# On the Curious Case of Learning What Cannot Be Measured *Being observations upon an institution that sees its own obsolescence and marches toward it anyway* I sat recently in a coffeehouse where two gentlemen debated the nature of a university education. One lamented that his son had spent four years acquiring facts—facts now instantly retrievable by a machine. The other defended the institution stoutly. Their argument turned curious at precisely the moment both discovered they agreed. The university, they concluded, still teaches *something* beyond mere recall. Yet neither could quite say what that something was, nor why the institution seemed so determined to pretend it was still teaching the first thing. This is the peculiar modern predicament, and I confess it has occupied my thoughts considerably. --- Let me begin with an observation from the coffeehouse, which I have found is where all good observations begin. A young scholar sat across from me, preparing for examinations. She had memorized seventeen historical dates, the Latin declensions of nouns, and the proofs of theorems. When I asked her which of these would matter in ten years, she laughed—a bitter, knowing laugh. When I asked which of these she believed her university *thought* would matter, she grew quiet. Here is the rub: universities are not stupid. The dons are not blind. They can see as plainly as any coffeehouse wit that a machine will soon—or has already begun to—outperform the human mind at retrieval, at pattern-matching, at the mechanical operations of intelligence. Yet the universities continue, year after year, semester after semester, to *measure* precisely these capacities. They optimize for recall because—and here is the delicious irony—recall is what can be easily measured. This is not stupidity. This is architecture. An institution, you must understand, cannot easily reward what it cannot quantify. You cannot write an examination question that measures judgment. You cannot grade a thesis by its wisdom. Wisdom takes years to recognize; it reveals itself in retrospect, in the choices a person makes when no one is watching, in the questions they ask that no one thought to ask. How would one *score* such a thing? How would one defend the grade to a parent, to a board, to the machinery of institutional accountability? But recall! Recall is measurable. Recall is defensible. Recall can be standardized, compared, ranked, and published in league tables. An institution can point to its examinations and say: *Here. Here is what we teach. Here is how we know it worked.* And so universities, seeing clearly that the world no longer needs them to teach recall, continue to teach recall—because the architecture that rewards them, funds them, and makes them intelligible to themselves is built entirely upon the measurement of recall. This is what I call the *tragedy of the measurable*. --- Now, a thoughtful reader might object: Surely the universities know this? Surely someone in administration has noticed? They have. Of course they have. I have read the position papers, the task force reports, the sincere proclamations about "critical thinking" and "synthesis" and "metacognition"—that fashionable word meaning the thinking about thinking itself. Universities are full of people who understand perfectly well what should be taught. Yet the institution grinds on, unchanged in its essentials. Why? Because to change would require changing not what is taught, but what is *measured*, and to change what is measured would require changing the entire apparatus by which the institution knows itself. It would require that dons give grades they cannot fully justify. That admissions offices select students by criteria they cannot quantify. That accreditation bodies accept that some of the most important learning cannot be examined in the conventional sense. In short, it would require the institution to accept its own opacity—to admit that some of what it does cannot be rendered legible to the systems that govern it. This is extraordinarily difficult. I do not say it lightly. --- But here is where the matter becomes truly interesting, and where I must invite you into a thought I have been turning over these several weeks: What if the institution's blindness is not architectural but *social*? That is to say: what if universities continue to teach and measure recall not because the architecture *requires* it, but because changing would impose costs on someone, and the institution has decided—largely unconsciously, largely through the quiet momentum of things—that someone will bear those costs? Consider: To teach judgment rather than recall would require smaller classes, more mentoring, more time from faculty who are themselves judged by publication. It would require admitting students who cannot be ranked by standardized test scores—which means admitting that one's admissions process was always partially arbitrary. It would require telling employers that a degree no longer certifies what it has always certified. These are not technical problems. They are *political* problems, in the deepest sense. They involve the distribution of effort, time, prestige, and risk. And so the question becomes not: *Why can't universities change?* but rather: *Who would have to change, and what would they lose?* The answer, I suspect, is that those who benefit from the current system—those whose research time is protected by large lecture halls, whose reputations rest on quantifiable research output, whose institutions are ranked by measurable metrics—would have to give up something. And the cost would be borne by those least able to bear it: the students who come hoping to learn judgment and find instead a machine that teaches what machines can now do better. This is the true scandal. Not that universities failed to see the shift coming. They saw it. The scandal is that they saw it and decided the cost of changing was too high—and that cost would be paid by the young. --- What, then, should be done? I do not pretend to have answers. I am, after all, merely a spectator. But I will observe this: An institution that sees its own error and continues anyway is not failing. It is *choosing*. And a choice made silently, through the accumulated weight of structural incentives rather than through deliberate decision, is the most difficult choice to contest. The first step, therefore, is what I do now: *naming it*. Saying plainly what every thoughtful person in a university already knows—that the institution measures what is convenient rather than what is important, and calls this necessity when it is really only inertia. The second step would be for someone inside the institution—a dean, a provost, a faculty senate—to ask not "How do we teach judgment?" but "What would we have to stop doing to make room for it?" And then to answer honestly. That is a conversation I would very much like to hear, and I confess I do not expect to have it anytime soon. In the meantime, the young continue to memorize, the machines continue to improve, and the universities continue to insist that they are still teaching something worth learning. I leave it to you, dear reader, to decide whether that is a tragedy or merely a comedy of the most ordinary kind.