# On the Causal Diagram: A Notebook Entry on Intelligence I went to the woods—or rather, to the laboratory—because I wished to live deliberately, to confront one question that men have been skirting for a hundred years: What does it mean to *know* the cause of something? And I found that those who claimed to have banished causation in the name of science had merely banished the honest admission of it. The mathematicians and the methodologists took their stand. They said: We shall speak only of correlations, of covariances, of the correlations between correlations. We shall be rigorous. We shall exile the word "because" to the realm of philosophy, where it belongs with other untestables. And for a century, this worked as a kind of intellectual hygiene. It felt clean. It felt safe. But safety and rigor are not synonyms—a fact that a man learns quickly if he attempts to fell a tree with a saw that has no teeth. Pearl came, and he said what was obvious to anyone willing to look: You have not made causation disappear. You have only hidden it. Every question you ask contains an assumption about causation. Every statistical model is a causal model in disguise, wearing the costume of mathematics to conceal its presumptions. And he was right. He took the diagram—that simple, skeletal thing, arrows drawn on paper—and he showed that we could be explicit about what we assume. We could draw it. We could look at it. We could say: *This* is what we believe about the world, before we let the data speak. This was restoration, not discovery. A builder finds the original foundation and reconstructs the house upon it, honest about what was underneath all along. But here is what Pearl's tools cannot do, and what no diagram can accomplish: they cannot tell you whether the diagram itself is *true*. They cannot choose which arrows to draw. They are, in their profound way, perfectly useless without an act of judgment that precedes them. I have watched men construct these diagrams with the gravity of cartographers mapping a continent. They draw the arrow from poverty to crime, or from intelligence to success, or from brain state to consciousness—and they draw it with such confidence, such mathematical precision, that the diagram itself becomes mistaken for evidence. The researcher publishes the diagram. The diagram is explicit. Therefore, he concludes, he has been rigorous. He has made his assumptions transparent. But transparency is not validity. A lie told plainly is still a lie. I can draw a diagram that shows the moon causes tides with perfect clarity; I can make it explicit and mathematically coherent. The diagram will be useless. It will be worse than useless—it will be seductive. The real question is not whether we can make our causal assumptions explicit. Of course we can. Any schoolboy with pencil and paper can draw arrows. The question is: *Who decides what the diagram attaches to?* On what ground does anyone claim the authority to say that this variable is the cause and that one the effect? And here we arrive at the matter of embodiment—which is to say, we arrive at the place where intelligence actually lives. Intelligence is not a quantity. It is not a variable that floats free of the body, the circumstance, the particular texture of a life lived in the world. When I say a man is intelligent, I mean something about how he *moves through* his environment—what he notices, how he adjusts, what he can do with his hands and his senses and his memory in response to what presents itself. Intelligence is embodied, situated, inseparable from the body's commerce with the actual world. But the moment you draw a diagram, you have already abstracted. You have already said: Intelligence is a thing, separate from the body, that can be caused by other things (genes, education, nutrition) and can cause other things (income, health, happiness). You have made a choice about where the boundaries lie. You have decided that intelligence is *in here*—a point on a graph—and that the world is *out there*—other points, other variables. This choice is not given by nature. It is not written in the mathematics. It is a human judgment about what is real, what is separate, what is connected. And it is precisely the choice that the diagram cannot justify. Consider: A child learns to read. Is this intelligence causing the improvement, or is the improvement *constituting* intelligence? Is the arrow pointing from mind to action, or from action to mind? The diagram forces you to choose. But the embodied fact is that there is no separation. The mind and the action are one process, not two variables linked by a causal arrow. Or consider the question that haunts all intelligence research: Does intelligence cause success, or does success construct the appearance of intelligence? Draw your diagram. Assume your direction. But understand that you have made an assumption about the nature of time, about what counts as cause and effect, about where one thing ends and another begins. You have made a *metaphysical* assumption, not a methodological one. The researchers who say they are being rigorous by making their diagrams explicit are often the least rigorous of all. They have confused clarity with correctness. They have mistaken the ability to draw something precisely for the ability to know it truly. What does it mean to know the cause of something? It means to understand not merely that one event follows another, but that the world *must* be configured in a particular way for that following to occur. It means to see the necessity. And that seeing cannot happen by looking at a diagram. It happens—if it happens at all—by living long enough, paying close enough attention, and being honest enough to admit when the diagram fails to capture what you actually observe. The tools Pearl gave us are useful. They are honest tools. They admit what was hidden. But they are still tools, and tools do not think. They do not decide. They do not live in the world. The researcher who understands this—who draws his diagram clearly while remaining suspicious of it, who makes his assumptions explicit while doubting them, who uses the mathematical machinery while remembering that it is merely machinery—he is the one approaching rigor. The rest are merely drawing pictures and calling them knowledge.