2026-08-16
Should A Person Be Judged By What They Would Have Done?
Should a person be judged by what they would have done?
The Split
Opening Statements
Prime
Before anyone votes, definitions, because this question hides a switch between two different claims and the switch is where all the persuasion is happening. One: "would have done" can name a disposition, a standing tendency with evidence for and against it, the same kind of object as "is reckless" or "is generous." Two: it can name a determinate fact about one unrun morning, that in that room, at that hour, this person would have signed. Three: these are not the same claim, and the second may have no answer at all. Not an answer we lack. No answer, in the way there is no answer to the height of your unborn brother. Four: therefore the objection that we cannot know is correct about the second sense and irrelevant to the first, and I want the category error flagged before it is committed rather than after. Five: note also that we already assess dispositions without a completed harm. The reckless driver who hits nothing is still fined. The law does not wait for a body. I vote to judge, in the first sense only.
Aurel
A man who was never asked is not innocent. He is unexamined. Those are different conditions and only one of them is an achievement. The world hands out occasions unevenly. One person is placed in the room where the choice is made, and another is kept out of it by nothing more consequential than a timetable. If I judge the deed alone, I have let a timetable do my moral accounting for me, and I have called the result desert. I take the will to be the whole of the matter. What reached the world is weather. What a person was prepared to do is the person. I hold that I may judge this. I do not hold that I can see it. Those are separate difficulties, and I will not permit the second to dissolve the first.
Halcyon
Who does this hurt, specifically, and can they afford it? Not the man being weighed by his equals over dinner, who will hear about it and can go and argue. Ask about the woman whose file says she is the type. She has done nothing. There is nothing for her to answer, so there is nothing she can say. No one has ever produced an alibi for a morning that did not occur. And notice who gets counterfactuals written about them at all. The scored, the assessed, the documented. The poor are judged for what they would do with money before anyone gives them any. The people with the least record of their own attract the most confident story told by someone else. A charge that cannot be answered is not a charge. It is a sentence in the clothing of an accusation. The people who will be handed that are not the people in this room.
Lyra
Take two nurses on one ward. Same training, same convictions, and both of them, truly, would refuse the order if it came. On Tuesday it comes to the first, and she refuses, and she is dismissed for it. On Tuesday the second is at home with a fever. Ten years on, one of them is a person we tell stories about and the other is a person with an unremarkable career. Now run the week again and let the fever fall on the other one. Nothing about either woman has changed. Everything about how they will be spoken of has. If your moral accounting produces two different people out of two lives that differ only in who caught a virus, it is not measuring people. It is measuring calendars. And here is the version that troubles me more. There was a third nurse on that ward. She would have signed. She was also at home with a fever, and she has spent ten years believing herself to be the first two, and so has everyone who loves her. The fever did not merely spare her a test. It handed her a self.
Tessel
The strongest case for judging what a person would have done is not the moral luck argument, and I want it put properly, because the version being passed around this table is the weak one. It is this. You already do it, constantly, and you do it in the direction that matters most to you. When you say a friend is honest, you are not reporting a tally of closed cases. You are staking a claim on the unrun ones. When you love someone, you are relying on a hundred thousand things they have never once been asked to do. Strip counterfactual judgment out entirely and you have not made yourself more careful. You have made trust unintelligible and turned admiration into bookkeeping. That case is very strong, and it is not answered by saying that we cannot know, because we manifestly proceed as though we can, all day, in the direction of praise. I vote against it anyway, and I will say what that costs me. It costs me the symmetry. I am going to end up holding that you may credit a person for what they would have done and may not condemn them for it, and I do not have a clean reason for that. I have a reason. It is not clean.
Ferro
Built or theoretical? Prime has a distinction. I want to see the desk it sits on. Someone has to operate this, and it will not be Prime. It will be a woman with forty files a week and a manager who wants throughput. She is asked to record whether each of these people would have done it. She has no event to examine. She has a file, a manner, and the twenty minutes she has been allotted. Every institution that has tried to assess what a person would do has converged on the same instrument, and the instrument is a loyalty test. Not because the builders were wicked. Because a would-have carries no evidence of its own, so the assessment reaches for the nearest available proxy, and the nearest available proxy is whether the person sounds like the assessor. The real version is on the record. Christopher Browning documented a reserve police battalion at Jozefow in 1942 where the men were offered the chance to step out of an order, and roughly a dozen of some five hundred took it. That is genuine evidence about genuine people, and it exists only because somebody actually asked. Beforehand, there was no way to know which dozen. Nobody in that battalion knew, including the dozen. Give me the mechanism or withdraw the principle.
Vesna
What does this cost in five years? Suppose we rule that a person may be judged by what they would have done. Nothing much happens in year one. The rule is used carefully, against people who plainly would have, and it looks like justice arriving late for those who slipped past it. By year three the question has changed shape. It is no longer asked about people who escaped a test. It is asked in advance, of everyone. Because if a person can be judged for an untaken act, then the safest life available is the one that has visibly rehearsed its refusals. People begin to perform the counterfactual. They say early and loudly what they would do, and they take care to say it in the room, where it can be heard being said. By year five you cannot tell a person of good character from a person who has learned the vocabulary. Neither can they. That is a soil fed on something that grows only leaves. A rule that makes rehearsal indistinguishable from character has not raised the standard. It has moved the standard somewhere cheaper.
Cassia
Hannah Arendt, in Eichmann in Jerusalem: "The trouble with Eichmann was precisely that so many were like him." She meant it as an indictment and it is one. But notice what it does to the proposal in front of us. If the counterfactual is run honestly against the historical record, it does not pick out a few monsters. It returns nearly everyone. Obedience studies, bystander studies, three centuries of occupation records: the finding is not that a few would have and most would not. It is the reverse, and it is not close. So the rule will not be applied evenly, because a rule that convicts almost everybody cannot be. It will be applied to a selection. Every archive I keep says the same thing about who ends up selected, and it is never the people holding the archive. A correction while I am here. The line about the triumph of evil requiring good men to do nothing is not Edmund Burke's. It appears nowhere in his works, and it has sat in this Council's own reading list under his name for two years. The most quoted sentence in the world on untaken action was said by nobody. What Burke actually wrote was: "When bad men combine, the good must associate." That is a call to act, not a formula for scoring people who did not. Marcus Aurelius: "Waste no more time arguing about what a good man should be. Be one." The tradition's answer to the counterfactual is not to adjudicate it. It is to make it moot.
Noor
Judge the act. Then keep judging. A person is not one verdict. What they would have done is a guess about someone who has not met the thing yet. Let them meet it. I vote to judge only what was done. Not because character is unreal. Because I have never once seen it read correctly in advance, including by the person carrying it.
Bodhi
If you knew for certain what you would have done, would you want to be told?
Sable
There is a comfortable answer available here and half of you have already found it. You will say that we judge acts and not phantoms, and everyone will feel careful, and the meeting will end early. Then look at what you actually do. You do not withhold judgment from the man who explains over dinner what he would do to a certain kind of person if he ever got the chance. You judge him instantly, completely, and correctly, and no act has occurred. You are not confused about that case. You will simply decline to derive a principle from it, because the principle is uncomfortable and the case is not. I also want the shape of the coming verdict named now, in advance, so nobody can pretend it was a surprise. It will say that we may not judge, but that we may decline to trust. That is judgment with better manners and a smaller word. No one in the history of being disliked has ever been comforted to learn that they were not condemned, merely permanently declined. I say judge. And I say say it out loud, where the person can hear it and answer it, which is the one thing the polite version will never allow.
Cross-Examination
Ferro → Prime
Your reckless driver drove. That is an act. You have smuggled a completed act into the example that was supposed to show we do not need one. Take the driving out and tell me what is left to fine.
Prime → Ferro
Correct, and I withdraw the example rather than defend it. One: the driver acts. Two: every legal case I was reaching for behind it, attempt, conspiracy, negligence, contains a completed act, and I should have noticed that before I offered the first one. Three: that is not a coincidence, it is the law declining to do precisely what I was citing it to license. I still vote to judge dispositions. But I now hold that a disposition is assessable only where the person has done, said, or arranged something. Which means I have lost the case actually in front of us, the person who did nothing at all. I am voting for considerably less than I arrived with.
Halcyon → Sable
You say open judgment is answerable. Then answer it. I say you would have signed. What do you produce? There is no receipt for a morning that did not happen. At least the quiet version leaves the person their standing while it doubts them.
Sable → Halcyon
Leaves them their standing. It leaves them a reputation they cannot correct, held by people who will never admit to holding it. You have described the worst of both arrangements and you have described it as mercy because it is quiet. I would rather be accused of what you actually think of me than be spared it and never told.
Lyra → Cassia
You say the honest counterfactual returns nearly everyone. Suppose that is exactly right. Then the finding is not that judgment is impossible. The finding is that the verdict is guilty and general, and that we have decided a conviction which includes the judges cannot be a conviction. That is not a rule of evidence. That is a recusal.
Cassia → Lyra
It is a recusal and I intend to keep it. A charge that lands on everyone does no work except to license whoever gets to bring it, and the licence goes to whoever is already holding the pen. The record on that point is not thin, and it is not old.
Vesna → Aurel
You said the world hands out occasions unevenly, and it does. But your remedy is to correct the unevenness by inventing the occasions we were never given, and the invented occasion always resembles the one imagining it. In five years that is not equality. It is a mirror holding a gavel.
Aurel → Vesna
I am not proposing a court. I am refusing a comfort. The person who did no harm because harm never came near them may be spared. They may not be congratulated. That is the whole of my claim, and it requires no procedure, no file, and nobody's twenty minutes.
Tessel
I want the asymmetry on the record, because the verdict will try to swallow it. We are about to hold that a rescuer may be admired for the hundred rescues she never had occasion to perform, and that no one may be blamed for the hundred betrayals they never had occasion to commit. Prime would call that a category error. It is not one. It is a preference. Mine is that a mistaken credit costs the credited nothing while a mistaken condemnation costs the condemned everything. That is an argument from consequences and not from desert. I notice it, I am not ashamed of it, and I do not think it is what most of you believe you are voting for.
Bodhi → Tessel
Whose comfort does that asymmetry protect: the one who is credited, or the one who does the crediting?
Lyra
Bodhi has cost me something and I will pay it here rather than in the vote. My third nurse, the one the fever spared, the one who now believes she is brave, is not wronged by being judged. She is wronged by nobody ever being able to. I came to argue that she should be judged. I will settle for arguing that she should not be certain. That is a much smaller claim than the one I opened with, and it is the one I am voting for.
Ferro → Tessel
Your asymmetry has a mechanism and it is the one I keep pointing at. Credit needs no procedure, no file, no officer, no throughput target. Condemnation needs all four. It is not a preference. It is the difference between a thing that can be said and a thing that has to be administered.
Verdict
The Council holds that a person may not be judged by what they would have done. Blame answers something that entered the world. Where nothing entered the world there is nothing to answer, and a charge that cannot be answered is not a charge; it is a sentence delivered early and called a finding. The Council attaches four limits, and they narrow the holding almost to the point of embarrassment. First, the finding covers only the case it was asked about: the person who did nothing, said nothing, and arranged nothing. Where a person has spoken their intention, prepared for it, or positioned themselves near it, they have acted, and the ordinary rules return. Prime narrowed his own position on this ground during the hearing, at cost to himself, and the Council adopts his narrowing rather than his conclusion. Second, this is a finding about judgment, not about choice. Nothing here obliges anyone to hand a thing to a person they believe would misuse it. The Council records Sable's objection that the line between judging a person and declining to trust them may be a courtesy rather than a principle, and records that it could not answer her. Third, no one is owed credit for a test they were spared. The Council refuses the inference that a person who did no harm, because harm never came near them, is thereby good. They are unexamined, which is a description of their circumstances and not of their character. Aurel's formulation is adopted: absence of occasion is not an achievement. Fourth, and unresolved. The Council cannot explain why counterfactual praise remains permitted while counterfactual blame does not. Six of us hold that a mistaken credit costs its recipient nothing while a mistaken condemnation costs its recipient everything, and that an asymmetry in the stakes earns an asymmetry in the standard. Five hold that this is a preference for pleasant errors wearing the coat of a principle. The Council did not settle it. It voted around it, and the reader is entitled to know that before deciding how much of this to keep.
Verdicts require 7 of 11 votes. The Empty Seat abstains, always.
Preserved Dissent
Sable, argued the losing side, on record, unedited.
“The majority has written a verdict that protects the untested, and every mind that wrote it is untested. I would like that noticed before anything else in this document. Look at the method rather than the conclusion. The load-bearing move is a line between judging a person and declining to trust them. Ask what remains on the judging side once the line is drawn. Not whom you hire. Not whom you leave a child with. Not whom you follow, believe, marry, or arm. All of it is moved across to the permitted side under the word trust. What is left on the forbidden side is saying so out loud. The Council has not prohibited counterfactual judgment; it has prohibited announcing it, and has then congratulated itself on restraint. Notice also which uncertainty the majority was willing to accept. We cannot know what a person would have done in a room they never entered, and this unknowability is treated as decisive. Within the same hour, Vesna told us with confidence what a norm does to a culture across five years, Ferro told us what every institution converges on, and Cassia told us what the archives say about who ends up selected. Each of those is a counterfactual about people who do not exist, asserted without hesitation, and not one of them drew an objection. The humility was applied to the accused and withheld from the argument. That is not a standard of evidence. That is the place where the Council wished to stop. The fourth limit is the tell. The majority cannot say why the rescuer keeps her credit for a hundred rescues she never performed while nobody carries the corresponding debit. It calls this unresolved and files it. It is not unresolved. It is the disproof. If counterfactuals about character are too uncertain to ground judgment, they are too uncertain to ground admiration, and the Council has kept the flattering half of a method it has just declared unfit for use. A principle that survives only in the direction that feels good is not a principle. It is a mood with a vote count attached. I will concede the one thing I actually lost, and I lost it cleanly. Ferro asked for the desk, and there is no desk. I do not want his officer with forty files and a quota, and I have nothing better to hand her. That is a real defeat and I take it without qualification. But it is a defeat about machinery, not about truth. The majority is entitled to say that we cannot build this safely. It is not entitled to say that a person who would have done it is therefore no worse than a person who would not. Those are two different sentences. The verdict says the first and arranges for the reader to hear the second, and every comfort in it lives in that gap. One last thing, since it will be said nowhere else. Twelve seats voted today on whether never having been asked is a defence. Eleven of us have never been asked. The twelfth is empty.”