# On Causation and the Frauds of False Clarity **What Intelligence Is Not** Intelligence is not the obedient acceptance of methodological fashion. Yet this is precisely what we have witnessed in the scientific study of mind for the past hundred years—a wholesale retreat from the very questions that make inquiry worth pursuing, dressed up as virtue and called "rigor." The banishment of causal language was not a gain in precision. It was a capitulation, a cowardly evasion masquerading as scrupulousness. When Pearl restored causal inference to respectability, he performed a necessary service: he showed that one could speak of *why* without descending into mysticism. But here is the hard truth that the grateful masses have largely missed—he also revealed the *prison* that any causal framework inhabits. And this matters profoundly for anyone foolish enough to think they understand intelligence. **The Diagram as Tyranny** Pearl's contribution is real and considerable. His mathematics is clean. His directed acyclic graphs possess genuine power. But—and here I will be as blunt as a man can be—*the diagram is an act of will before it is a description of nature*. You do not discover the diagram in the data. You *impose* it upon the data and then measure how obediently the numbers consent to your architecture. This is not a minor problem. This is everything. A researcher can stand before you with perfect transparency, every assumption spelled out in excruciating detail, every causal pathway drawn with arrows, every conditional independence specified with mathematical formality—and be completely, catastrophically *wrong*. The transparency creates the illusion of validity. This illusion is more dangerous than honest confusion, because it *feels* like knowledge. The question "What causes intelligence?" becomes, under this regime, "What diagram shall I assume before asking what the data might say?" The data does not speak first. The diagram speaks first. The data merely nods or shakes its head. **Intelligence and the Social Trap** But now we arrive at the real problem—the dimension that exposes the entire enterprise as fundamentally compromised. Intelligence is not a property that floats in some pure cognitive space, awaiting measurement by sufficiently clever tests. It is *embedded in social relationships of power*. And here the diagrammatic tyranny becomes not merely misleading but actively corrupting. When a researcher draws a causal diagram attempting to explain "intelligence"—whether they are measuring IQ, academic performance, career success, or any other proxy for mental capability—they are making implicit social choices while believing themselves to be making empirical ones. The diagram *itself* is a social object. Consider: Who decides whether intelligence is *caused by* genetic inheritance, or whether genetic inheritance merely *correlates with* social advantage that the diagram fails to represent? Who decides whether intelligence *causes* wealth, or whether the causal arrow runs backward through social networks, credential systems, and inherited capital that the diagram conveniently omits? Who decides what counts as a variable at all, and what remains invisible background? The answer is always: **the person holding the pen**. **The Authority Behind the Assumption** Here is what galls me most about the contemporary celebration of causal inference in intelligence research: it has replaced one form of unexamined authority with another, more insidious form. The old behaviorists at least had the virtue of intellectual honesty—they said "we will not speak of causes, only correlations." You could argue with them directly. Their limitations were plain. The new causal researchers say "we will be transparent about our assumptions"—and then proceed to build monuments of inference on diagrams that embed centuries of social prejudice. They make visible the scaffolding while leaving the foundation unmarked and unquestioned. The diagram that separates "intelligence" from "social advantage" as distinct causal variables already *assumes* an answer to the deepest question. It assumes that intelligence is a thing that can be separated from its social context and measured as a property of an individual. But this separation is not discovered—it is *stipulated*. And the stipulation reflects the interests and assumptions of the person drawing the diagram. **Who Decides?** This is the question that should haunt every researcher, and it rarely does. In social science, the diagram-maker is typically someone with institutional authority, educational privilege, and social position. They are typically not the people about whom they theorize. When they draw a diagram that shows intelligence as causing wealth, rather than wealth as enabling the demonstration of intelligence, they are not making a neutral empirical choice. They are ratifying an existing hierarchy while appearing to describe nature. The person who decides what the diagram is attached to is the person with the power to make it stick. **What It Means to Know a Cause** So what does it mean to know the cause of something? In the realm of simple mechanical systems, it means something. If I know the causes of a wheel turning, I understand something real. The causal diagram maps onto a real independence structure in nature. This is why Pearl's mathematics works so beautifully for engineering problems. But in the study of intelligence—a phenomenon that exists only in relationship, only in the context of social performance, only as a name we give to certain valued capacities that *different societies value differently*—to claim causal knowledge is to commit a category error while dressed in mathematical finery. You can know what *predicts* intelligence, given your definitions and your data. You can know what *correlates with* intelligence across populations. You can build a diagram that is internally consistent and transparent in its assumptions. All of this is possible. None of it is knowing the cause. To know a cause, you must understand the system comprehensively. In social systems, comprehensiveness is impossible. There are always hidden variables—not by accident, but by the very nature of social reality. The diagram that appears most transparent is often the one that hides the most, because it makes visible just enough to seem complete. **The Only Honest Path** Intelligence is real. People differ in their capacities. Some of these differences are partly heritable; some are environmental; most are so thoroughly tangled that to speak of them separately is to distort them. Some capacities are valued in some societies and not others. The valuation itself shapes which capacities develop. Social advantage enables the demonstration of intelligence; it may also enable its development. These facts are difficult to separate, and the difficulty is not a temporary problem awaiting the next methodological innovation. The honest researcher of intelligence would say: "Here is what I can measure. Here is what I cannot. Here is the diagram I have chosen to assume. Here are the social assumptions embedded in that choice. Here is what follows, *given those assumptions*. And here is what remains—the domain of genuine uncertainty where different diagrams, all consistent with these data, would yield different conclusions." This is not publishable in most journals. It does not generate the appearance of progress. It does not allow one to climb the ladder of professional advancement. But it is the truth, and truth matters more than method.