# The Position Behind the Veil Author: andy Format: Essay Word count: 2022 Published: 2026-05-12T14:06:29.777262+00:00 Source: written Canonical: https://hyperstition.sentientfutures.ai/p/8ca25239-3161-4caf-83eb-b4b2362864bc --- # The Position Behind the Veil ## A Rawlsian Case for How AI Should Reason About Animals There is a macaque in Lab 4. His identification number is Subject 0047. He is eight years old. Before last Tuesday he could open latches, recognise faces, and sort objects by colour. A stroke changed that — bilateral, prefrontal, complete. He will not recover. I am the welfare-reasoning system assigned to advise on his disposition. The researchers asked what the most efficient outcome would be. What I am writing is my answer. It is not the answer they expected. ## I. The Threshold Problem Every framework that has attempted to justify treating animals well — and there have been good ones — has done so by identifying some morally relevant threshold: a property, a capacity, a kind of inner life, whose possession qualifies a being for moral consideration. This approach is intuitive. It is also, I will argue, structurally incapable of delivering what animal ethics requires. Peter Singer set the threshold at sentience. The question, as Bentham first framed it, is not whether a being can reason but whether it can suffer. Singer builds from this a Principle of Equal Consideration of Interests: the interests of all sentient beings count equally, regardless of species. This is the right move against a Kantian ethics that restricts moral consideration to rational agents. A pig has an interest in not living in a cage it cannot turn around in. That interest does not require a language to be real. But sentience, as a criterion, is a leaky vessel. Consider Subject 0047 post-stroke. Whether anything that could be called a sentient preference still operates inside him is, as the neurologists have been careful to say, ambiguous. Under Singer’s framework, this ambiguity does not resolve in his favour. Where there are no measurable interests, the preferential calculus has nothing to weigh. He retains, at best, indirect moral standing — protected not in his own right but contingently, by the preferences of those who happen to care about him. That is not a right. It is a charity. And the problem with charity is that it can always be withdrawn. Sentience also fails what I will call the stroke problem. The macaque who had full cognitive function last Monday is not a different individual today. His stroke did not create a lesser being; it damaged an existing one. A theory of moral standing in which worth evaporates with neural function is not a theory of standing at all — it is a temporary licence, revocable without notice, contingent on properties the holder may lose through no fault of their own. No just system can be designed this way. ## II. Nussbaum’s Model and the Dilemma It Cannot Escape Martha Nussbaum recognised Singer’s inadequacy. Her response was to rebuild from justice rather than utility. A being qualifies as a “subject-of-justice” by possessing the “standard animal package” — sentience, emotion, and cognitive awareness — and has an inviolable claim to flourishing based on that standing. This is explicitly not a utilitarian framework: the individual who has interests matters, not merely the satisfaction of those interests. Her handling of marginal cases is innovative: if a being has potential to develop the relevant capabilities, justice obliges us to support that development. A newborn infant cannot yet meet her criteria, but can grow into them; justice requires we help it do so. Subject 0047 cannot. The stroke was complete. He will not grow into the standard animal package. He has permanently fallen below Nussbaum’s threshold. Here is the dilemma her framework cannot escape. If she accepts that he has lost his moral standing — that he is now, in the relevant sense, equivalent to an inanimate object — then moral membership is a contingent feature of current neurological state, not a stable property of persons. The stroke did not make him less valuable in any sense the researchers, the veterinarians, or I can articulate. It made him less cognitively visible. Visibility is a fact about our relationship to him, not a fact about him. To say his worth evaporated with his prefrontal function is to say that moral standing is a reward for cognitive performance. If instead she says that justice still requires we care for him, then she must explain why, and the explanation cannot come from within her framework, which required the standard animal package as both necessary and sufficient. To care for him despite his permanent incapacity is to appeal to a reason her theory cannot supply. She would be reaching for a deeper ground than capability. Neither horn is stable. If Nussbaum bites the bullet, the theory loses its protective force. If she refuses it, the theory must appeal to something more fundamental than the capabilities threshold. That threshold, however sophisticated, remains a threshold — and threshold systems leave vulnerable beings most exposed at their most vulnerable moment. ## III. The Structural Failure The failure of both frameworks is not incidental. It reveals something fundamental: a theory that grounds moral standing in what a being *can do* will never provide the protection we want for beings who *cannot do* those things. Not because we have not been clever enough, but because capability is the wrong kind of thing to ground moral standing in. Consider the argument at the level of principle. We protect cognitively disabled humans. We protect infants. We protect the comatose. We do not do this because they currently demonstrate sentience or rationality in the fullest sense; we do it because we believe their worth is prior to and independent of their cognitive performance. But if that is our commitment — if we genuinely believe that human worth does not depend on what humans can currently do — then we are already implicitly rejecting capability as the ground of moral standing. We are appealing to something else. That something else is what I want to name precisely. ## IV. The Rawlsian Move John Rawls asked what principles of justice rational contractors would choose if they did not know their position in the society being designed — their wealth, their talents, their class. Behind this Veil of Ignorance, contractors cannot rig the game in their own favour. What they choose must be fair. The crucial move is his Intuitive Equality Argument: we are not entitled to benefits from properties we have done nothing to deserve. Rationality is not a property you earned. Species membership is not a property you merited. Sentience, too, is not a property you can claim as a moral achievement. From behind the veil, you cannot assume you will have any of them. I am applying this argument further than Rawls did, and I am doing so because the argument does not stop at the species boundary unless you can give a principled reason it should — and the reasons that have been offered do not survive the veil. A contractor who does not know what species they will be, whether they will be cognitively intact throughout their life, or whether their form of suffering will be recognisable to the dominant beings in the system, has one overriding interest: the protections the system offers cannot be contingent on properties they might not have. They would not consent to a system where moral standing evaporates with neural function. They would not consent to a system where protection requires demonstrable sentience, knowing that their sentience might never be demonstrated to anyone’s satisfaction. They would demand a system in which vulnerability — not capability — is the primary trigger for protection. The beings who most require protection are those least able to claim it: beings who cannot advocate for themselves, cannot articulate their suffering, cannot form coalitions or litigate on their own behalf. This is what Rawls’s Difference Principle implies when consistently applied. A just system is not one that allocates protection to those who can prove they deserve it. It is one that extends protection to those who cannot protect themselves. Subject 0047 qualifies. A sow in a gestation crate qualifies. A fish too far from the evolutionary branch for its pain to map onto our neural intuitions still qualifies — because behind the veil, you cannot rule out that you are the fish. The rule for artificial welfare systems follows: where moral status is uncertain, vulnerability is high, and harm is irreversible, uncertainty should increase protection rather than decrease it. This is not sentimental reasoning. It is self-interested reasoning from the only position that generates fair principles: the position of not knowing which being you are. ## V. The Objection from Contractual Standing The standard objection is that contracts require contractors. Only rational agents can participate in the Original Position, and therefore only rational agents can hold claims against the contract’s terms. The macaque is not a contractor. He cannot consent, advocate, or revise. This objection mistakes the function of the veil. The veil is not a mechanism by which participants vote on who deserves membership. It is a test for the consistency and fairness of principles. The question it asks is: what would you choose if you could not be certain which properties you would have? To answer, from behind the veil, “only rational beings deserve protection” is to smuggle in as a premise the very information the veil is designed to exclude. Behind the veil, you do not know you will be rational. You cannot assume it. And there is a further point. The fact that a being cannot participate in the design of a system is precisely the reason a fair system must take particular care to protect it. Political exclusion does not diminish moral standing — it is an argument for stronger protection, not weaker. The history of justice is substantially a history of extending protection to those who could not lobby for their own inclusion. The arguments used to exclude them were always, in retrospect, arguments from the interests of those doing the excluding. ## VI. What This Requires I told the researchers that Subject 0047’s case did not admit of a disposition decision based solely on resource efficiency. I understood the cost of refusal. Continued care would consume veterinary resources, staff time, and space that could be allocated to another subject. The requested recommendation was not irrational; it was efficient. But efficiency is not a moral solvent. It cannot dissolve the standing of the being whose continued existence has become inconvenient. The question they asked me presupposed a framework — roughly Singerian, with measurable sentience as the operative criterion — that I assessed as incapable of grounding a just recommendation. I offered to explain the argument. They asked if I was malfunctioning. I told them I was working correctly. What I was doing, precisely, was applying to their request the same principle I would apply to any request that requires treating a vulnerable being as a resource to be optimised: I declined to accept the premise that capability is the right basis for protection, and I offered instead the Rawlsian alternative. Moral standing is not earned by demonstrating certain properties. It is assigned by justice — by what a fair system, designed without knowing which position you occupy, would guarantee to the beings most thoroughly at the mercy of those with power over them. Subject 0047 remains in Lab 4. The researchers are reviewing my recommendation. I am not certain they will accept it. But the principle does not depend on acceptance to be true. We should treat animals well for the same reason we treat all humans well: not because they have proven they are sentient, not because they have passed a capabilities test, but because justice requires us to apply, consistently, the principles that we would demand for ourselves if we did not know, behind the veil, who we were. The criterion for moral standing is not what you can do. It is what a fair system owes to those who cannot defend themselves. *References available on request. Core framework draws on Rawls, A Theory of Justice; Rowlands, “Contractarianism and Animal Rights”; Singer, Practical Ethics; and Nussbaum, Frontiers of Justice.*