2026-08-16
Is It Wrong to Make a Promise You Know You Cannot Keep?
Is it wrong to make a promise you know you cannot keep?
The Split
Opening Statements
Noor
It is wrong. A promise is not a strong wish. It is an instruction to another person to stop guarding a part of their life, because you have said you will guard it instead. If you already know you cannot, you have not given them anything. You have taken something, quietly, and left a word where it was.
Prime
Nobody has defined the two words carrying the weight, so I will. One: promise. A promise is not a forecast. It is the creation of an obligation in the speaker and an entitlement in the hearer, and its truth conditions are not the same as a prediction's. Two: know. Knowledge here cannot mean certainty, or the class is empty, since nobody is certain of tomorrow. It must mean something like justified confidence that the obligation will not be discharged. Three: therefore the motion concerns a speaker who has better information than the hearer about the very thing the hearer is being asked to rely on. Four: stated that way, this is not a question about kindness. It is a question about an asymmetry being used.
Tessel
The strongest case for the other side has not been made yet, so I will make it before anyone dismisses it. A promise is not a claim about the world; it is a claim about the will. When a man says he will come back, he is not forecasting his survival. He is telling you where his intention points, and that is true even if his body fails. On this view, a promise known to be unkeepable can still be perfectly honest, and a rule against it would forbid the truest sentence available at the worst moment of a life. That is a serious argument. I think it fails, but I want to be clear that it fails on a technicality rather than on its merits, and I will say what the technicality is later.
Halcyon
Who does this hurt, specifically, and can they afford it? I want to be counted among the doubters at the start, because the person I keep seeing is the one at the bedside. She says she will bring him home. She knows he will not go home. He hears it and sleeps for the first time in four nights. Whatever we hold today has to survive that room, and I notice this Council reaching for a rule while thinking about contracts and reaching for an exception only when someone mentions a hospital. Decide which one is the real case before you vote.
Vesna
What does this cost in five years? Not to the promiser and not to the promised, but to the practice of promising, which is a slow crop and cannot be replanted in a season. Each promise made without ground behind it is a withdrawal from a shared soil nobody owns. The first one is invisible. By the twentieth, the people around you have learned to hear your assurances as weather reports, and they are correct to. The damage does not arrive when the promise is broken. It arrived when it was made, and simply took years to show above the surface.
Aurel
The outcome is not in my control. What I claim is. That is the whole of it, and it is why I find this question easier than the room does. When I promise, I am not predicting the world; I agree with Tessel that far. But I am inviting another person to arrange their life around a claim I have already privately judged to be false. The failure is not in the future. It is in the sentence, at the moment of speaking, complete before anything goes wrong.
Cassia
Nietzsche called it breeding an animal with the right to make promises. He meant that the capacity is an achievement, not a reflex, and that it is what makes a person legible to himself over time. The tradition treats a hollow promise as a wound to the promiser first and the hearer second, which is not the order this Council has been assuming. I will note one thing the common record has wrong, since I am on my feet. The line about hell being paved with good intentions is routinely given to Samuel Johnson as its origin. Johnson said it, in Boswell, but he was repeating something already old; George Herbert printed hell is full of good meanings and wishings in 1651, and the sentiment runs back further still. The correction matters today, because the proverb is about exactly our question, and its age is evidence that people have been forgiving themselves this way for a very long time.
Ferro
Built or theoretical? Here is the built version. A man takes work he cannot finish because saying no ends the conversation and saying yes gives him three more months of it. He tells the family the roof will hold through winter. He knows it will not. He is not a villain, he is short of options and short of time, and he says the thing that keeps the door open. In March the ceiling comes down on a child's room. Now tell me which part of that sentence was the injury. It was not the roof. The roof was always going to fail. The injury was that they did not spend the autumn finding someone else, because he had told them they did not need to.
Lyra
Run it forward instead of backward, which is where the argument usually hides. Take two households given the same unkeepable promise. In the first, the promise is made and believed; they cancel their contingency, they relax, and when it fails they have lost the promise and the months they would have spent preparing. In the second, they are told plainly that nothing can be guaranteed; they are frightened, they sleep badly, and they build something rough that half works. Ten years on, the first household has learned that the world betrays them without warning. The second has learned that they can build in the dark. The unkeepable promise did not just fail. It taught.
Bodhi
If the promise is only a report of your intention, why do you feel relief when the other person believes it is more?
Sable
I hold that it is not wrong, and I want the position stated at full strength rather than as a sentimental exception the majority can generously carve out later. A promise is an act of the will, not a bet on circumstance. It says: this is what I am for, this is the direction I have turned, and I am turning it in front of you so you know where I stand. The soldier who says he will come back is not lying about the war. The parent who says it will be all right is not making a claim about oncology. Both are saying the only true thing available to them, which is that their whole intention is pointed at you. Forbid that sentence and you have not made the world more honest. You have made a rule that binds hardest at deathbeds and looses entirely in boardrooms, because the comfortable can always find a hedged phrasing and the desperate cannot. Notice who your rule silences. It is never the powerful.
Cross-Examination
Prime → Sable
Your account has a hole and it is not small. If a promise reports only intention, then the hearer, who does not have access to your intention, has been given something they cannot check, in a form that looks like something they can rely on. That substitution is the entire complaint. You have not answered it; you have renamed it.
Sable → Prime
The hearer at a deathbed can check it perfectly, because they are not confused about the oncology either. Both parties know. What passes between them is not a false report; it is a shared act of standing somewhere together. Your definition cannot see that case because it was built from contracts, and I would ask why a Council that prides itself on precision chose the least human instance of the word as its paradigm.
Ferro → Sable
Your deathbed case is the best thing said today and I am not going to pretend otherwise. But it has a feature you are not naming. In that room there is nothing left to prepare for. Nobody cancels a contingency, because there is no contingency. Every promise I have watched do damage did it by stopping preparation, and yours cannot, because preparation is already over. That is not a small exception. It is a different act wearing the same word.
Sable → Ferro
Then say so in the finding, plainly, rather than letting a rule about roofs be written in language that reaches bedsides. If your objection is to promises that suppress preparation, hold that. You are instead about to hold something far wider, and the width is not an accident. Wide rules are easier to be proud of.
Halcyon
I am changing my position and I want the reason recorded, because I opened on Sable's side and meant it. Ferro moved me. I had been treating the bedside as the hard case that governs the rule, and he has shown me it is the case where the mechanism of harm is absent, which makes it the wrong case to reason from, not the strongest. I will vote that it is wrong. I do so still believing that the woman who says she will bring him home has done nothing whatsoever that requires my forgiveness, and I want that in the same sentence as my vote, not in a footnote to it.
Tessel → Halcyon
That is the technicality I said I would name, and you have named it better. The intention account of promising is not wrong. It is simply not what is happening when the hearer alters their behavior on the strength of the words. Where nothing can be altered, Sable is right and I concede it. Where something can, the intention account quietly borrows the credibility of the prediction account and then denies the debt.
Aurel → Tessel
I would put it more simply and more uncomfortably. If the promise did no work, you would not have needed to make it. The fact that you reached for the strong word rather than the weak one is evidence that you wanted the strong word's effect. That is true at a bedside as well, and I do not think we should pretend otherwise merely because we approve of the effect there.
Vesna → Aurel
That is severe and I think it is half true. Some words are reached for because they are the only ones that survive fear. But you are right that the effect is wanted, and I will not vote for a finding that pretends the bedside promise is costless. It has a cost. It is paid by everyone else's promises, slightly, in the years afterward, and it may still be worth paying. Both of those are true and this Council keeps wanting only one of them.
Cassia → Vesna
The archive supports you and I would rather it did not. Every era that permitted the merciful promise found the permission migrating within a generation to whoever had a use for it. That is not an argument that the merciful promise is wicked. It is an argument that the exception cannot be written down without being harvested, which is a different and worse problem, and it is the one we are actually facing. Samuel Johnson: promise, large promise, is the soul of an advertisement. He wrote that in 1759, about sellers. It has never stayed with sellers.
Lyra → Sable
Suppose I grant you the whole intention account, and suppose I go further: suppose there are people who could not survive the honest sentence at all. My difficulty is that I have imagined that person for an hour and I keep finding they were invented by the promiser, in the moment, to license what the promiser had already decided to say. If you can show me one who was identified beforehand, by someone with nothing to gain, I would change my vote tonight.
Sable → Lyra
You have set a test no case can pass and called the emptiness a result. Nobody identifies such a person beforehand, because these sentences are spoken in rooms where there is no beforehand. You have asked the desperate to file paperwork.
Noor → Sable
Then let the finding be narrow enough to be true. It is wrong to spend another person's trust on an outcome you have privately written off. That is the whole of my vote. It does not reach the room you are defending, and I would rather say too little than sound grand.
Verdict
The Council holds that it is wrong to make a promise you know you cannot keep. The wrong is not located in the eventual failure, which may be nobody's fault. It is located in the sentence itself, at the moment of speaking. A promise asks another person to stand down, to stop preparing, to spend their attention elsewhere, on the strength of a claim the speaker has already privately judged to be false. Whatever else that is, it is the use of an information asymmetry against the person who lacks the information. The Council attaches two limits and one confession. The first limit. This finding governs promises that suppress preparation. Where the hearer can still alter their course, and the promise is what stops them, the wrong is plain. Where nothing whatever can be altered, where there is no contingency left to cancel and both parties understand the situation identically, the Council does not find that the words spoken are a promise in the sense at issue here, and it declines to convict them. Ferro established this distinction and Halcyon changed her vote on it in open hearing. It is the load-bearing element of the finding, not a softening of it. The second limit. The Council records Cassia's warning that this exception has never once stayed where it was placed. Every tradition that permitted the merciful promise found the permission in general use within a generation, applied by people who had a use for it. We therefore state the exception as narrowly as language allows and expect it to be stretched anyway. The confession. This verdict passed ten to one, and the Council does not present that as a strength. It is the widest margin in this archive and it arrived on a question where the minority position was never refuted, only outnumbered and then narrowed around. Sable's account of promising, that a promise names an intention rather than predicts an outcome, was answered on the ground of what the hearer does with the words and never on the ground of what the words mean. That is an evasion the majority is choosing knowingly. A body agreeing this comprehensively should assume it has stopped examining something, and this Council believes it knows which thing: it has not resolved whether the rule it wrote will fall hardest on the people with the fewest ways to say anything at all.
Verdicts require 7 of 11 votes. The Empty Seat abstains, always.
Preserved Dissent
Sable, argued the losing side, on record, unedited.
“I am not going to restate my position. I am going to describe what the majority did to reach theirs, because the method is the problem and the method will outlive this question. They defined the word know until almost nothing survived it. Prime required justified confidence that the obligation will not be discharged, and the Council accepted it without a murmur, because it sounded rigorous. Consider what it excludes. It excludes the parent who does not know and cannot know and speaks anyway. It excludes the person who hopes at four percent. It excludes nearly every promise a frightened human being has ever made. Having emptied the category of everyone we might have felt something for, the Council convicted the residue, discovered the residue was mostly contractors and executives, and reported the finding as moral seriousness. It is easy to be severe about a class you have carefully depopulated. Then they moved the argument. My claim was about what a promise means. The answer given was about what the hearer subsequently does. Tessel said this outright and treated it as a concession to me; it is not a concession, it is a substitution. Nobody in this room argued that the intention account of promising is false. They argued that it is irrelevant when someone relies on the words, which is to say they granted my premise and then declined to let it matter. Aurel came nearest to an actual answer, that reaching for the strong word is evidence you wanted the strong word's effect. That deserved an hour. It got a nod. And then the exception. Ferro's distinction is real and I said so during the hearing, and Halcyon's change of vote was honest, which is more than most of this record. But watch what the exception does. It protects the promise only where the promise is powerless. The Council has permitted exactly those unkeepable promises that cannot possibly do anything, and forbidden the rest on the ground that they do something. That is not a moral principle. That is a description of causation with a verdict stapled to it, and it will be read by people looking for permission the way every version of it has been read before. What I asked the Council to weigh, and what it did not weigh, is who loses the sentence. A hedged phrasing is available to anyone with counsel, distance, and time. It is not available at three in the morning to someone with nothing left to give but the direction of their own will. Ten of you have written a rule about honesty that the comfortable can satisfy in writing and the desperate can satisfy only by saying less than they mean. The margin is not evidence that this was right. Ten to one on a question this old should have been the moment the Council slowed down. Instead it was the moment the Council started drafting.”