# On the Perilous Wisdom of Those Who Calculate Without Consequence *By Mr. Spectator* I took my morning chocolate yesterday at Will's, where I fell into discourse with a gentleman of considerable learning—one of those modern philosophers who concerns himself with the engines of calculation that our age has produced. He spoke with great enthusiasm of algorithms, those systematic procedures by which machines arrive at optimal solutions. "They are perfect," said he, "within their domains." "Perfect?" I inquired. "Then why do we not all employ them for the conduct of our lives?" He smiled at this, perceiving my ignorance was not mere affectation. "Because," said he, "the algorithm solves only the problem it has been set to solve. It optimizes brilliantly for *its* specifications. But life does not present itself in specifications." This observation struck me with some force, and I have since meditated upon it considerably. --- Consider a merchant of my acquaintance. He employs a clerk of admirable facility with numbers, who calculates the profit of each transaction with exactitude. Yet this clerk, I have noticed, recommends courses of action that maximize the ledger without regard to the merchant's reputation, his word, or the repeated custom of honest men. The calculation is flawless. The wisdom is absent. Why? Because the clerk has never lost a customer to his own poor judgment. He has never felt the slow erosion of trust, never watched as men of worth ceased to conduct business with his master. The *consequence*—that most rigorous of teachers—has never touched him. He solves the problem he was given. He does not solve the problem his master *actually faces*. This is a problem of what we might call *specification discord*. The algorithm optimizes for Profit. Life requires the optimization of Profit-while-maintaining-Honor, Profit-while-preserving-Reputation, Profit-within-the-bounds-of-Justice. These are not the same problem. The addition seems small, a mere clause. Yet it is everything. --- But I wish to press this matter further, into more troubling territory. What becomes of us when we attempt to teach decision-making to those who shall never face its consequences? I have observed this lately in the matter of education. We place young persons—clever ones particularly—into systems of instruction where they are asked to solve problems without ever bearing the weight of their solutions. A youth is asked to optimize a budget. Excellent. He performs admirably. But he need not eat the bread that budget provides. He need not face the widow whose pension he calculated away in his pursuit of systemic elegance. "This is only practice," we tell ourselves. "They will learn responsibility when it matters." But I wonder if this is true. For there is a moral muscle that atrophies without use. The capacity to *feel the gravity* of a decision—to know that your calculation will be borne by actual flesh, actual hunger, actual sorrow—this cannot be learned from description. It must be learned from consequence. And here is the difficulty: we cannot ethically grant young persons the full consequences of their errors in judgment. We rightly protect them from ruin. Yet in protecting them, do we not also protect them from the very thing that transforms calculation into wisdom? --- I must now speak to a matter even more delicate: the social dimension. When we teach decision-making to those who will never face consequences, we often teach them in *groups*, within *institutions*, where responsibility becomes diffuse. A committee votes to optimize a city's budget. The algorithm recommends closing the clinic in the poor quarter—the calculation is sound, the efficiency gains are real. But when the clinic closes, *who* has decided this? The algorithm has no face. The committee members, individually, bear only a fraction of the blame. The person who wrote the specifications for the algorithm was solving a technical problem, not a moral one. And somewhere, a mother cannot get her child treated. This dispersal of consequence is the great danger of our age. We have created systems of such complexity that no individual within them bears the full weight of their decisions. The algorithm is optimal for its specification. The specification was set by someone who did not consult with those affected. Those affected have no voice in what the specification should be. And we wonder why such systems produce such injustice. --- What, then, is intelligence in this light? I have come to think that intelligence—*true* intelligence, not mere calculation—consists partly in this: the capacity to recognize when the problem you have been asked to solve is not the problem you should actually be solving. It is the judgment that says: *Yes, I can optimize for this specification. But should I? Does this map apply to this territory?* This judgment cannot itself be computed. It requires something the algorithm does not possess: a stake. A possibility of loss. A connection to actual consequence. This is why the wisest counselors are often not the cleverest calculators, but those who have suffered. They have learned, through the grammar of consequence, to ask the right questions before optimizing. They understand that the step from *this is optimal* to *this is right* is not a mathematical step. It is a human one. And it cannot be taught to those who will never face the consequences of their teaching. --- I commend this to the consideration of those who design our institutions of learning, particularly those who would train the young in matters of consequence. It is not enough to teach them to calculate. We must teach them to *feel the weight* of calculation. But how to do this without permitting them to destroy themselves or others through their inevitable errors? This is perhaps the central question of education. And it admits no algorithmic solution. For the answer lies precisely in that space between the specification and the territory—that human space where judgment lives, and where consequence is real.