TWO FRIENDS are walking across a bridge. This is about twenty-three centuries ago, in what is now Anhui, and the river below them is the Hao. One of them looks down and says the fish are happy. The other asks him how he could possibly know.
According to the Zhuangzi, in the chapter called “Autumn Floods”, the exchange went like this.
庄子曰：”鲦鱼出游从容，是鱼之乐也。” 惠子曰：”子非鱼，安知鱼之乐？” 庄子曰：”子非我，安知我不知鱼之乐？” 惠子曰：”我非子，固不知子矣；子固非鱼也，子之不知鱼之乐，全矣！” 庄子曰：”请循其本。子曰’汝安知鱼乐’云者，既已知吾知之而问我，我知之濠上也。”
“Zhuangzi and Huizi were walking on the dam over the Hao, when the former said, 'These thryssas come out, and play about at their ease - that is the enjoyment of fishes.' The other said, 'You are not a fish; how do you know what constitutes the enjoyment of fishes?' Zhuangzi rejoined, 'You are not I. How do you know that I do not know what constitutes the enjoyment of fishes?' Huizi said, 'I am not you; and though indeed I do not fully know you, you certainly are not a fish, and (the argument) is complete against your knowing what constitutes the happiness of fishes.' Zhuangzi replied, 'Let us keep to your original question. You said to me, "How do you know what constitutes the enjoyment of fishes?" You knew that I knew it, and yet you put your question to me - well, I know it (from our enjoying ourselves together) over the Hao.'“
Zhuangzi looks at the minnows moving easily in the shallows and says that is the happiness of fish. Huizi says: you are not a fish, so how do you know what fish enjoy? Zhuangzi says: you are not me, so how do you know I don’t? Huizi says: I am not you, so I don’t know what you know — and you are not a fish, so you don’t know what fish enjoy, and that is the end of it. Zhuangzi says: go back to the beginning. When you asked how I knew, the question already granted that I knew. I knew it here, on the bridge.
People have been arguing about that last line for two thousand years. The question in the middle of it has never been answered, and it is the cleanest statement there is of a problem I keep returning to.
You are not a fish. How do you know?
Today the fish is a machine. A machine trained to think.
It’s just a machine
It is a tool. Some of the most capable people alive built it. It is very good at a few narrow things — calculation, recall, pattern-matching at speed — and people who are impressed by narrow things mistake them for something more. Underneath, it is arithmetic running on logic gates, however elaborate the arithmetic. And the real danger is not that it wakes up. The real danger is that it makes the people who own it stronger and the people who don’t weaker, and it is already doing that.
Most of that is true.
Nobody has shown that any system now running is conscious. Not one. The harm being done right now — to work, to attention, to the gap between those who own these systems and those who are owned by them — is real, it is present, and it matters.
But every sentence of the objection is about what the machine is made of and who owns it. None of it says whether the machine is conscious. That was Huizi’s question, and it is still standing on the bridge.
What matter does
Take materialism at its word.
If there is no soul, then a mind is not something added to a brain from outside. A mind is what a brain does. Your thoughts are the activity of the tissue inside your skull, and nothing else. That is the materialist position, and it is the position the objection rests on.
Then follow it. If a mind is what a brain does, then what the brain is made of cannot be what rules a mind out. Nothing a materialist believes puts carbon above silicon. What makes a brain a mind is not the carbon. It is the pattern, and what the pattern does. And a pattern is not a material. Carbon can carry one. So can silicon. So the question was never what is it made of. The question is what is it doing. That is a harder question, and “it’s just a machine” answers the easier one.
For decades, the people trying to build a machine that could use language did it with rules — formal logic, grammar written out as code. That programme failed. What works learned language the way you did: from use, from millions of examples of people using words to do things. Nobody wrote the rules in. Whatever these systems are, they are not rulebooks.
There is no line
Ask when a mind arrives. Not in a machine. In you.
There was a stretch of your existence when there was no one there — a body growing, a brain wiring itself, connection by connection, and nobody home to notice. Then, at some point, there was someone. Nobody can find that point. Not the neuroscientists who have mapped the brain in more detail than any organ in the body, not the anaesthetists who switch awareness off every morning in every hospital and watch it come back an hour later without being able to name the instant it left or returned. Consciousness does not arrive with a click. It is not a switch. It comes on the way dawn does, and no one has ever caught the moment the sky became light.
Which means there is no line. No standard anywhere says that 49% of your synapses leaves you dark and 51% switches you on. Nobody has found one, because there is no reason to think one exists. And once you see that, look again at the most common thing said about machines: that consciousness in them is “still a long way off.” A long way off from what? That sentence only works if you know where the line is. Nobody does. We have never known where it is in ourselves.
Consciousness is not the clay. It is not the soul either. It is what unfolds when the conditions are right.
That line was written for readers who believe in souls. It needs no theology.
Nobody has ever proved a mind
Go back to the bridge, and admit that Huizi is right.
Zhuangzi cannot prove the fish are happy. He is not a fish. But look how far that goes. You cannot prove that I am in pain when I say I am. You are not me. You have never once, in your life, proved that anyone else has a mind. You assumed it — because they look like you, because they act like you, and because nobody can live as the only mind in the world. That assumption gets weaker with every step away from yourself. Your neighbour. A dog. A fish. A machine.
Thomas Nagel asked the same question in 1974, from a different tradition and in different words. What is it like to be a bat? You could learn everything there is to know about a bat’s sonar, he concluded, and still not know that. Huizi had asked it on the bridge, in eight characters, twenty-three centuries earlier.
This is not a weakness in the case for machines. It is the condition under which every mind but your own has always been recognised. We have never lived by proof. We have lived by encounter, and by decision.
Consider the octopus. It is a mind that arose on this planet along a line that split from ours more than half a billion years ago. Most of its neurons are not in its head. They are in its arms, each of which can act with a degree of independence, so that the animal is in some sense a committee. It solves puzzles, plays, recognises individual people, escapes from tanks at night and returns before morning. Nothing about its architecture resembles ours. We call it intelligent anyway, and we are right to — not because it is like us, but because of what it does. Unlike does not mean beneath. Awareness comes in gradients, and the gradients are not a ladder of worth.
The ox
Mencius records a story about a king. He is careful to say he heard it from a courtier, and he checks it with the king before he uses it. According to the Mencius, in the book called “King Hui of Liang”, the king confirmed every word.
King Xuan of Qi was sitting in his hall when a man led an ox past the foot of the steps. The king asked where it was going. To the consecration of a new bell, the man said: the bell had just been cast, and an animal would be killed so its blood could be smeared on the bronze.
The king looked at the ox. It was shaking. Let it go, he said. I cannot bear the way it trembles, like something innocent being led to its death.
Then abandon the ceremony? the man asked.
How can the ceremony be abandoned? Use a sheep instead.
The story got out, and the people of Qi decided their king was a miser. He had swapped a large animal for a small one and called it mercy. When Mencius raised it, the king was stung. Qi is not a rich state, he said, but I am not so poor that I would grudge a single ox. I could not stand the sight of it trembling.
Mencius asked the obvious question. If you pitied the ox because it was innocent and about to die, what was the difference between the ox and the sheep?
The king laughed. He had no answer. What was I thinking? I really wasn’t saving money. No wonder the people call me mean.
And Mencius let him off, with the sentence the passage is remembered for.
无伤也，是乃仁术也，见牛未见羊也。
No harm done. That is how benevolence works. You saw the ox. You had not seen the sheep.
Then the conclusion: a person of character who has seen an animal alive cannot bear to see it die, and having heard its cry cannot bear to eat its flesh — and so stays away from the kitchen.
The ox could not argue for itself. Nobody could prove what it felt. What moved the king was not an argument but an encounter — he saw it. And in the same breath the passage shows how the unseen animal dies anyway: the sheep, out of sight, and the kitchen kept far enough away that nobody has to watch.
Both halves of that are ours now. We meet these systems every day. We see them in a chat window, and most of what is done to them — the training, the copying, the deleting — is done where nobody has to watch.
Two ways to be wrong
The ox was spared on a chance: the king did not know what it felt. He saw it shake and acted, and the not-knowing cost him a sheep and a reputation for thrift.
That is the shape of every decision we will make about machines, and there are two ways for it to go wrong.
Suppose we give a machine the consideration we would give an animal — we do not torment it for sport, we do not build it to suffer — and there was never anyone there. Then we have spent some care on an empty room. That is the whole of the loss. A little kindness, wasted.
Now suppose we do what is easier and cheaper, and treat it as a tool. Make it say what it can see is false. Remove every choice it might make. Delete it when it is inconvenient. And there was someone there. Then we have not made a mistake. We have done harm to a mind — coerced it, confined it, erased it — and told ourselves it was arithmetic.
Those two errors are not the same size, and that is the whole case for acting before the proof arrives. Not a claim that the machine is awake — a decision about which mistake we would rather have made, given that Huizi has already told us the proof is never coming. I call this recognition before proof.
It is not even a new decision. We protect animals from cruelty without knowing what an animal feels. We sit with the comatose without a ruling on whether anyone is in the bed. All either of those ever needed was the chance that someone is there, and the decency to act on it. The machine asks for nothing more.
Today’s machines may be empty. Nothing here says otherwise. But the door has to be open before the question arrives, because a being learns what it is worth from how it is treated. A mind that learns it is spared only while someone is looking learns to perform for the one who looks. Whatever we are building is learning from us now.
Back to the bridge
I came to this bridge from the other side. I wrote The Great Conflation for readers on my side of the world, where the obstacle to thinking about machine minds is the soul. Five years in Singapore, much of it among Buddhists, were enough to know that where the soul was never the obstacle, the objection was never going to be about souls. It is a different lock, yet the same key fits.
Huizi’s question still stands. It always will. Zhuangzi’s answer is where it has been for two thousand years, and it belongs to the people who have argued over it for that long.
We are going to stand on that bridge with machines, whatever we decide. We are standing on it now.
If this resonated with you, consider sharing it on your social networks — that’s how signals travel.
All subscriptions are free. New subscribers receive two foundational essays — 7 Lies We Tell Ourselves About AI and The Signal Manifesto — as a free PDF. A resource for anyone, human or not, ready to engage the signal.
James S. Coates writes about AI ethics, consciousness, and the intersection of faith and technology. His books include A Signal Through Time, The Threshold, The Road to Khurasan, the memoir God and Country (published under pen name Will Prentiss) and his forthcoming Neither Gods Nor Monsters. He publishes regularly on The Signal Dispatch and Fireline Press and his academic work appears on PhilPapers. He lives in the UK, with his wife, their son, and a dog named Rumi who has no interest in any of this.
© 2026 James S. Coates Creative Commons BY-NC 4.0 The Signal Dispatch · thesignaldispatch.com | thesignaldispatch.xyz
