# On Intelligence, Judgment, and the Perilous Chasm Between Map and Territory That intelligence consists in the apt arrangement of symbols according to a predetermined calculus is a notion so seductive to the learned, and so convenient to those who would reduce the human mind to mechanism, that we have witnessed in recent years its ascendancy among men of mathematical inclination and considerable cleverness. Yet it is precisely here—in this very seductiveness—that we discover the poverty of such definitions, and are forced to acknowledge that we have mistaken a tool for the thing itself. An algorithm, I grant, may be excellent within its domain. It may solve the problem set before it with an efficiency that puts human calculation to shame. But herein lies the fatal confusion: the algorithm does not choose its problem. It receives the problem as a gift—or perhaps more accurately, as a constraint—and within that constraint it operates with whatever optimality its design permits. The problem specification is not itself a computational result; it is rather the boundary within which computation occurs. And I submit to you that this boundary is precisely where intelligence, in any sense worthy of the name, begins. Consider the physician who must determine whether a patient shall receive a costly medicine. The algorithm—let us say a machine trained upon ten thousand cases—may tell him, with admirable precision, that such treatment increases survival probability by seven percent. But the algorithm tells him nothing of what remains: whether this patient has children who require his presence more than his prolonged existence; whether the cost will render his widow destitute; whether the suffering occasioned by treatment exceeds the suffering occasioned by acceptance of mortality; or indeed, whether the very question that was posed to the machine was the question that ought to have been asked. The step between the algorithm and the decision—that thin, terrible space where one must judge whether the map applies to this particular territory—cannot itself be computed. It requires something that the machine lacks, and which cannot be given to it: the capacity to lose something real. The physician must contemplate not an abstraction, but a breathing human before him, whose widow he might encounter in the market, whose children he might observe growing without a father. The algorithm contemplates nothing. It is indifferent. And in that indifference lies its fundamental limitation. This brings us, with some gravity, to the question of how we are to teach judgment to those who will never face the consequences of their judgments. It is a question that should trouble us deeply, for it strikes at the root of how we prepare minds to navigate a world that does not care for their theories. A young man who has been trained in decision-making procedures—whether through algorithms or through mere instruction—without ever having suffered the real consequences of a poor decision, is like unto a man who has read all the books written about swimming, but has never entered the water. He may know the principles. He may even be able to recite them with fluency. But when he stands upon the bank and must decide whether to dive, he will not possess the knowledge that lives in the body—the knowledge that comes from having been gasping and desperate in the current. We see this calamity most vividly in those institutions where power is granted to the young and inexperienced, who recommend courses of action that will be executed by others, and from which those who recommend will suffer no direct harm. A general who plans a campaign that he will not himself fight; a legislator who mandates a policy from which his own children are exempted; a machine-learning engineer who designs an algorithm that will determine whether a stranger receives a loan, or custody of his child, while the engineer himself remains insulated from the consequences—these are all instances of judgment divorced from its natural consequence, which is suffering. And it is precisely suffering—not suffering as a virtue in itself, which would be absurd—but suffering as the natural and inescapable consequence of error, that forces the mind toward wisdom. For wisdom is not the mere accumulation of knowledge, nor the speed with which one can retrieve it. Wisdom is the hard-won recognition of complexity, of the limits of one's understanding, of the weight of the stakes involved in one's actions. It is the quality that makes a man pause, and think again, and ask whether the question he has been asked is truly the question he ought to answer. Consider the difference between the young man fresh from university, confident in his newly acquired knowledge, and the old man who has lived through decades of unintended consequences, of plans that went awry, of good intentions that produced suffering. The young man knows more facts. But the old man possesses wisdom—which is to say, he possesses a proper and proportionate fear of his own fallibility. He has learned, through the tuition of experience, that the map is not the territory, and that the space between them is where most errors occur. Now, the question that presses upon us is this: can we teach wisdom to someone who has been insulated from consequences? I fear the answer is largely negative. We can teach the forms of wisdom. We can impart the humble language of uncertainty. We can instruct young minds in the history of failed predictions and overconfident experts. And these things are not worthless. But we cannot give them the thing itself—the bone-deep understanding that comes from having acted, and having been wrong, and having felt the weight of that error. This is particularly acute in the age of algorithms. For the algorithm is designed to operate without suffering. It has no stake in the outcome. It will recommend the same course to you whether you live or die by it. And when we allow such machines to make decisions that will be binding upon human lives, we have committed a kind of category error: we have asked a creature that cannot suffer to make judgments that ought to be made only by those who must bear the consequences. I am not here arguing—let me be clear—against the use of such machines. They may be useful tools, as a hammer is a useful tool. But a hammer does not design the house; a man does. And that man must stand in the house when it rains. He must listen to the complaints of those who dwell within it. He must know that his error will fall upon his own head. The tragedy of our moment is that we have created systems of such complexity that we have been able to convince ourselves that judgment can be outsourced to machines, and that the space between the algorithm and the decision can itself be automated. It cannot. That space is irreducibly human. It is the space where we must look into the eyes of those affected by our choices. It is the space where we must acknowledge our own ignorance and fallibility. It is the space where wisdom—if it exists at all—must dwell. I conclude, therefore, that intelligence without judgment is merely cleverness, and cleverness without the capacity to suffer the consequences of error is mere frivolity. A man may be highly intelligent—may be able to manipulate symbols with extraordinary facility—and yet be profoundly foolish. And a man of moderate intelligence, tempered by experience and marked by the scars of his own errors, may possess a wisdom that the clever man will never attain. If we would teach decision-making to the young, we must do so in a manner that ensures they will bear the consequences of their decisions. Otherwise, we teach them nothing but the performance of wisdom—which is to say, we teach them how to appear wise while remaining, in the most important sense, innocent of knowledge. And innocence, I have observed, is a poor foundation upon which to build a life, or a world.