# On the Causes of Things, and Who Pretends Not to Know Them **A Meditation on Pearl, Pretense, and the Tyranny of Diagrams** Intelligence, I have long maintained, is the capacity to see things as they actually are—not as we wish them, nor as our methodological prejudices demand. It is the refusal of comfortable falsehood. Yet I find myself contemplating a peculiar modern perversity: scientists have spent a century *expelling the very word for reality* from their vocabulary, then congratulating themselves on the purity of their exile. They called it rigor. I call it cowardice dressed in mathematics. Judea Pearl—a man of genuine intellectual courage—has restored to us the language of causation. He has given us tools: directed acyclic graphs, do-operators, counterfactual reasoning. These are not small things. They are the recovery of human speech from the sterile prison of correlation. But here is the rub, sharp and unavoidable: **Pearl has merely transferred the problem. He has not solved it. He has made it visible.** This is progress. But let us not mistake progress for salvation. ## The Diagram as Confession When a researcher produces a causal diagram, they are not—despite what they may tell you at a faculty seminar—discovering the skeleton of nature. They are *confessing*. They are saying: "This is what I believe about how the world hangs together. This is the architecture I am willing to bet my reputation upon." This confession is *necessary*. Transparency is good. I have no quarrel with a man who shows his work. But transparency is not validity. This is the cardinal error of our age, committed daily in a thousand research papers and policy documents. A researcher can be—and frequently is—completely explicit about a model that is completely wrong. You can draw your diagram with perfect clarity, specify your assumptions with crystalline precision, apply Pearl's machinery with flawless technical competence, and arrive at conclusions that are *false*. The tools do not protect you from the fundamental human predicament: you must *choose* the diagram before the data can speak. And your choice reflects not what nature demands, but what you believe, what you fear, what you are paid to believe, what your colleagues believe, what your culture believes. The diagram is a mirror. We think it is a window. ## The Question That Cannot Be Evaded Here is the question that neither Pearl nor his disciples adequately confront: **Who decides what the diagram is?** Not the data. Data is mute until you speak to it. Data will not contradict your diagram; it will only fail to confirm it. And in that silence, in that mere failure to confirm, lies the entire problem of human knowledge. The diagram is decided by: - The researcher's prior convictions - The funding source's priorities - The profession's established wisdom - The culture's blind spots - The era's fashionable assumptions - The researcher's honest confusion about matters genuinely obscure These are not separable. They interpenetrate. And they are *especially* consequential when we turn to questions of intelligence itself. ## The Social Dimension, Where Diagrams Become Dangerous But now we arrive at the matter that stirs my blood: the social dimension. Here the diagram ceases to be a technical convenience and becomes an instrument of power. Consider: We ask, "What causes intelligence?" We produce a diagram. We draw arrows pointing from genetics, from environment, from nutrition, from schooling, from motivation. We apply Pearl's apparatus. We calculate. We publish. We declare victory. But notice what we have done. We have *fixed* a diagram—a representation of cause—before examining the most consequential fact: that intelligence itself is not a thing that *sits there* waiting to be measured. It is a *social construction*, saturated with judgment, embedded in power relations, and fundamentally contested. When we ask "What causes intelligence?" in a particular society, at a particular moment, we are not asking a neutral causal question. We are asking: "What explains why some people possess the qualities that *this society* has decided to call intelligent?" The diagram you draw will inevitably embed the prejudices of your age. If your age believes that intelligence is primarily a matter of processing speed, your diagram will reflect this. If it believes intelligence is resilience, or creativity, or obedience, or the capacity for irony—your diagram shifts. The causal arrows remain, but they point toward different things. And here is the democratic scandal: **the people whose lives are explained by the diagram rarely get to vote on what the diagram contains.** A psychometrician in 1920 drew a diagram that placed certain populations lower on the intelligence scale. The diagram was explicit. The assumptions were stated. The mathematics was—for that era—rigorous. The causal story was clear: these groups possessed less of what intelligence is. And this diagram, this transparent confession of belief, was then used to justify immigration restriction, forced sterilization, and the organization of society itself. The diagram was not wrong because it was vague. It was not refuted because its assumptions were hidden. It was wrong because the society that created it had built its premises upon a foundation of power and wished to preserve that power. **No amount of causal transparency protects you from this.** ## What It Means to Know a Cause Let me be direct: to know the cause of something is not primarily a technical achievement. It is a *moral and political one*. To know that X causes Y, genuinely, requires: First, that you have imagined the full range of possible causal mechanisms, not merely the ones your discipline permits. Second, that you have submitted your diagram to the judgment of those most affected by its claims—which means, in the case of intelligence, to the communities whose capacities it purports to explain. Not as subjects in an experiment, but as judges of whether the diagram captures something true about their own lives. Third, that you have held your diagram *lightly*—that you have remained willing, even eager, to discover that you were wrong about the fundamental structure. That you have not fallen in love with your own architecture. Fourth, and most importantly, that you have asked: *What interests does this particular diagram serve?* Whose power does it consolidate? Whose resistance does it explain away? In the domain of intelligence, this last question is not incidental. It is everything. ## The Diagram Remains a Choice Pearl has given us better tools for making our choices explicit. This is valuable. But do not mistake the improvement of tools for the solution of the problem they address. The diagram you choose to draw about the causes of intelligence will be *your* diagram. It will bear your fingerprints, your assumptions, your hopes, your fears. It will be more transparent than previous diagrams, yes. But it will not be any less yours. And if you are honest—if you possess what I would call genuine intelligence—you will acknowledge this constantly. You will hold your diagram as a *provisional map*, not a final territory. You will remain alert to the possibility that the questions you are not asking are more important than the ones you are. You will notice, with something approaching shame, how easily your scientific apparatus aligns itself with the prejudices of your employers. This is not a counsel of despair. It is a counsel of honesty. We *can* know causes. We can reason about them. We can build better models and test them against reality. Pearl's work genuinely improves our capacity to do so. But we cannot escape the fundamental fact: **before the data can speak, someone must decide what it will be allowed to say.** In the case of intelligence, that someone has historically been the powerful, deciding about the weak. Transparency has not stopped this. Better mathematics will not stop it. Only a sustained political and moral vigilance—a refusal to accept any diagram that serves the interests of those who already have power—can do that. This is the form of intelligence that matters: the capacity to see the diagram itself as an object worthy of suspicion, and to ask, always, *Who benefits from this story?*