# On the Causes of Things: A Discourse Upon Knowledge and Its Pretenses *Being observations made in the coffeehouse, with some remarks upon modern philosophy and the curious predicament of those who study the mind.* --- I have observed that mankind labours under a peculiar vanity in matters of understanding. We wish to know *why* things happen, yet when this wish grows inconvenient to our method, we simply declare the wish itself unscientific. Thus does rigour become not the pursuit of truth, but the avoidance of difficult questions dressed in the language of probity. A gentleman recently explained to me the work of one Pearl—a mathematician of considerable ingenuity—who has restored to Science a tool it had renounced. For near a century, we natural philosophers had agreed that to speak of *causes* was to descend into speculation. Better to observe correlations, to measure associations, to remain safely within what the data itself could utter. This was called rigour. It was, in fact, a kind of cowardice made respectable. But here is where the matter grows curious, and where I confess the coffeehouse fell silent. Pearl's contribution is this: *If* you draw a diagram—a map of which thing influences which—*then* you may derive what causes may be inferred from your observations. The tools are provably correct. The logic is airtight. And yet (here is the rub that troubles a Spectator of modest understanding) the entire edifice rests upon an assumption you must make *before* the data speaks at all. The diagram is not discovered. It is *assumed*. It is *believed*. It is, in the most literal sense, drawn from the air. This is rather like saying: "If you assume the universe is arranged thus-and-so, I can prove to you that the universe is arranged thus-and-so." The circularity is elegant, but it is circularity nonetheless. --- Now, a man might ask—and I have asked—what separates such a diagram from the old metaphysical speculation we were meant to have abandoned? The answer, I'm told, is *transparency*. The researcher announces his diagram openly. He does not hide it in the bowels of his method. He does not pretend it emerged from the data, pristine and innocent. He *declares* it. This strikes me as progress of a sort. It is easier to quarrel with what is said plainly than with what is smuggled in disguise. Yet I cannot help observing that a man may be perfectly explicit about a model that is perfectly wrong. Transparency is not truth. Honesty about one's assumptions does not make the assumptions correct. I knew a merchant once who kept his ledgers with great clarity—every false entry written fair and plain. His accounts were models of transparency. He was, nonetheless, bankrupt. --- Here is where the matter touches the Social dimension, and where I believe the true difficulty resides. Who, pray, decides what the diagram shall be? In the natural philosophy of falling bodies or planetary motion, the diagram is relatively safe. The forces involved do not care who observes them, and they do not alter their character based on who is watching. But when we turn our instruments toward human nature—toward intelligence, capacity, potential—the diagram becomes a *choice* laden with power. Consider: A researcher wishes to understand intelligence. Before the data can speak, she must draw a diagram. Does intelligence flow from genes, or from circumstance? From individual effort, or from social advantage? Does it flow *backward* from success to talent, or *forward* from talent to success? Each diagram will permit the data to tell a different story. Each is transparent. Each is wrong in different ways. But here is the particular sting: The diagram is not drawn in solitude. It is drawn within a society—one that already believes certain things about intelligence, about who possesses it, about what it is for. The researcher, however rigorous, does not stand outside these beliefs. She swims in them as a fish swims in water, often unaware of the medium itself. A diagram that seems obviously true to a man of privilege may seem obviously false to a man of none. The transparency of the method does not resolve this. It merely makes the disagreement explicit. --- I shall venture a thought that may displease the rigorous. Perhaps the question "What does it mean to know the cause of something?" cannot be answered by mathematics alone, because it is not fundamentally a mathematical question. It is a *social* question. It asks: In this community, with these purposes, given these values, what explanation shall we accept as sufficient? What diagram shall we agree to work within? This is not to say that anything goes—that all diagrams are equally valid. Some diagrams will prove fruitful; others will lead to ruin. Some will align with how the world actually works; others will not. But the test of a diagram is not purity of method. It is *consequence*. Does it lead us to act wisely? Does it help us flourish? Does it serve justice, or entrench injustice? When we study intelligence—that most laden of terms—we are not merely studying a natural phenomenon. We are, whether we acknowledge it or not, choosing what kind of society we wish to be. We are deciding what we shall count as a human excellence. We are drawing a diagram of human worth. The researcher who presents this diagram with perfect transparency, yet without acknowledging that she has drawn a diagram at all, has merely made her power invisible rather than eliminated it. She has dressed her choice in the garments of necessity. --- I confess that I do not know how to resolve this. Pearl's tools are genuine achievements. They restore to us a language we had foolishly abandoned. Yet they do not solve the anterior problem: the problem of the diagram itself. Perhaps the answer lies not in method, but in *plurality*. Let us have many diagrams, drawn by many hands, emerging from many traditions and communities. Let us see what each reveals, what each conceals. Let us quarrel over them openly in the coffeehouse, as gentlemen and ladies ought. Let us be rigorous about our methods, yes—but also humble about what our methods can accomplish. For in the end, to know a cause is not merely to possess a correct diagram. It is to possess a *shared* understanding—one that a community has deliberated upon, tested against experience, and chosen to live within. That understanding must be transparent, yes. But it must also be *political*, in the finest sense: it must be something we have decided together. Until we acknowledge that we are making such decisions, we shall continue to mistake our choices for discoveries, and call this progress. --- *Having set down these reflections, I remain, as ever, your humble Spectator—observing, questioning, and uncertain whether even this discourse has not itself been drawn from invisible assumptions. Perhaps that uncertainty is itself the beginning of wisdom.*