# On Causation, Knowledge, and the Proper Bounds of Inquiry To inquire what intelligence truly is—and more perilously, what causes it—is to stand at that precipice where the natural philosopher must acknowledge the difference between the precision of his instruments and the rectitude of his conclusions. I confess myself implicated in this error, for I have long been the sort of man who believes that clear language and rigorous method must together produce truth. Experience, that stern schoolmaster, has taught me otherwise. The banishment of causal language from the sciences was, we are told, an act of methodological virtue. For nearly an age, the learned men of natural philosophy declared that to speak of *why* things occur—to assert that X causes Y—was to indulge in unscientific fancy, that such language belonged to the metaphysical nursery from which rigorous minds must graduate. They would content themselves with correlation, with mathematical description, with the humble catalogue of what occurs when. This was called *rigor*. But rigor, I have come to believe, is a word much abused by those who mistake the narrowing of vision for the clarification of it. Enter Judea Pearl, a man whose contribution to human knowledge deserves the serious reckoning it has received. He has performed an act of restoration: he has shown us that causal reasoning need not be the enemy of mathematical exactitude, that one may speak of causes with the same precision one brings to equations. He has given us *tools*—diagrams, algorithms, formal procedures—by which we might navigate from observation toward understanding. For this, he merits our gratitude. Yet here lies the rub, and it is a rub that no amount of methodological sophistication can entirely smooth away: **Pearl's tools are provably correct only given assumptions that precede the data, not those that emerge from it.** The diagram—that crucial architecture upon which everything depends—must be *assumed*. It is not discovered in nature like a fossil or deduced from first principles like a theorem. It is *chosen*. Let me be explicit about what this means, for clarity here is not mere academic nicety but a matter touching upon the very possibility of human knowledge. When a researcher proposes a causal diagram—asserting that intelligence depends upon genetics, environment, and education in such-and-such relationship—he has not thereby made his model true by making it transparent. Transparency is indeed not validity. A man may draw his lines with perfect clarity, label his nodes with admirable precision, explain his assumptions with exemplary candor, and yet be *completely wrong*. The diagram may be a beautiful lie, a perfectly articulated fantasy. The rigor lies in the logical consequences that follow *from* the diagram, not in the diagram itself. This is not a defect in Pearl's framework—it is rather the revelation of what was always the case, now made visible. The scientist has always made assumptions; now, at least, he is compelled to make them *explicit*. This is progress, though a progress that brings with it a kind of melancholy knowledge: that the foundation of inquiry rests not upon bedrock, but upon the ground of human judgment. --- ## On Causation and Its Proper Understanding To know the cause of something is not what the naive mind supposes it to be. We imagine that causation is a kind of mechanical pushing, a billiard ball striking another, and that to know a cause is to observe this pushing directly. But even in the physical world, this is an illusion. We never observe causation itself; we observe only constant conjunction, regularity, the succession of events. What we call *cause* is our mind's interpretation of this pattern—an interpretation that is useful, that allows us to predict and to act, but which is not simply *read off* from nature. When we speak of the causes of intelligence, we speak of something still more obscure. Intelligence is not a billiard ball. It is a capacity, a tendency, a way of engaging with the world that manifests itself in a thousand particulars. Does genetics cause intelligence? In what sense? The genes do not contain intelligence as a chest contains coins. Rather, they establish conditions within which intelligence may develop. Does environment cause intelligence? Again, we must ask what we mean. The poverty-stricken child deprived of books and instruction will not develop the same intelligence as the child in a library—but is this because the environment *causes* intelligence, or because it *permits* or *prevents* its development? The grammar of causation grows uncertain. Here is where the causal diagram reveals both its utility and its limitation. By forcing us to specify *how* we believe the world works—which variables influence which, in what sequence—it compels us to make our tacit assumptions explicit. This is genuinely valuable. But the diagram cannot escape the fundamental problem: **the person who draws it must decide what variables to include and how to connect them.** And this decision is not made by the data; it is made by the researcher, drawing upon his prior beliefs, his theoretical commitments, his sense of what matters. --- ## On the Question of Authority: Who Decides What the Diagram Is? This brings us to a question that is not merely technical but profoundly political and moral: **Who decides what the diagram is?** In the sciences as they are currently practiced, this decision falls to those with credentials, with institutional authority, with access to publication and funding. A researcher at a prestigious university proposes a causal model of intelligence; it is published in a respected journal; it is cited by others; it becomes, by a kind of social gravity, accepted wisdom. But the authority that validates the diagram is not the authority of nature itself—it is the authority of the scientific community, which is to say, the authority of human beings with their prejudices, their fashions, their institutional interests. I do not say this in a spirit of cynicism. The peer review process, imperfect though it is, represents one of humanity's better attempts to regulate inquiry. But we must not confess the regulation and then forget that it is regulation—that it is the exercise of human judgment, not the discovery of natural law. Consider: in an earlier age, causal diagrams of intelligence would have included "racial essence" as a fundamental variable. This was not because the data demanded it, but because the researchers—respectable men, learned men—believed it to be true. Their diagrams were transparent; their methods were rigorous by the standards of their time; their conclusions were *completely wrong*. The diagram had been drawn by those with power, and it reflected their prejudices. Today, we like to believe we are more enlightened. And in some respects, we are. But the fundamental problem remains: the diagram is still drawn by human beings, still reflects their prior commitments, still embeds assumptions that are not themselves validated by data. The only difference is that we are now more aware of this fact—or at least, we ought to be. --- ## On Wisdom, and the Proper Bounds of Inquiry Here we arrive at what may be the most important question of all: What is the relationship between intelligence and wisdom? The sciences, as they have developed, have become increasingly expert at answering questions of a particular kind: *Given these assumptions, what follows?* This is a form of intelligence—a rigorous, systematic, powerful form. But it is not wisdom. Wisdom is something different. It is the capacity to know not merely what follows from one's assumptions, but whether those assumptions are worth making. It is the ability to perceive the limits of one's knowledge, to recognize the difference between precision and truth, to act rightly in the face of uncertainty. Wisdom asks: *Is this question one I ought to be asking? Is this the right way to frame the problem? What am I leaving out?* A man may be highly intelligent—may be skilled in mathematics, in logic, in the manipulation of data—and yet be entirely lacking in wisdom. Such a man may construct a causal diagram of extraordinary complexity and mathematical sophistication, may derive conclusions with perfect rigor, and may be profoundly mistaken about something essential. He may optimize for the wrong thing. He may solve a problem that did not need solving while ignoring a problem that desperately does. The study of intelligence—and this includes the causal study of what produces intelligence—has become increasingly a matter of technical expertise. We have graphs and algorithms, databases and statistical procedures. We have, in Pearl's framework, a way of formalizing causal reasoning that is mathematically elegant. All of this is to the good. But it is not enough. What is required, in addition to intelligence, is wisdom. And wisdom cannot be formalized, cannot be reduced to a diagram, cannot be derived from data alone. Wisdom is the fruit of experience, of moral seriousness, of the willingness to acknowledge one's own fallibility and the fallibility of one's methods. --- ## The Proper Understanding of Knowledge To return to the original question: **What does it mean to know the cause of something?** It means, I submit, something more modest and more difficult than we often suppose. It means to understand a relationship between things that is stable enough to rely upon, clear enough to articulate, and honest enough to acknowledge its own limits. It does not mean to possess certainty. It does not mean to have escaped the need for judgment. It does not mean to have transcended the human condition and achieved a view from nowhere. When a researcher constructs a causal diagram and derives conclusions from it, he has done something valuable—but what he has done is not to discover the truth about the world. He has rather constructed a *model*, a representation that captures certain aspects of reality and necessarily omits others. The model may be useful. It may lead to predictions that prove accurate. It may guide action in beneficial directions. All of this is to the good. But the researcher should know—and should tell others—that this is what he has done. He has not read the mind of God. He has not escaped the human condition. He has made assumptions, drawn conclusions, and must stand ready to defend both. The banishment of causal language was, in its way, an honest mistake. The natural philosophers of the last century believed they were being rigorous by excluding causal talk; they were actually impoverishing their language and obscuring their assumptions. Pearl has done us a service by restoring causal reasoning to respectability—by showing that one can speak of causes with mathematical precision. But this restoration brings with it a responsibility. We must now be more transparent about the fact that causal reasoning, for all its rigor, rests upon foundations that cannot themselves be validated by the methods of natural science. The diagram must be chosen. The variables must be selected. The relationships must be specified. And these acts of choice and specification are human acts, fallible acts, acts that require wisdom. --- ## Conclusion: On the Limits of Method In the end, to ask what intelligence is, and what causes it, is to ask a question that no amount of methodological sophistication can entirely settle. Intelligence itself is implicated in the inquiry—for it is intelligent beings who are asking the question, who are drawing the diagrams, who are deciding what matters and what can be left aside. We are, in a sense, trying to understand ourselves by means of the very capacities we are trying to understand. This is not a defect in our method; it is the human condition. We cannot step outside ourselves to achieve a perfect view. We can only proceed with as much clarity, honesty, and rigor as we can muster, while remaining always aware that we are the ones doing the proceeding. The tools that Pearl has given us are excellent tools. Let us use them. But let us use them with wisdom—with an awareness of their power and their limits, with honesty about our assumptions, with humility about what we claim to know, and with genuine concern for how our knowledge affects the lives of those about whom we presume to theorize. For in the end, intelligence divorced from wisdom is a dangerous thing. And wisdom requires that we never forget: we are all of us implicated in the inquiry. We cannot escape it. We can only hope to conduct it well.