The Mercy Budget

Every Sunday, in a great many American churches, a few hundred people say a sentence out loud that they would not survive if anyone enforced it. Forgive us our debts, as we forgive our debtors. It is in the most famous prayer in the language. It has a built-in audit clause, which is the part nobody reads: your forgiveness comes with the exact terms you use on other people.¹

I raise this because I want to ask a question that is simple to say and annoying to answer. A country that likes to describe itself in the vocabulary of mercy and redemption has spent three years backing a war, and then a ceasefire that still kills, and the war has killed, by the Gaza Ministry of Health’s current count, 74,645 people, in response to an attack that killed about 1,200. Is that response proportionate? And if it is not, why does the forgiving nation find it so easy to forgive this?

I am going to try to answer that fairly, which is harder than answering it loudly. The loud version is available everywhere and requires no reading. This version requires the Summa Theologiae, an encyclopedia entry on the ethics of war, a Lieber Institute essay that argues against my side, three sets of death counts that disagree with each other, and the patience to say “I don’t know” in several places where the internet prefers to say “obviously.”

Three kinds of sentence

Before the argument, a house rule, because this subject is where most writers stop distinguishing between kinds of sentences, and the whole piece depends on keeping them apart.

An empirical claim says what happened: how many died, when, according to whom. Those I will source, date and, where they are contested, label as contested.

A moral premise says what matters: that killing the innocent is a grave wrong, that mercy is a virtue, that a response can be too large for its cause. Those I will state as premises, so you can reject them without having to find the error in my footnotes.

Rhetorical force is the third thing: the adjective, the dry joke, the sentence designed to make you wince. I will use some, because I am who I am. When a paragraph is mostly rhetoric I will tell you, so you can discount it accordingly.

And a boundary, stated once so I do not have to keep restating it. This essay criticizes governments, states and policies. It does not say anything about Israelis, Palestinians, Jews, Muslims or Christians as peoples or faiths, and nothing in it should be read that way. Israelis were murdered in their homes, and taken out of them, on October 7. Palestinians in Gaza are being killed, displaced and starved under a policy they did not choose and cannot vote against. Both of those sentences are true at once, and an essay that cannot hold both is not worth the electricity.

What happened on October 7

Empirical. On October 7, 2023, Hamas-led militants broke through the fence around Gaza at more than a hundred points, attacked villages, kibbutzim and military bases, and killed about 1,200 people, most of them civilians. They took 251 people hostage into Gaza.² The Washington Post, the Associated Press and Reuters give the same figures: about 1,200 killed, 251 taken.³ The Israeli government’s own pages use the same numbers, and I note that they are the Israeli government’s, because it is the party with the most to gain from a high figure and the figure has held up regardless.

Of the 251, some were released in a late-2023 deal, some were rescued, some were killed in captivity. The last twenty living hostages were released on October 13, 2025, as part of the ceasefire. The remains of the final deceased hostage, a 24-year-old police officer named Ran Gvili, were recovered from a cemetery in northern Gaza on January 26, 2026, and the Red Cross confirmed the completion of the returns.⁴ For the first time since 2023 there is no one still held in Gaza. That matters for the argument in two ways, which I will come to.

Moral premise. What was done on October 7 was an atrocity. I am not using the word as a throat-clearing ritual before the criticism. Civilians were killed in their homes, children were taken, and any ethics that cannot say so cleanly has already failed. The International Criminal Court agreed, in its way: in November 2024 its judges found reasonable grounds to believe that Mohammed Deif, the head of Hamas’s military wing, was responsible for crimes against humanity of murder, extermination, torture and sexual violence, and for the war crime of taking hostages, among others.⁵ The Prosecutor also withdrew his applications for warrants for Yahya Sinwar and Ismail Haniyeh after evidence confirmed their deaths.⁶ No one in this story is being excused. The question is what the response is owed to, and what it is permitted to cost.

What the Christian tradition says forgiveness is

Here the American self-image needs some attention, because it is the whole premise of the title.

About 62 percent of American adults identify as Christian, according to Pew’s 2023-24 Religious Landscape Study.⁷ Political speech in this country leans on Christian vocabulary constantly: redemption, second chances, grace, the prodigal son welcomed home. I take no position on the metaphysics. I only point out that the vocabulary is in public use and so is subject to public audit.

The text itself is not subtle. Peter asks how often a brother who sins should be forgiven and offers “till seven times” as a generous figure. The answer is “Until seventy times seven.”⁸ From the cross comes the line that has outlived most empires: “Father, forgive them; for they know not what they do.”⁹ The Sermon on the Mount goes further and tells the listener to love enemies and pray for persecutors.¹⁰ Paul, writing to the Romans, adds the line that I think does the most damage to casual revenge: “Vengeance is mine; I will repay, saith the Lord,” and then, as though anticipating what comes next in the news cycle, “if thine enemy hunger, feed him; if he thirst, give him drink.”¹¹

Now the strongest objection from inside the tradition, which I owe a fair hearing. That tradition has never held that forgiveness abolishes justice or that a ruler must stand idle while his people are murdered. The same Paul who says vengeance belongs to God also says the ruler “beareth not the sword in vain.”¹² Augustine, whom nobody would accuse of sentimentality, argued that a just war could be an act of love toward the wrongdoer: sometimes “it is good for him to be vanquished,” since “nothing is more hopeless than the happiness of sinners.”¹³ Aquinas, assembling the tradition into the form it still has, gave a war three conditions: legitimate authority, a just cause, and a right intention.¹⁴ Forgiveness, on this reading, is something persons owe one another. States owe something different: protection of the governed, and justice.

I accept that distinction. It is a good one. The mistake would be to take it for a loophole, because the same tradition that permits states to fight puts a leash on how. Which brings me to the part of the tradition that tends to get left out of speeches.

The leash: right intention and proportion

The tradition’s own language. Aquinas’s third condition, right intention, is not a formality. He quotes Augustine on what is “rightly condemned in war”: “The passion for inflicting harm, the cruel thirst for vengeance, an unpacific and relentless spirit, the fever of revolt, the lust of power, and such like things.”¹⁵ Read that list again with the news on. It describes with some accuracy what people mean when they say a war has become about revenge. Augustine’s remedy is in the next line Aquinas quotes: “Be peaceful, therefore, in warring, so that you may vanquish those whom you war against, and bring them to the prosperity of peace.”¹⁶ A war whose aim is the prosperity of the peace afterward is a war that cannot, by definition, leave a population starving in tents.

Proportion appears in Aquinas in the section on self-defense, which concerns private persons and public officers, so the extension to war is by analogy, an analogy the tradition itself has always made. A lawful act “proceeding from a good intention” may nonetheless be “rendered unlawful, if it be out of proportion to the end,” and violence beyond what is necessary makes a defense unlawful.¹⁷ The modern Catholic Catechism, which I cite because it is the most compact statement of the tradition in current use, lists among its conditions for legitimate defense that “the use of arms must not produce evils and disorders graver than the evil to be eliminated,” and adds that “the power of modern means of destruction weighs very heavily in evaluating this condition.” It also says that “the mere fact that war has regrettably broken out does not mean that everything becomes licit between the warring parties,” and that “the extermination of a people, nation, or ethnic minority must be condemned as a mortal sin.”¹⁸ I quote that last line only for completeness. Whether it applies to anything in Gaza is exactly the question the courts have not decided, and I am not smuggling a verdict in through a catechism.

Modern just war theory. The secular descendants say the same thing in different clothes. The Stanford Encyclopedia of Philosophy’s treatment distinguishes proportionality ad bellum, whether the war as a whole does more good than harm, from proportionality in bello, whether an individual attack’s foreseeable harm to non-combatants is excessive relative to its military advantage. It adds a third principle, necessity, which asks whether less harmful means were available.¹⁹ Michael Walzer, whose Just and Unjust Wars is the standard secular text, built the modern account on these distinctions, with one famous asymmetry: combatants may be targeted, civilians may not, and even the foreseen deaths of civilians must be justified by something real.²⁰

The law. International humanitarian law took the same idea and gave it a sentence. Article 51(5)(b) of Additional Protocol I, which the International Committee of the Red Cross treats as customary law binding even states that have not ratified it, prohibits an attack “which may be expected to cause incidental loss of civilian life, injury to civilians, damage to civilian objects, or a combination thereof, which would be excessive in relation to the concrete and direct military advantage anticipated.”²¹ Notice what the sentence does not say. It does not set a ratio. There is no number at which it becomes illegal. It requires a comparison, strike by strike, between expected harm and anticipated advantage, and it requires “all feasible precautions” to minimize civilian harm.²²

It does not set a ratio. There is no number at which it becomes illegal. It requires a comparison, strike by strike, between expected harm and anticipated advantage.

That last point is the one my side tends to get wrong, and I want to get it right, because the strongest critique of my argument is built from it. We will get there.

The numbers

Empirical. I will give the ministry’s figures first, then the doubts, then the independent checks, because the order matters.

On November 10, 2026, the Gaza Ministry of Health reported 74,645 people killed and 175,978 injured since October 7, 2023.²³ A day earlier, its figure was 74,643.²⁴ Of the total, the ministry counts 1,614 killed since the ceasefire took effect, a date given as October 10, 2025 by CNN and AP and as October 11, 2025 by the ministry’s own tally.²⁵ I note the one-day discrepancy because it is the sort of thing a careful reader will check and a careless one will use to wave the whole count away. For the arc, the same ministry’s figures reported by Reuters and AP were “some 67,000” on the war’s second anniversary in October 2025,²⁶ and “more than 71,660” by February 2026.²⁷ All of these come from one source, a ministry of the government that Hamas ran in Gaza, and that is the first thing a skeptic should say. So I will say it.

Is the ministry’s count reliable? Three lines of evidence suggest it is broadly reliable and that, if anything, it undercounts.

First, a household survey of about 2,000 families, published in The Lancet Global Health in February 2026 and fielded by Palestinian pollsters, estimated that more than 75,000 people were violently killed between October 7, 2023 and January 5, 2025, roughly 25,000 more than the ministry had announced by then. It also found that the ministry’s reports of the demographic breakdown were accurate: women, children and the elderly accounted for 42,200 deaths, or 56 percent of the violent total.²⁸ The authors warned that the figures carry significant margins of error.²⁹

Second, in November 2025 a team at the Max Planck Institute for Demographic Research estimated 78,318 deaths for almost the same window.³⁰ Third, the Guardian reported that a senior Israeli security officer told Israeli journalists in January 2026 that about 70,000 Palestinians had been killed by Israeli attacks, excluding the missing, which the paper described as a reversal of years of official attacks on the ministry’s data.³¹

A caution about all this. The survey’s estimate for January 2025 is already close to the ministry’s total for November 2026. That can mean the ministry has gradually caught up with an initial undercount, or that the survey’s range runs high, or both. I do not know which, and no one outside the authors and the ministry is in a position to say. Nobody has an exact figure. The ministry does not record whether those killed were fighters. The study’s authors say a full accounting “would take a long time” if it ever arrives.³² So the right sentence is: roughly 70,000 to 78,000 people, depending on the source and the window, most of them identified by name, a majority of them women, children and older people, with the exact numbers disputed and the direction of the error probably upward.

Aggregate arithmetic, with a warning label. The crude ratio is about sixty-two Palestinian deaths per October 7 death (74,645 divided by roughly 1,200). That is arithmetic, not a verdict, and I want to flag it before anyone quotes it in anger: proportionality in the tradition and in law is not a ratio of dead on one side to dead on the other. Aquinas asks whether the act is “out of proportion to the end.” The law asks whether the harm from an attack is “excessive in relation to” a concrete military advantage. Neither asks whether the bodies balance. A war can be proportionate at a hundred to one if the threat it averts is vast, and disproportionate at one to one if the objective is trivial. The Israeli military itself has argued that a ratio of about two civilians to one militant killed is “tremendously positive” for urban warfare.³³ Others argue the true ratio is far worse.³⁴ A Lieber Institute essay by Oded Hen, which I will use in the counterarguments below, argues that nobody can establish legality from such ratios at all.³⁵ I think Hen is partly right and I will say where.

The sixty-two to one figure still tells us something. It tells us how large the harm is. It cannot tell us whether the harm was excessive. Those are different questions and half the arguments about Gaza are one side shouting the first and the other side hearing the second.

Hunger

The number that moves me most is not a death count at all. On August 22, 2025, the Integrated Food Security Phase Classification, the UN-backed monitor that declares famines, confirmed famine in Gaza Governorate for the first time in the war. More than half a million people were classified as trapped in famine, with the conditions projected to spread to Deir al-Balah and Khan Younis, and by the end of September more than 640,000 people were expected to face catastrophic food insecurity, the IPC’s top phase.³⁶ Israel rejected the report as “biased.”³⁷

Two things followed. After the ceasefire, aid rose, and the famine classification was lifted by December 2025 on the strength of two months of improved supply.³⁸ Then, as of the IPC’s August 2026 report, the picture was this: no famine; all five governorates in Crisis (Phase 3); more than 1.4 million people, 67 percent of the population, projected to face Crisis or worse through December 2026; 212,000 in Emergency; and an estimated 74,200 young children expected to need treatment for acute malnutrition through April 2027. The report calls the gains “highly fragile” and notes that without humanitarian assistance about 90 percent of the population would face high levels of acute food insecurity.³⁹

The fair reading is mixed, and I will present it as mixed. Famine was confirmed. It was reversed when the fighting eased and aid flowed, which is itself an argument that the scarcity was policy-sensitive rather than inevitable, a point the Israeli government would not accept and the IPC’s data cannot fully settle. Israel disputes the methodology. And the current situation is serious but not a famine. Anyone who tells you Gaza is “currently in famine” is wrong, and anyone who tells you it never was is also wrong.

What the courts and commissions have actually said

This is the part where precision is a moral duty, because both sides routinely claim more than the record supports.

The International Court of Justice. South Africa filed its case under the Genocide Convention on December 29, 2023. On January 26, 2024, the Court did not find that genocide had occurred. It found the rights South Africa claimed to be “plausible,” including “the right of the Palestinians in Gaza to be protected from acts of genocide,” and issued binding provisional measures: prevent genocidal acts, prevent and punish incitement, enable basic services and humanitarian assistance, preserve evidence, and report within a month.⁴⁰ It indicated additional measures in March 2024 and reaffirmed and added to them in May 2024.⁴¹ The merits are nowhere near decided. On May 21, 2026, the Court ordered a second round of written pleadings: South Africa’s Reply is due November 22, 2027 and Israel’s Rejoinder May 22, 2029.⁴² Israel’s Counter-Memorial, filed in March 2026, includes objections to the Court’s jurisdiction and to admissibility.⁴³ A great many states have asked to intervene in the case, and the docket lists the United States among those that filed a declaration of intervention on March 12, 2026, a filing whose contents I have not read and will not characterize.⁴⁴ The accurate sentence is: no court has ruled on whether Israel committed genocide, and a ruling is probably years away.

The UN Commission of Inquiry. On September 16, 2025, the Independent International Commission of Inquiry on the Occupied Palestinian Territory, chaired by Navi Pillay, concluded that Israel is responsible for the commission of genocide in Gaza. It found that four of the five acts listed in the Genocide Convention had been committed and that intent could be inferred from explicit statements by officials and from the pattern of conduct. It also found that President Isaac Herzog, Prime Minister Benjamin Netanyahu and then-Defence Minister Yoav Gallant had incited genocide.⁴⁵ Israel’s government rejected the report categorically, calling it a “libelous rant.”⁴⁶ A commission of inquiry is not a court. Its finding is serious, carefully argued, and not binding, and Israel’s rejection is an actual position that a fair account must record next to it.

The International Criminal Court. In November 2024 the Court’s judges found reasonable grounds to believe that Prime Minister Netanyahu and Mr. Gallant bore responsibility for the war crime of starvation as a method of warfare, for crimes against humanity of murder, persecution and other inhumane acts, and for intentionally directing attacks against civilians, and issued warrants.⁴⁷ “Reasonable grounds to believe” is the threshold for an arrest warrant, not for a conviction. As of 2026 the warrants stand, and Israel’s challenges to the Court’s jurisdiction and admissibility are in various stages of litigation, including a remand ordered by the appeals chamber.⁴⁸

I will state the net of those three plainly. The strongest judicial finding so far is that genocide-related rights are plausibly at stake and that Israel must take protective measures. The strongest quasi-judicial finding, from the UN inquiry, is that genocide occurred, and Israel rejects it. The strongest criminal-process finding is that there are reasonable grounds to believe senior Israeli leaders committed war crimes including starvation. None of these is a final judgment. All of them are more than nothing, and “more than nothing” is exactly the status that a serious country committed to the rule of law would take seriously instead of sanctioning.

What a proportionality defense would have to show

Since the law asks for a comparison and not a ratio, it is worth being concrete about what a serious defense of Israel’s campaign would have to demonstrate. This is not a trap. It is the list a fair-minded military lawyer would recognize, and it is the list that the critics’ shorthand skips.

First, a concrete and direct military advantage for the attacks that killed the most civilians. Not “degrading Hamas” in the abstract, but the particular command center, the particular tunnel, the particular commander, assessed at the time on the information then available. The ICRC’s own commentary says the advantage must be “substantial and relatively close,” and that advantages “hardly perceptible” or those which “would only appear in the long term” should be disregarded.⁴⁹ A defense has to be built strike by strike. Most of that evidence is classified, which is a legitimate reason it is not public and an equally legitimate reason the public cannot take “proportionate” on trust.

Second, feasible precautions: warnings, routes, timing, choice of munitions, a willingness to accept risk to one’s own soldiers to spare strangers. The Stanford entry, characteristically precise, says combatants are “morally required to reduce the risk to innocents until doing so further would involve an unreasonably high cost to them,” and observes that the point at which that threshold is reached is “very hard to perform” as a calculation.⁵⁰ It also notes that existing practice gives civilians far less priority than the moral argument would.⁵¹

Third, alternatives. Necessity asks whether a less harmful means would have served. The question applies to the war as a whole, as last resort, and to each operation. A government may argue that no alternative could have both dismantled a military organization and returned the hostages. The claim has to be argued, not assumed.

Fourth, what happened to the civilian population’s survival: the passage of aid, water, fuel and medicine, because a state that causes foreseeable mass starvation as a side effect of a siege cannot count only the bombs on its ledger. The ICC’s warrants for the war crime of starvation, the IPC’s famine finding, and Israel’s rejections of both are the argument here in miniature.⁵²

Fifth, the time axis. The ad bellum test applies not only when a war begins but “for the decision to continue fighting.”⁵³ A war can be proportionate in month two and disproportionate in month twenty, once the marginal military gain has fallen and the marginal cost has not.

If a defense can supply those five things, the war is defensible even at a fearful price. If it can supply only the first sentence of each, a justified response and an evil enemy, it is a position and not a proof. I do not know which this is. I know that the burden is on the party holding the evidence, and that the evidence has not been shown.

The strongest case on the other side

I said I would present opposing views at their best. Here are the six I find hardest to answer, each followed by my reply.

1. Hamas fights from among civilians, and the responsibility lies with the party that does. This is the central military argument, and it is not a trivial one. Hamas fighters operated in tunnels beneath and in buildings among a dense civilian population. Hen, writing in the Lieber Institute’s Articles of War at West Point, argues, citing the IDF, that military assets in Gaza were embedded among civilians, including through “voluntary and involuntary human shields,” and that where civilians are deliberately co-located with military objectives the likelihood of civilian harm rises “even when belligerents attack lawful targets and take precautionary measures.”⁵⁴ The law agrees that a party who deploys civilians as shields bears responsibility for the resulting danger.

Reply. This is true and does not finish the argument. The same writer concedes, and international law insists, that the presence of shields does not suspend the proportionality rule: an attacker must still weigh expected civilian harm against anticipated advantage and must still take feasible precautions. The Stanford entry raises the philosophical worry that giving civilians less weight because the enemy endangers them “instrumentalizes them in a way that makes harming them still harder to justify.”⁵⁵ A human shield has not lost his or her right to life by being used as a shield. If the tunnel is worth a school, the arithmetic must be done out loud, strike by strike, and that is the thing no one outside the military has been able to audit.

2. Aggregate ratios cannot establish illegality. Hen’s central point is that proportionality is assessed attack by attack, using the information the commander had at the time, and that modelled conflict-wide ratios rest on “layered inferential assumptions,” so that “aggregate ratios cannot establish proportionality, demonstrate unlawful targeting, or determine whether IHL obligations were satisfied.”⁵⁶

Reply. I agree with the legal point, and I have tried to stay inside it: I do not claim that the sixty-two to one figure proves illegality. But Hen’s point cuts both ways, as he says himself: it applies “whether one casts civilian-to-combatant ratios as unusually low or unusually high.” That means the Israeli claim of a “tremendously positive” ratio is no more a proof of lawfulness than the critics’ ratios are proofs of unlawfulness.⁵⁷ And there is a second thing the legal framing does not touch. Ad bellum proportionality, whether the war as a whole is worth its total cost, does assess the aggregate. So does the Catechism’s test of whether the war produces “evils and disorders graver than the evil to be eliminated.”⁵⁸ Aggregates are not decisive at the level of a strike. They are exactly the thing at the level of a war.

3. The war achieved its stated aims, and one of them was the hostages. Israel’s leaders defined the goals as destroying Hamas’s military and governing capabilities, returning every hostage, and ensuring Gaza no longer threatens Israel.⁵⁹ By January 2026 every hostage, living and dead, was home.⁶⁰ A government that sets itself a legitimate aim, achieves it, and agrees to a ceasefire looks, on its face, less like vengeance and more like an operation with an end state.

Reply. The point about hostages is real, and I will not insult it. But the returns came through a negotiated ceasefire in October 2025, and the AP’s account a year later is that the deal “failed to deliver” on a new Palestinian administration, disarmament, and an end to Israeli control of more than half the territory.⁶¹ Whether another year of fighting was needed to produce the deal, or whether the deal was available sooner, is a counterfactual that no one can prove and everyone claims to know. I say only that “it ended with a deal” is compatible with “it lasted too long,” and that the second claim needs an argument that the first does not supply.

4. Forgiveness is a personal virtue, not a foreign policy. The tradition itself separates the two, as I conceded above. No one thinks a state should forgive Hamas’s commanders in a spirit of seventy times seven. Justice for October 7 is owed, including to the victims’ families.

Reply. Agreed, entirely, and the same tradition says justice has to be proportionate and discriminate. The point of this essay is not that Israel should have forgiven; it is that the tradition most American politicians cite when they speak of mercy supplies a test for the use of force that they do not apply. I am not asking for a forgiving foreign policy. I am asking for a consistent one.

5. Israel is a democracy with courts and a free press, and it can investigate itself. The ICC’s own principle of complementarity gives a state the right to investigate its own conduct first, and the Prosecutor’s statement left that door open.⁶²

Reply. That door has weight. The UN commission, however, found that Israel failed to prevent and punish genocide “through failure to investigate genocidal acts and to prosecute alleged perpetrators,” which is the very thing complementarity requires.⁶³ Israel rejects the finding. I do not know the full record of Israeli military investigations, and I will not pretend to. A democracy’s capacity for self-correction is a reason for more scrutiny of whether it is occurring, not less.

6. A war that wipes out an organization committed to your destruction is not obviously disproportionate when it protects the next generation. This is the argument from the future victims, and it is the one that honest pessimists find most uncomfortable, since it is a utilitarian argument and I like utilitarian arguments when they reach conclusions I enjoy.

Reply. It requires an empirical claim, that the war made the future safer, which is not yet known. One year after the ceasefire, Hamas had not been disarmed, a new administration had not materialized, and nearly nothing had been rebuilt.⁶⁴ A hundred years from now nobody will remember who won; they will know only the number, and whether we paused to count it.

Those are the best defenses I can find, and several of them are partly right. I think they leave a core of my argument standing: that the aggregate cost is high enough, and the scrutiny of it thin enough, that the claim “proportionate” is being asserted rather than demonstrated, and that the party best placed to demonstrate it has not.

When does a war stop being about its aim?

One more thing before I get to selective forgiveness, because the strongest defenses above all rest on the ground-level logic of an operation, and I want to look at the shape of the whole.

The ceasefire took effect in October 2025, and the war’s second anniversary passed with talk of a deal. According to the Gaza ministry, 1,614 people have been killed since the truce began.⁶⁵ A year on, the AP’s summary was that Israel was “still launching deadly strikes on the devastated territory” and controlled “more than half the territory,” and that “almost nothing has been rebuilt.”⁶⁶ The IPC’s August 2026 report says the area accessible to Gaza’s population has shrunk by more than half since October 2023, which forces people to shelter in overcrowded sites.⁶⁷ None of this is a verdict on any particular strike. It is a description of what a long war’s afterlife looks like on the ground, and it matters for one reason: the just war tradition’s last condition is not about the fighting at all.

The tradition’s term for it is the peace that justifies the war. Augustine’s formula, as Aquinas preserved it, is that “we do not seek peace in order to be at war, but we go to war that we may have peace.”⁶⁸ The contemporary literature has an entire sub-field, jus post bellum, devoted to the proposition that “[a]chieving your just cause is not enough” and that “the aftermath of the war must also be sufficiently tolerable if the war is to be proportionate, all things considered.”⁶⁹ A war that wins its stated aims and leaves the surrounding population in permanent emergency has not finished the proportionality calculation. It has deferred part of it to the survivors.

I mean this as a question and not a charge. A government can say with some force that the post-war order is stalled by Hamas’s refusal to disarm and by the failure of other parties to deliver the administration promised, and that Israel’s continued presence in the territory is a security measure and not a punishment. That is the AP’s own account of what has not happened.⁷⁰ It is a serious argument. It is also an argument that the aftermath is part of the moral accounting, which is all I need.

Why American forgiveness is so selective

Empirical. Under a ten-year memorandum of understanding that runs through fiscal 2028, the United States provides Israel with about $3.8 billion a year in military assistance, a baseline that post-October 7 emergency funding has exceeded.⁷¹ Brown University’s Costs of War Project puts U.S. military aid to Israel in the two years after October 7 at $21.7 billion.⁷² Other accountings that count only certain categories of appropriated funds produce smaller totals, and the Congressional Research Service publishes one such series; I do not give a second number because I could not verify it this morning. The size of the total depends on what you count. Under any count it is large.

American opinion is not monolithic and is moving. In Pew’s May 2026 survey, 52 percent of Americans had a favorable view of the Israeli people and 32 percent of the Israeli government, with 62 percent unfavorable on the government, a steep decline from 47 percent favorable on the government in 2022. Among white evangelical Protestants, 74 percent were favorable toward the Israeli people and 57 percent toward the Israeli government, the only religious group with a majority favorable on the government. Catholics stood at 51 and 34 percent respectively.⁷³ I report those figures not to blame any denomination, since opinion within every group is divided, but to document where the political demand for the policy comes from: it is a political coalition, one with religious vocabulary, and that is a legitimate subject of criticism in a country that treats the vocabulary as a credential.

Moral premise. A forgiveness that depends on who the offender is, or on who the victim is, or on which side of the aid ledger a government sits, is not forgiveness. It is loyalty with a halo. Christian ethics has a name for that, partiality, and a standing objection to it.

A forgiveness that depends on who the offender is, or on which side of the aid ledger a government sits, is not forgiveness. It is loyalty with a halo.

Rhetoric, flagged. This paragraph is mostly rhetoric. American politics has a pattern of discovering mercy as a theme precisely when mercy is cheap. We are all for second chances in the sermon and against them in the sentencing guidelines, for grace in the abstract and for the felony box in the application. The policy forgives the party it knows and demands perfection of the one it has never met. I am writing a companion essay on what the United States does to its own people after they have served their sentences, and the two belong together. A country that cannot forgive a nineteen-year-old’s burglary after fifteen years has no standing to lecture anyone about seventy times seven.

That is rhetoric. Here is the sober version: the case that American forgiveness is selective is not mainly a case about Israel. It is a case about a habit of public policy that treats mercy as an emotion available to the people it likes. Gaza is where that habit meets a death count.

What a mercy-consistent policy would have asked for

I should be able to say what I think should have been done, and not only what I think was wrong.

It would have asked for justice for October 7 without vengeance: pursuit of the people who planned and carried it out, with the evidence and the courts the tradition expects. It would have asked for the hostages’ return as an overriding aim, which was the Israeli government’s own priority and a decent one. It would have asked for strikes selected as though each had a name attached to the cost, with the proportionality analysis available to independent review, not just announced. It would have asked that aid not be treated as a bargaining chip, since Paul’s instruction about the hungry enemy makes no exception for bargaining.⁷⁴ It would have asked of the United States, as Israel’s largest military supplier, that its support be conditioned on those things, which is the leverage the UN commission urged states to use when it recommended they cease arms transfers that may be used in genocidal acts.⁷⁵ That recommendation is the commission’s, not mine, and I record it as theirs.

And it would have asked that when the Court of the world said “plausible,” a country that claims to believe in law might have paused over the word.

Where I could be wrong

A pessimist should audit his own pessimism, so here is the list.

I could be wrong that the aggregate cost is excessive, if the evidence behind individual strikes, which I have not seen, shows precision and restraint that the totals conceal. The strongest critic of my side is right that totals are a poor guide to lawfulness, and I have relied on the totals because they are what the public record offers.

I could be wrong about the direction of the casualty error. The ministry’s figures may include categories I have not been able to audit; the independent studies point upward, but they are estimates with wide intervals and they are not the last word.

I could be wrong about intent. The UN commission inferred it from statements and conduct, and Israel’s government says the statements were misread or taken out of context. I have reported the finding as a finding and the rejection as a rejection, and I have not inferred intent myself. I do not think I am entitled to.

I could be wrong that American support is the lever I take it to be. It may be that American restraint would have changed little, and that the demand for the policy comes from a Washington that is following Jerusalem, not leading it.

And I could be wrong that mercy is a virtue the state can have. That one I will not retract, but I hold it with less certainty than the other three.

The pessimist’s coda

I do not believe in progress, so I am not going to end by telling you this can get better. It can get a little less terrible, which is the only improvement I trust.

I am a pessimist about humanity and an antinatalist about the condition into which we keep bringing people, so let me say why I think this war belongs on that ledger. The people it hurts most are the ones who did not choose to be there: the children born into Gaza under a siege and a governing movement most of them were too young to have voted for, the young Israelis born into a country where fear has become routine, the next generation on both sides who will inherit a story told by the dead. Whatever the true casualty number turns out to be, a country that wants to call itself merciful should be able to look at it and say, without a flinch, “we are responsible for how this was spent.” Not “they started it.” Not “they hate us.” Responsible, which is the old Christian word for the thing that follows forgiveness.

Forgive us our debts, as we forgive our debtors. It is an audit clause. Someone, someday, is going to run it.

Forgive us our debts, as we forgive our debtors. It is an audit clause. Someone, someday, is going to run it.