hamas operative that” Israel sought to kill, “it was permissible to kill up to 15 or 20 civilians.” If “the target was a senior Hamas official with the rank of battalion or brigade commander, the army on several occasions authorized the killing of more than 100 civilians” as a side effect of the attempt to kill him (Abraham 2024). For the killing of 20 Gazan civilians to be justified as the lesser evil, one must assume that any junior Hamas operative would, if not killed, have killed more than 100 Israeli civilians. Since that is a wholly implausible assumption to make, any attempt to defend these standards of proportionality in the killing of innocent people must appeal to a form of justification that is more permissive than a lesser-evil justification.

The Israelis’ assumption seems to be that combatants are justified in giving substantial priority to the lives of their fellow citizens over the lives of ‘enemy’ civilians in the assessment of proportionality. Some philosophers have sought to explain the permissibility of partiality more generally by appealing to what is variously known as an ‘agent-centered prerogative’ or an ‘agent-relative permission.’ (Scheffler 1982) This is a moral permission to give some degree of priority to one’s own interests or concerns – that is, to give them greater weight than they have impersonally, or greater weight than one is required to give to the equivalent interests or concerns of others. It is hard to deny that there is such a permission in the case of choices that are not governed by a moral constraint, such as a choice between saving one’s own life and saving the lives of two strangers. But I doubt that there is an agent-relative permission that overrides the constraint against harming. Suppose that the only way that one can prevent oneself from being killed by a culpable attacker will not only kill the attacker but also kill an innocent bystander as a side effect. I believe that it would be impermissible, because disproportionate, to kill the innocent bystander – though one would presumably be largely excused, or not blameworthy, if one were to do so (for opposing views, see Davis 1984; Quong 2020). One has no more moral reason to give priority to oneself than a stranger would have. Indeed, the reason that a stranger would have to defend one from the unjust threatener is a prima facie moral duty, whereas one’s own reason would, apart from any reason concerned with the effects of one’s death on others, be prudential rather than moral.

There is, however, a related form of justification, grounded in relations that people bear to one another, that I think can override the constraint against harming an innocent person even in the absence of a lesser-evil justification. Suppose that one’s only means of preventing one’s child, either biological or adopted, from being killed by a culpable attacker will unavoidably kill an innocent bystander as a side effect. In this case, one has a moral reason to save one’s child that is much stronger than both one’s prudential reason to save oneself and one’s moral reason to save a stranger. If one is permitted to kill an innocent bystander as a side effect of saving one’s own child, it is not because one has an agent-relative permission to give priority to those about whom one happens to care more. It is because of the objective moral significance of the special relation one bears to one’s child, and the duties grounded in that relation.

Whether it is permissible to kill an innocent stranger as a side effect of saving one’s own child is controversial. I think it may well be, but I am not confident about this judgment. And I doubt whether it could be permissible to kill two children as a side effect of preventing one’s own child from being killed. Suppose, however, for the sake of argument, that the parent-child relation is sufficiently morally important to override the constraint against harming when it is necessary to kill two other children as a side effect of preventing one’s own child from being wrongly killed. Call this form of permission a special relations justification. It is far more permissive than a lesser-evil justification. Whereas a lesser-evil justification may permit one to kill one innocent person as a side effect of saving no fewer than five innocent people, the special relations justification in this case permits one to kill two innocent people as a side effect of saving only one. In this case, then, wide proportionality is the limit to what can be justified on the basis of a special relations justification.

It is not obvious, however, how a special relations justification might apply to the action of combatants in war. The amount of harm to innocent people that a special relation between an agent and those he might defend can justify, beyond what could be justified by a lesser-evil justification, is variable. It depends on the nature of the special relation, some of which are morally far more important than others. The parent-child relation is, perhaps, the paradigmatically morally important special relation. It is capable of justifying more harm to innocent people than, for example, the relation between a person and her neighbor, other considerations being equal. Similarly, the relation of co-nationality between a combatant and a civilian whom the combatant does not know personally is not as significant morally as the relation between a parent and child. Thus, even if a parent has a special relations justification for killing two children as a side effect of saving her own, it does not seem that a combatant could similarly have a special relations justification for killing two ‘enemy’ civilians as a side effect of saving a single fellow citizen who is unknown to the combatant.

There is, moreover, a further question that I cannot address here, which is whether the special relations justification that applies to combatants fighting in a just war applies equally, or at all, to combatants fighting in an unjust war. But this question need not be addressed here, as it may not arise when a war is unjust precisely because the harms inflicted on enemy civilians exceed those that could be justified by a special relations justification if the war were otherwise just.

One might argue here that an Israeli combatant is an agent of the Israeli state, which has a duty to defend its citizens. The special relation in this case is thus not co-nationality only but is also that of a professional defender and a person whom the defender has a duty to defend. But it remains unclear how much this sort of special relation can justify. A paid bodyguard also has a duty to defend his employer against unjust threats but would not be justified in killing two innocent bystanders – or even one innocent bystander – as a side effect of saving his employer’s life.

One might nevertheless argue that the duties of states are different, and stronger. To test this supposition, suppose that an American soldier stationed in Germany during the occupation following the Second World War were to see a German criminal who was about to kill an American girl on the street. The only way the soldier could prevent the killing would be to kill the criminal, but doing so would unavoidably kill two German girls as a side effect. I cannot believe that the soldier’s professional duty to defend Americans, together with his relation of co-nationality with the American girl, could be sufficient to justify his killing two as a side effect of saving one.

So far, then, I find little basis for the assumption that Israeli combatants have a special relations justification that makes it permissible for them to kill two Gazan civilians as a side effect of saving one Israeli civilian. I will nevertheless assume, for the purpose of argument, that they have a special relations justification for killing five Gazan civilians as a side effect of saving one Israeli civilian. This assumption is far more permissive than a lesser-evil justification for killing an innocent person, according to which one must save at least five innocent people to justify killing one. It is also, as I have indicated, considerably more permissive than what could reasonably be justified by appeal to the special relations between combatants and their fellow civilian citizens.

7 Proportionality in the Gaza War

8 Necessity

Not only has Israel’s war been highly disproportionate in the wide sense; it has also been unnecessary in the just war sense, in that there were morally far better means of ensuring Israel’s security, the most important of which are listed in the previous section. A strong case can be made that the war has been not only less effective than those listed means but also counterproductive, even with regard to the immediate aim of preventing more Israelis from being killed by Hamas. According to the Israeli authorities, 730 IDF soldiers and 67 Israeli police officers have been killed in the war as of early October 2024.^[8] Also, given the amount of physical devastation in Gaza, it would be surprising if there were no Israeli hostages buried beneath the rubble along with countless Gazan civilians. Other sources indicate that the total number of Israeli civilians killed in the war up to October 4th is 915.^[9] It is highly doubtful that Hamas could have killed so many Israelis had Israel implemented even the first few of the defensive measures in the list above rather than invading Gaza. If that is right, the war has resulted in the deaths of more Israelis – none of whom would have been combatants had the invasion not occurred – than would have been killed by Hamas had those defensive measures been implemented without the invasion.

Another respect in which the war has been counterproductive is that it has vastly exacerbated the grievances that Palestinians have against Israel, as well as in many cases the hatred they feel for Israelis. Even if ‘only’ 24,000 Palestinian civilians have been killed by Israel, and even if fewer than 16,000 of those have been children, and even if only 54,000 adult civilians and children have been wounded, there are vastly more who are the bereaved or devastated parents, children, spouses, siblings, and friends of those victims. It would be noble of them if these people – probably well over a million – were not motivated to seek retribution; but few of us are that noble. It is likely that for every Hamas militant Israel kills, many more young men inflamed by hatred will be recruited to Hamas or to whatever organization might emerge in its place.

The alternative measures would have been much less costly to Israel in economic and diplomatic or reputational terms as well as. And it is not as if these alternatives were morally optional. The first five in the list above were duties that the state owed to its own citizens – duties the government violated because of the complacent belief that Israel could continue indefinitely to suffer small attacks by Hamas that would kill a few civilians, to which it would respond with far greater violence, thereby keeping Hamas at bay for some further years until the next small attack occurred. The last few measures on the list above are duties that Israel has owed to the Palestinians.

I cannot, of course, prove any of the foregoing counterfactual claims. They are necessarily speculative. But if I am right that the fulfilment of the state of Israel’s duties to its citizens and to others would have been more effective in protecting the lives of Israelis than a war that has killed and wounded tens of thousands of Gazan civilians, it is close to a necessary truth that Israel’s fulfilment of its duties would have been a morally better means of achieving its just cause than the war has been.

One final comment about the requirement of necessity. Like the requirement of proportionality, necessity is, in its normal application, prospective. But, in an important article, the Israeli philosopher Daniel Schwartz has shown that necessity has a neglected retrospective dimension as well (Schwartz 2020). Suppose, he argues, that an innocent person is threatened by a murderer. This person can safely flee and have the murderer arrested. But she instead intentionally eliminates the option of retreating from the confrontation. She now has only two options: kill the murderer or be killed by him. Killing was unnecessary but she has wrongly made it now necessary. This complicates the question whether it remains permissible for her to kill the murderer in self-defense. More importantly for our purposes, it surely diminishes the amount of harm that it would otherwise have been permissible for her to cause to innocent bystanders as a side effect of her own self-defense. Since I believe that the threat to Israel from Hamas could have been largely prevented if, in the past, Israel had fulfilled its duties of justice to the Palestinians, I also believe that the retrospective dimension of the requirement of necessity is highly important in understanding the morality of the Gaza war. But this is too complicated an issue to explore here.^[10]