Over the past six weeks, one piece of ordinary reasoning has been taken apart — a piece of reasoning so familiar that most of us have never heard ourselves use it. It is the reasoning that lets everyday comfort sit alongside a clear conscience, and it runs roughly like this. The suffering of distant strangers is real, and regrettable; but it is not the same case as a child drowning in front of me; and the proof that it is not the same case is that I am a decent person, and I do not feel like a monster.
We have now examined both halves of that reasoning, and neither survived in the form we found it. The claim that the two cases are simply different turned out, on inspection, to rest on five separate assumptions — about distance, visibility, certainty, numbers, and whose job it is — each of which had to be defended on its own, and none of which proved as solid as the feeling had suggested. The appeal to conscience failed for a different reason. "I do not feel guilty, therefore I am not guilty" assumes that conscience is a reliable instrument; and we established that we cannot tell, from the inside, whether a clear conscience is tracking the moral facts or merely reporting what it is comfortable to believe. How loud the feeling is settles nothing either way.
I should say at once what tonight is not. It is not the verdict. On the first night I said I did not have one, and that is still true. What the six weeks have removed is not the possible answers but the easy versions of them — the versions that asked nothing of the person holding them. What remains is a set of positions, each of which can genuinely be held, and each of which carries a genuine cost. Tonight's business is to state those costs as precisely as we can. Not the answer, but a full and honest account of what each answer costs — including the one answer we have so far declined to examine directly: that Singer was simply right.
This is Philosophy for Us — philosophy for everyone, no degree required. This is the seventh and last of seven episodes on what we owe a stranger.
Last week I set out two arguments — and they were arguments proper, not the raw feelings we had already examined and set aside. The first, from the philosopher Frances Kamm, was that the claim that distance never makes any moral difference is a universal claim, and a single clear counterexample refutes it — so nearness may, after all, carry some real moral weight. The second, from the tradition of care ethics, refused the whole framework rather than one clause of it: morality begins not with the world weighed impartially from no particular standpoint, but with the particular person in front of you who needs your help.
So tonight is a different kind of episode, because there is no new argument left to build. The task is to set the positions side by side, state the cost of each, and then — since the series ends tonight — hand the question back to where it has belonged all along.
Let us begin with the shape of the difficulty, because it is now clearer than at any point in the series, and seeing it clearly is most of tonight's work.
You came into this series holding three commitments at once, probably without noticing that they were in tension. The first is the principle — the one established at the pond in the first episode, in Singer's formulation from his 1972 paper: if we can prevent something very bad from happening, at a cost not comparable to the thing prevented, we ought to do it. That principle was not taken on authority; it was read off your own certain judgement about the drowning child, and it withstood every objection we brought against it. The second commitment is your life — the actual one, with its ordinary comforts, its spending, and the people you love more than you love strangers. The third is your good opinion of yourself: the settled belief that you are a decent person, and not a monster.
What the series has shown is that these three commitments cannot all be kept. If the principle is true and you go on living this life, then the good opinion must go: on the arithmetic, you are a person who lets children die for the sake of comforts, and the absence of guilty feeling does not alter the arithmetic. If instead you keep the life and the good opinion, then the principle must fail somewhere — and you must be able to say where, and why. Since almost no one, honestly surveyed, will give up the life, the real contest is between the principle and the good opinion. And the only honest way to keep the good opinion is to find a genuine flaw in the principle: not to feel your way out of it, but to argue your way out, and to accept what the argument costs.
That is what the positions we spent three weeks building amount to. Each one names a different place at which the principle is said to fail. Each was built as strongly as its actual defenders would build it. Tonight I am not going to re-argue them. I am going to state, for each, what accepting it commits you to. There are four.
The first position holds that the argument proves too much. When the honest application of a principle would leave the person applying it with almost nothing — when it demands the surrender of nearly everything for as long as there is preventable suffering anywhere — that is not a discovery that morality is extraordinarily severe. It is evidence that the principle is defective. The cost of this position is an obligation to specify. If the demand stops somewhere short of everything, you must say where it stops, and why the limit falls at that point rather than another. And wherever you set it, there stands a child whose life would have been saved by the next small amount you have declared yourself entitled to keep.
The second position grants the principle its general authority but claims a permission: a moral entitlement to give one's own life, projects and attachments more weight than a strictly impartial calculation would allow, simply because they are one's own. The cost here is that the permission comes with no principled boundary. Without one, the category of what I am entitled to keep expands by unnoticed degrees, until it coincides with what I wanted in the first place — and the only auditor of that expansion is my own conscience, an instrument this series has given us specific reasons to doubt.
The third position is Kamm's: that distance makes some moral difference. Its cost is twofold. First, what the argument delivers, even if it succeeds, is a reduction and not a release — some discount on what is owed to the distant, and a modest, fragile one; it does not return you to the comfortable position you started from. Second, holding it means making a judgement: that the intuition favouring the near is a piece of genuine moral insight, rather than one of the oldest partialities in human life given a respectable philosophical form. That judgement may be right. It must be defended, not assumed.
The fourth position is the care tradition's: that morality begins with the particular person before you, not with an impartial survey of the world. Its cost was stated last week and has not gone away. The distant child has no face in front of you; a morality built entirely from near relations has a gap in it at exactly the place where she stands. "I owe what reaches me, and no more" may well be the truth about love. It is also, historically, the formula by which every large-scale indifference has been defended. A defender of this position must explain why it is the first and not the second.
Four positions, then. Each can be held; each has been given its strongest form; each carries a cost; and none of the costs is small, and none is obviously greater than the others. The question they jointly pose cannot be answered by anyone on your behalf: which premise of Singer's argument are you prepared to deny — and can you defend the denial aloud, to a thoughtful person who denied a different one? But before that question can fairly be put, there is a fifth position to examine. It is the one we have left alone all series, and its cost should be known before anyone chooses.
The fifth position is that Singer was right.
We have spent three weeks looking for the flaw in his argument, and it is easy to forget that accepting the argument is itself one of the options. You are not obliged to find a false premise. You may conclude that there is none — that the principle holds as stated, indifferent to distance, and that the conclusion, however hard to live, is true, because the difficulty of living a truth is not evidence against it. This position is unique among the five in owing no account of where the principle breaks, because it says the principle does not break. It is not naivety, and it is not zealotry. It is consistency.
What sets this position apart from a seminar exercise is that it has been lived — at scale, by real people, with real money, within the last two decades. The movement called Effective Altruism was, in effect, Singer's principle made operational: if you are going to do good, do the most good, measured as carefully as possible, directed wherever it saves the most lives per unit of money, with your personal sentiments about particular causes set aside. Some of its members took ordinary jobs and gave away a large fraction of their income, year after year. Some adopted the strategy called earning to give: pursue the highest income you can, precisely to have more to give away. Whatever else may be said of them, these were the people in this story who took the principle most seriously. They were not looking for the flaw. They accepted the conclusion and tried to live it.
Honesty therefore requires that we ask of this position the same question we asked of the other four: what did it cost the people who held it? The answer arrived in public, within the last few years, and in two quite different forms. They must be kept apart, because they are different objections, and the difference matters.
The first cost emerged from inside the movement. In late 2022 the cryptocurrency exchange FTX collapsed, in what proved to be fraud on an enormous scale, and its founder, Sam Bankman-Fried, was eventually imprisoned. He had been the movement's most visible example of earning to give — a young man reported to be making billions in order to give billions away — and he was among its largest funders. The money was, in large part, stolen. What happened next is to the movement's credit, and it is also the sharpest form of the objection. The movement did not defend him, and did not defend itself. William MacAskill, the philosopher most responsible for building Effective Altruism, had personally championed Bankman-Fried; and MacAskill said publicly that he felt rage at the fraud and, in his own words, a kind of self-hatred at having been deceived. Beneath the personal responses, a colder question began to circulate within the movement itself: does the injunction to do the most good, by whatever honest means of acquiring money, contain a defect? Does organising a human life around maximising a single number make a person, in the end, credulous towards anyone who promises a larger number? The sharpest critic of this position here is not Kamm and not the care tradition. It is the movement itself, examining what it had built.
The second cost comes from outside, and it must not be folded into the first, because the first is a movement being honest about one man, and the second is an objection the framework itself is poorly equipped to see. It runs as follows. Effective Altruism asks, with great rigour, where money should go. It is much quieter about where money comes from. It is possible to perform the calculation perfectly — the most lives saved per unit, audited, optimised, delivered — while the entire operation still depends on, and is funded from within, the economic arrangements that produced the suffering in the first place. The fortune that pays for the mosquito nets was made within the same global system that helped impoverish the region the nets are sent to. A framework of impartial maximisation has no term in which to register that fact. It can accept money from any source, direct it to the best outcomes, and leave entirely untouched the distribution of power that generates the need it so efficiently meets. On this objection, an exclusive attention to giving well can become — without anyone intending it — a way of making the underlying arrangements respectable. That is not a claim about Bankman-Fried. It is the charge that the whole procedure — optimise the donation and ignore its origins — may be a more sophisticated version of the very complacency this series began with.
Two costs, then, attach to the fifth position: one that the people inside found and stated publicly, and one that the framework is structurally slow to notice. And we must be exact about what neither cost shows. Neither shows that Singer was wrong. The child is still in the pond; the principle still holds at the water's edge; nothing in these events touches the argument of the 1972 paper. That a movement's chief benefactor proved to be a fraud does not refute a thesis about beneficence — true ideas have been championed by liars before. And the structural critique does not say: do not save the child. It says that saving this child, with money made inside the system that endangered her, is not the clean conscience it appears to be. So the fifth position remains open, as the other four do, and Singer remains standing, as he has stood all series. What is gone is only the last comfortable belief a listener might have carried into tonight — that this position, at least, the position of simply doing what the argument says, was the one that cost nothing. It costs something. Its cost is merely of an unusual kind: not that you will give too little, but that you can give everything the argument asks, calculate correctly, and still find that your hands are not clean, because part of what is wrong with the world is not of a kind that any donation reaches.
Now that all five positions are in view, there is one more thing to see before anyone chooses — something about the whole set that could not be seen until every position was laid out at once. And it changes how all five of them look.
We now have all five positions in view at once, and something can be seen from here that could not be seen earlier. The five are not all of the same kind.
Three of them — the argument proves too much; some of my life is legitimately my own; distance makes some difference — concede the fundamental picture. They agree that morality is, at bottom, a matter of impartial weighing: what is owed, and to whom, assessed from no particular standpoint, with every life counted once. Their disagreement with Singer is about quantity. They accept his terms and dispute the amount. The fourth position does something different. It denies that there is any such standpoint to weigh from. Morality, it says, is not a survey of the world conducted from above it; it is what holds between particular people. That is not an amendment to Singer's picture. It is a rival picture — with, we should remember, its own unpaid cost, the gap where the distant child stands.
And the existence of that fourth position tells us something about how we have been proceeding. Throughout this series, without ever quite saying so, we have treated Singer's argument as the fixed standard — the neutral measure against which every other position was tested, so that each alternative appeared as an attempted escape from it, succeeding or failing. It felt that way because the argument was so clean, and so hard to fault, that it seemed like plain truth, and everything else seemed like special pleading. But the fact that the fourth position can be stated at all shows this was never accurate. Singer's argument does not begin from neutral ground. Its very first step — adopt the point of view of the universe; stand back from your own life and count every life equally — is not a neutral starting point. It is a substantive position. It has a history, which we traced last week: it is the secularised descendant of a religious doctrine, the equal worth of every soul, with the theology removed. And it has a rival tradition just as old — the one your untutored instinct still speaks — in which concern is strongest for the near and fades honestly with distance. So a person who adopts the fifth position, and accepts Singer's conclusion, is not declining to choose among frameworks. They are choosing one: a particular, datable, contestable framework, exactly as much a position as the other four. No such standpoint outside all the traditions has been produced — every candidate offered as neutral, including the view from nowhere itself, has turned out to be one of the views.
That conclusion can sound like despair, and it is not, because it leaves intact the most durable thing this series has produced. You have in fact been using it since the fourth episode. Tonight it should be given its proper name.
You have spent seven weeks with an argument you could not fault and could not live. Every premise looked true; every step followed; and the conclusion describes a life no human being actually leads. Most people, presented with such an argument, say that the conclusion is obviously absurd and that the argument can therefore be ignored — and that is not reasoning; it is refusal dressed up as reasoning. What you can now do instead is this. Singer's argument is valid: the conclusion genuinely follows from the premises. That is precisely why it has been so difficult. But validity is not truth. A perfectly valid argument built on a false premise transmits the falsehood faithfully, step by faultless step, to a false conclusion. So when a valid argument delivers a conclusion no one can live, the right response is neither obedience nor dismissal. The unliveable conclusion is evidence — evidence, pointing back up the chain of reasoning, that one of the premises, all of which looked true, is false. The work is to find which.
That is the first half of the method. The second half is what keeps it honest, and it is the reason none of this has been an easy way out. Establishing that some premise must be false does not tell you which one, and it does not entitle you to pick the convenient one and stop. The unliveable conclusion gives you grounds to reject a premise. It does not give you grounds to stop looking. You must trace the argument, name the exact premise you deny — the demand has a limit; or some of my life is my own; or distance counts; or the impartial framework itself is mistaken — and then accept, explicitly, the cost of denying it. Which premise you reject is not a mood, and not a preference. It is a moral commitment, one you must be able to defend aloud to a thoughtful person who rejected a different premise and is asking you why. "I deny this one; here is what the denial costs; and here is why I will bear that cost rather than one of the others." That sentence is philosophy. The shrug never was. You came into this series able to disagree with a conclusion. You leave it able to do the actual work: find the premise, and accept the cost of denying it.
Let me say, then, where we have arrived, at the end of seven weeks.
We began with a certainty: that the child in the pond and the child overseas are simply, obviously, different cases. On examination the difference was never one thing; it was five unstated premises, each of which had to be defended separately. We then went looking for the flaw in Singer's argument, and what we found was not a refutation and not a proof. It was five positions: four ways of saying where the principle fails, and one way of saying it does not — each of them open, each of them costly, and the costs now stated in full.
I am not going to say which position is the right one. Not from coyness, and not to make you do the work by withholding an answer I possess. There is no answer being withheld. No expert, no philosopher and no broadcaster knows which premise you should deny; that knowledge does not exist to be had from outside your own life. And anyone who claims to know is claiming a standpoint no one in this series has been able to produce — a neutral standpoint above all the positions, from which to award the verdict. Whoever chooses is already inside one of the traditions. Choosing does not escape that fact; choosing is the honest acknowledgement of it.
What you have, in place of a verdict, is this. Seven weeks ago you could feel the difference between the pond and the famine, and you were certain it was obvious. You can no longer call anything in this territory obvious — and that is not a loss; it is the entire gain. Where there was a single unexamined story, there are now five explicit positions, each with its cost itemised, together with the knowledge that one of them must in practice be yours, and that none of them is free. And you have the method: a valid argument is not yet a true one; when reasoning you cannot fault delivers a conclusion you cannot live, that is neither a command nor a scandal but evidence against a premise — and a task, which is to find the premise, and a cost, which is the price of denying it. That method applies far beyond this question, and it will still be usable when every name mentioned in this series has been forgotten.
On the first night I said that I had no verdict, and that I was not speaking from some settled position above the problem, because I bought the shoes too. That is still true tonight. I hold one of these five positions, and I pay its cost, whether or not I say so on air. This series was never instruction. It was two people working through the same problem from the inside, which is the only place it is actually lived.
This has been the first of these series, and there are more to come. The next one turns the examination around. Throughout these seven weeks we have trusted one thing without question: you — the self that weighs the positions, makes the choice and pays the cost. Next time we ask whether that self is as solid as the pond seemed to be: whether the "you" who chooses is a fixed thing at all. But that is another series.
Tonight, the five positions are in front of you, and their costs have been stated in full. So here is the question I have been working towards since the water closed over those expensive shoes, and it is the only one I brought: which premise will you pay to deny — and have you honestly counted what it costs?
Thanks for listening.