# On Intelligence, Causation, and the Comfortable Lie of Rigor What is intelligence? I shall tell you plainly: it is the ability to see *through* things — not merely to describe their surface, but to grasp what makes them move. And this is precisely what we have, for a century, been trained to *refuse*. The history is shameful. Science, grown fearful of metaphysics, adopted what it called rigor: a deliberate amputation of the language of cause. We were to be content with correlations, regressions, the mere notation of how things rise and fall together. We wore this constraint like a badge of honor. *This* is scientific! we declared, as if the refusal to ask "why" were somehow more honest than asking it. It was cowardice dressed as method. Judea Pearl saw this plainly and would not accept it. He restored to science what it had shamefully discarded — the vocabulary of causation, the diagram, the structural equation. And he built something remarkable: a calculus of cause, mathematically rigorous, provably correct *given your assumptions*. This is genuine progress. This is intelligence put to work. But here is where the matter grows dangerous, and where most practitioners of this new science show themselves to be fools: **The diagram is not discovered. It is *assumed*.** This is not a flaw to be corrected. It is the permanent condition of human knowledge. And the moment you make that assumption explicit — the moment you draw your causal graph on paper — you have committed an act that is *entirely* social, political, and contestable. You have taken a stand. You have said: *This is how the world hangs together.* And you have done so before the data could possibly tell you whether you are right. --- ## The Tyranny of Transparency We are now told that *transparency* is the solution. Publish your diagram! Make your assumptions visible! Then we shall all see whether you are honest! This is sentiment masquerading as epistemology. A researcher can be — and frequently is — *entirely explicit* about a model that is *entirely wrong*. Transparency does not confer validity. A clearly drawn diagram of a false world is still a false world. The man who carefully maps the territory of his own delusion has not thereby escaped delusion; he has merely made it legible. The pretense that visibility solves the problem is particularly pernicious because it flatters us. We can see the assumptions, therefore we are safe. We can object to them, therefore the process is democratic. We can publish them, therefore we are rigorous. None of this follows. What we have done is shifted the burden of correction from the researcher to the reader — and called this *progress*. --- ## The Question of Who Decides But the deeper problem lies here: **Who decides what the diagram represents?** In the natural sciences — when one studies, say, the motion of particles or the properties of chemical compounds — the diagram often corresponds to something we can, with sufficient effort, inspect directly. The world *speaks back*. It resists. It forces revision. A false theory of gravity will not keep the planets in their orbits. But when we turn to human society, to the causes of poverty, inequality, intelligence itself, to the mechanisms by which power distributes itself — here the diagram is *never* merely a transparent representation of natural fact. It is an interpretation. It is *shot through* with values, histories, interests, and the unexamined prejudices of those who draw it. A researcher studies the causes of academic performance. She draws a diagram. Intelligence is here, at the center — a fixed quantity, largely heritable, distributed unequally. Socioeconomic status is here, upstream. Family structure is here. Effort is here. And from this diagram, conclusions flow: certain populations are naturally more capable; intervention is therefore futile; selection mechanisms are justified. All of it is perfectly transparent. All of it is explicitly reasoned. And all of it rests on a diagram that was *assumed* — not discovered — and that embeds within itself assumptions about human nature, social possibility, and justice that were never submitted to empirical test because they *cannot* be. Who decided that intelligence was the right variable to isolate? Who decided it was primarily heritable? Who decided that the causal arrow ran from genetics to outcome, rather than asking whether the very *concept* of intelligence as a singular, measurable quantity was itself a social construction, a tool of sorting and hierarchy dressed up as natural fact? The diagram conceals this history of decision-making. It makes contingent choices appear necessary. It says: *Here is how causation works,* when what it actually says is: *Here is one way of organizing our ignorance into the appearance of knowledge.* --- ## The Social Dimension — Where Intelligence Becomes a Weapon This is where the matter becomes urgent and where most contemporary discussions of causation, rigor, and transparency reveal their bankruptcy. In social systems — in education, employment, criminal justice, health care — causation is *never* a merely technical problem. It is fundamentally a question of *power*. Who benefits from a particular diagram? Whose interests are served by drawing the causal arrows in precisely this configuration rather than another? Consider: A researcher draws a diagram of recidivism. Prior arrests are here. Neighborhood poverty is here. Family instability is here. And from this diagram, we conclude that certain populations are *inherently* more criminal. We build prisons. We intensify policing. We tell ourselves we are following the causal logic. But we might have drawn the diagram differently. We might have placed at the center the deliberate choice to police certain neighborhoods with overwhelming force. We might have traced the arrow from that choice to the statistical probability of arrest. We might have shown how that arrest itself — the causal act — becomes the predictor of future arrest. We might have asked whether the "variable" we are measuring is *crime* or merely *police activity in certain zip codes*. All of these diagrams are transparent. All are mathematically rigorous. All are, in principle, testable. And yet they tell radically different stories about causation because they embed different assumptions about agency, about who acts and who is acted upon, about which facts are causes and which are mere symptoms. **This is not a problem that transparency solves. It is a problem that transparency *obscures*.** The moment you draw a diagram, you have already decided who the agents are. You have already located the causal power. You have already, in short, made a political choice about how the social world ought to be *understood* — and this understanding will shape policy, allocation of resources, and the distribution of blame and credit among human beings. --- ## What It Means to Know a Cause So I return to the original question: What does it mean to *know* the cause of something? It does not mean to possess a transparent diagram. It does not mean to have made your assumptions explicit. These are necessary — but they are not sufficient. They are the bare minimum of intellectual honesty, and they are frequently absent even in that minimal form. To know a cause is to understand not merely the mechanism but the *system of interests and alternatives* within which that mechanism operates. It is to ask not only "how does this work?" but "who benefits from my understanding it this way? What becomes *invisible* when I draw the diagram this way rather than another?" True intelligence — genuine intellectual power — consists in the ability to hold multiple causal models in mind simultaneously, to see how the choice between them is not determined by logic but by *values*, and then to choose deliberately, with full awareness of what one is choosing and why. This is harder than rigor. It is messier than transparency. It cannot be automated or reduced to a calculus. It requires judgment, which is to say it requires a human being willing to *take responsibility* for the diagram they draw and to defend it not merely on technical grounds but on grounds of *justice*. The social dimension is not a problem to be solved by better methods. It is the permanent condition of human knowledge. We are always, when we speak of causes in human affairs, making claims about how the world should be structured. The question is whether we do so honestly, with eyes open, or whether we hide behind the pretense that our diagrams are merely technical and therefore beyond moral scrutiny. Pearl gave us powerful tools. We have used them to make our assumptions visible. This is good. But we have not yet learned the harder lesson: that visibility is not innocence, and that a wrong diagram, clearly drawn, is still wrong.