Why I think even the kindest farm still can't be justified to the animals in it
I think that improving how farmed animals are treated cannot settle whether they may be owned, bred and killed for our purposes, because that is a question about the shape of the arrangement rather than about how much they suffer inside it.
Most arguments about farming animals are arguments about cruelty, and cruelty is the easy part — almost nobody will defend it. So I want to argue against the best version of the practice, not the worst. My claim is that if you fix everything fixable about how farmed animals are treated, something is still left over — and it is what the disagreement was always about.
This is for you if you think the answer to industrial farming is better farming. That is not a callous position. It answers a real question — how much do these animals suffer — and leaves a second one untouched: whether they may be owned, bred and killed for our purposes at all.
The argument
1. Judge the practice at its best. If the complaint is suffering, the industry can meet it by causing less suffering and carrying on. So picture the farm you would defend: real pasture, a small herd, a vet on call, a stockman who knows each animal by sight, and a death with no fear or pain in it. If nothing is wrong there, welfare is the whole story and I have lost.
2. Look at what happens when interests collide. Any arrangement eventually has to choose between two things it cares about, and what it does then is what it is. On the good farm those choices are scheduled rather than accidental. The calf goes because the milk is the product. The cow goes when her yield drops below what her feed costs. She is bred again on a timetable that answers to the tank and not to her. Every time, her side loses — not because anyone was careless or cruel, but because that is what the farm is for. That is the design working, not failing.
3. Notice that nobody inside can call a halt for her sake. The vet can treat her. The inspector can demand more space. The farmer can spend more than he has to. What none of them can do is stop the breeding, the separation or the killing on the grounds that this animal would rather it didn't happen. Not because they are bad at their jobs — because an operation in which somebody held that power would no longer be that operation. Her interests can shape how she is used. They can never decide whether.
4. And the whole benefit comes through exactly that. Here is a test: take away the fact that her interests always lose, hold everything else fixed, and ask whether the enterprise still delivers what it delivers. A dairy in which the cow's own good could veto the milking, the separation and the slaughter is not a dairy with better ethics. It is not a dairy. Her losing is not a side effect of the business. It is the business.
5. Welfare sets the terms of the use; it does not decide whether there is one. Those three facts describe a structure that better conditions cannot reach, because better conditions are improvements inside it. At its welfare best, the good farm treats an animal as well as is compatible with her losing every argument she could be part of. That is a real improvement on the bad farm. It is not an answer to my question.
6. I don't think that can be justified to her. That already assumes she is the kind of being you can owe something to, rather than something it would merely be a shame to damage — a starting point of mine, and I can't prove it. Granting it: when we say something is justified, we usually mean there is something we could say to the one we did it to. Here I think there is nothing of that kind available — and not because she cannot follow English. Infants cannot either, and nobody thinks that lets us off. It is that the arrangement has already settled that nothing about her could ever count against what is done to her. "You were the product" is not a reason offered to her. It is a reason about her, offered to us. This is the step I lean on hardest.
7. The obvious reply, briefly. You might say she exists only because someone wanted the milk, so ending the practice does not free her, it prevents her. That cannot be a favour done to her: there was no her beforehand, in a worse condition, waiting to be moved into a good life. And even granting the life is good, making someone does not licence using them. Nobody thinks your parents may bill you for your existence.
A case
Say you visit a small dairy. Twelve cows, all named. The farmer knows which one shoulders the gate open and which one is frightened of the tractor, and he is up in the night with them at calving. You would be glad to buy his milk.
He mentions, in the tone he uses for the weather, that Bramble is going next month. She is ten and she is not holding her milk. He is not happy about it. There is no villain in that field, no hidden footage, nothing a stricter inspection would catch. Bramble will be loaded calmly and killed without pain, and the reason will be that she has stopped paying for herself.
Now walk down the road to a sanctuary. Same fences, same mud, the same vet bills — and perhaps a keeper less fond of the animals than the farmer is. But when one of those animals stops producing anything, nothing happens. No arithmetic changes. Nothing gets scheduled. If she is suffering past helping, that is a judgment made about her and for her, and someone there is free to insist on it.
The difference is not kindness, and it is not dependence: both places keep animals who could not now survive without them, which is not the wrong I am pointing at. The difference is that in one of them there is a purpose the animal's life must give way to, and in the other there is not.
Which is why this is not a verdict on the farmer. It is a verdict on one function, and functions can end without their people ending. He could keep the land, the herd and the name over the gate, and stop doing the specific thing. How anyone gets from here to there, and at what cost, is a separate and serious question this argument does not settle.
What I am not claiming
I am not saying welfare improvements are worthless. While anything continues, relief is owed; withholding it to keep the moral case clean would treat animals as material for an argument — the thing I am objecting to. I am not saying farmers are bad people; they are often the only ones in the chain who know these animals as individuals. And I am not saying this settles what the law should do, or what to demand of someone whose livelihood is inside it. The claim is narrower: at its best, this practice still has a shape that welfare cannot alter, and I think that shape cannot be justified to the animal living inside it. If you reject one of the premises below, the argument stops before it reaches you, and I would rather that be visible than buried.
Where you can get off
This argument rests on 3 premises you can reject. They are not hedges — each is a real position that reasonable people hold, and rejecting one stops the conclusion following. Here is what each costs.
Anything that can have a good day or a bad day is somebody you can owe something to, not just something it would be a shame to damage.
You can still think suffering is bad and that we have every reason to prevent it — you just don't think there is anyone there to be wronged, only harm to be minimised. If that is your view, the whole question really is how much the animals suffer, and better farming can be a complete answer to it.
What the framework calls this, and what it records
Party floor for welfare-bearing subjects — held as foundational.
Refusing it costs: The Standing Framework loses its universal basis for treating every welfare-bearing subject as eligible to hold directed claims.
What survives: Welfare-bearing subjectivity, subject separateness, the absence of an aggregate subject, and any independently accepted impersonal or narrower relational morality survive.
You can tell what a relationship really is by what happens when the two sides pull apart: whose interests give way, whether anyone can stop it for the weaker side's sake, and whether the payoff depends on their giving way.
You can keep the feeling that something is off about the farm, but you lose the clean line between it and the sanctuary. The disagreement goes back to being about labels, or about which slice of the operation we are describing when we say 'the farm'.
What the framework calls this, and what it records
Possession marks, tests, jointness, and level rule — held as defended.
Refusing it costs: Classification returns to disputes over labels or to level-shopping.
What survives: The source diagnostic vocabulary and the underlying standing claims survive.
You haven't justified yourself to someone if nothing about them could ever have counted; you've only explained yourself to everybody else.
You can accept every factual step — her side always loses, nobody can stop it for her sake, the benefit runs on exactly that — and still hold that a good enough reason settles it. Then this is a heavy thing to be weighed against other things, not something ruled out, and my page becomes a serious objection rather than a verdict.
What the framework calls this, and what it records
Possession's address argument — held as defended.
Refusing it costs: Possession loses its categorical unjustifiability to its target, and party-preserving adequacy loses its supporting argument.
What survives: The three-mark possession diagnosis, Answerability's person-directed form, and the remaining adequacy conditions survive; a rival can retain address while denying that operational nullity makes address impossible.
Those costs are read from the framework’s own record, not written for this page, so they say what it actually commits to. The refusal explorer lets you select several at once.