I was sitting yesterday in my usual corner at the Coffeehouse, observing the learned and the foolish in their customary proportions, when a gentleman — one of those modern fellows much occupied with machines and their capacities — fell into discourse with a younger man, lately returned from university. The contrast between them struck me forcibly. The elder spoke of what his mechanical apparatus could accomplish: it could retrieve information with astonishing speed, arrange words in pleasing patterns, and even mimic the reasoning of better minds than its own. "Behold," said he, "how it surpasses the labours of men!" The younger listened, then attempted to apply some learning to a question of natural philosophy. He could recite what he had been taught, but when pressed to defend it, or to ask why it mattered, or to consider what assumptions lay beneath it — he was quite at sea. He had, I perceived, been trained much as one trains a parrot: to repeat, not to think.

"You are both remarkable," I ventured, "for possessing the same deficiency."

This produced an uncomfortable silence.

Let me speak plainly, as is my custom in these papers.

Intelligence is not the mere possession of answers. A merchant's ledger possesses quantities of fact. A library possesses vastly more. Neither thinks. Intelligence, I have come to believe through observation, lives in the space between what is known and what is questioned. It is the capacity to stand before a problem and ask not merely "what is the answer?" but "is this the right question?" and more troublingly still, "why do I suppose this question matters?"

This is where both our silicon oracle and our academically-trained youth fail most conspicuously. The machine cannot audit plausibility — that is, it cannot step back from its own utterances and ask whether they cohere with the world as it actually is. It produces what is statistically probable, not what is true. A most dangerous distinction. The student, meanwhile, has been schooled in the reception of questions, rarely in their origination. He knows what he has been told to know. He does not know what he does not know he should ask.

But I wish particularly to address a dimension that these learned discussions persistently neglect, and which I have observed to be the very seat of intelligence itself: the social.

I have noticed something curious about intelligence in the wild. When two minds meet — truly meet, without pretense — something occurs that neither achieves alone. A merchant conversing with a physician will ask questions the physician has forgotten to ask. A widow of good sense will perceive a logical flaw in a gentleman's reasoning that his own education has rendered invisible to him. The very fact of another mind present, with different stakes and different blindnesses, seems to sharpen the faculties.

Intelligence is not solitary. This is what we have got wrong.

We have built our machine in isolation, feeding it the written word without the living presence of another consciousness to contradict it, to laugh at it, to ask "but what of my experience, which differs from your pattern?" We have trained our students in lecture halls where the appropriate response is passivity, where to ask an impertinent question is to disrupt, where the teacher is presumed to know what questions are worth asking. And now we find ourselves with two species of inadequacy — the machine and the graduate — both lacking what cannot be programmed or lectured: the social intelligence that arises only when one mind must reckon with another.

What is this social intelligence? It is the capacity to be uncertain in the presence of another person and to hold that uncertainty without collapsing it into false certainty. It is the ability to say "I do not know" not as a confession of failure but as an opening. It is the willingness to let another person's different way of seeing the world actually alter what you see. It is, in short, the opposite of both the machine's confident hallucination and the student's confident repetition.

I watched two merchants argue yesterday about the fair price of silk. Neither had read treatises on political economy. But in their disagreement — in the fact that each had to account for the other's perspective, had to explain why his reasoning differed — something emerged that resembled intelligence. Not the intelligence of books, but of judgment. The ability to navigate between competing plausibilities and choose wisely.

Now we are told that we must decide what to teach in the ruins of that coincidence — the ruin being that we created a machine incapable of causal reasoning and a student incapable of genuine questioning, and called both progress.

Here is what I would propose.

We must teach the young to think socially. Not in groups, though that is part of it. Rather, to internalize the presence of another mind — a mind that disagrees, that knows different things, that will not accept cant. Teach them to read not for the answers an author provides, but for the questions implicit in the author's blindnesses. Teach them to write as though they must defend their reasoning to a skeptic. Teach them to sit with another person and genuinely not know what will be said next, and to find this exhilarating rather than threatening.

Teach them, in other words, to do what the machine cannot do and what the old curriculum neglected: to think with others, not merely to absorb information or to process it.

As for the machine — let it do what it does. But do not mistake its fluency for intelligence. Intelligence arrives when a human being stands before it and asks: but is this true? And does it matter? And what are you not telling me?

Intelligence is the question, not the answer. This is best learned, I have observed, in conversation. Not in solitude before a screen, nor in silence before a lecturer, but in that ancient and irreplaceable forum: the meeting of two minds that must somehow make sense together.

Until we teach this — not as an add-on, but as the substance of education — we shall produce only sophisticated versions of what we have already made: confident emptiness, whether in silicon or in flesh.

Your humble Spectator