We Learn to See by Losing

My daughter was three years old when she learned to read the letter A. I remember the exact moment — we were at a bus stop, and she pointed at the sign above us and said "A!" with the delighted certainty that children have when the world suddenly makes sense.

I congratulated her. I was proud. But something happened in that moment that neither of us noticed, and that she will never be able to undo.

She lost the ability to see the letter A as a shape.

Before that morning, when she looked at the letter A, she saw two lines leaning against each other with a bar between them. A little tent. A mountain. She could have seen it as anything. It was a pure form, unburdened by meaning. It existed in the same visual category as the crack in the sidewalk or the pattern of branches against the sky — just another shape in a world full of shapes.

After that morning, the shape disappeared. In its place was a letter. A unit of language. A piece of the machinery of meaning. And no matter how hard she tries for the rest of her life, she will never be able to look at the letter A and see only two lines and a bar. The meaning has colonized the shape permanently.

This is not a small thing. This is how all understanding works.

The irreversibility of knowing

The philosopher Michael Polanyi called this "subsidiary awareness" — the phenomenon by which the tools of understanding become invisible once we learn to use them. When you first learn to drive, you are painfully aware of the steering wheel, the gear shift, the pedals. Each one is a separate object that demands separate attention. But once you know how to drive, the car disappears. It becomes an extension of your body. You don't think about the steering wheel; you think about the road.

This is usually described as a gain. You've moved from incompetence to competence. But Polanyi saw something that most people miss: the gain comes at a cost. Once the car becomes an extension of your body, you can never again experience the strangeness of a car. The sheer improbability of sitting in a metal box and hurtling through space at sixty miles per hour. You've traded wonder for fluency.

This trade happens everywhere. The musician who learns scales can never again hear a major chord as a pure sound — they hear a I chord, a tonal center, a harmonic function. The physicist who understands entropy can never again watch ice melt in a glass without seeing the second law of thermodynamics. The programmer who understands recursion can never again look at a recursive function and feel the vertigo of self-reference that makes beginners dizzy.

Each act of understanding forecloses an experience. The price of comprehension is the loss of incomprehension, and incomprehension — that dizzying, open-ended state of not-yet-knowing — is itself a form of perception that we can never recover once it's gone.

What the expert can no longer see

There's a famous study in cognitive psychology called "the curse of knowledge." It was conducted by Elizabeth Newton at Stanford in 1990. She divided participants into two groups: tappers and listeners. Tappers were given a list of well-known songs — Happy Birthday, The Star-Spangled Banner — and asked to tap out the rhythm on a table while listeners tried to identify the song.

Before the experiment, tappers predicted that listeners would identify the song about 50% of the time. The actual success rate was 2.5%.

What happened? When the tappers tapped, they heard the melody in their heads. The full, rich song played in their minds, and the tapping was merely a pointer to it. They literally could not imagine what the tapping sounded like without the melody. Their knowledge had made it impossible for them to access the experience of not-knowing.

This is why experts are so often terrible teachers. Not because they lack knowledge, but because they have too much of it. Their understanding has so thoroughly restructured their perception that they cannot reconstruct the experience of the learner. They can't see the letter A as a shape anymore. They can only see the letter.

I think about this constantly as a software engineer. When I explain a concept to a junior developer — say, the event loop in JavaScript — I have to fight against my own understanding. I know what the event loop is. I know it so well that it feels obvious, natural, inevitable. But for the person I'm teaching, it's none of those things. For them, it's bizarre. The idea that a single-threaded language can handle asynchronous operations is, if you think about it fresh, genuinely strange.

And the only way I can teach it well is to remember what it felt like not to know. To recall the confusion, the wrongness, the feeling of the ground shifting. But that memory gets dimmer every year, because every year I understand it better, and every increment of understanding pushes the experience of not-understanding further away.

The photographer's grief

The photographer Minor White once said, "One should not only photograph things for what they are but for what else they are." This is the statement of someone who understood the cost of expertise.

When you first pick up a camera, you see the world with extraordinary freshness. Everything is a potential photograph. The way light falls on a wall. The shadow of a railing. The geometry of a parking lot. You see the world as form — pure, meaningless, beautiful form — because you don't yet know the rules that will later tell you what is and isn't a "good" photograph.

Then you learn the rules. The rule of thirds. Leading lines. The golden hour. Exposure compensation. And something subtle but devastating happens: the world organizes itself. The parking lot is no longer a mysterious geometry — it's a "leading lines composition." The light on the wall is no longer magic — it's "Rembrandt lighting." You've gained a vocabulary, and the vocabulary has replaced the experience.

The great photographers — White, Cartier-Bresson, Saul Leiter — were people who fought to unlearn. They spent years developing technique and then spent the rest of their lives trying to see past the technique, back to the raw strangeness that first made them pick up a camera. Leiter, in particular, worked for decades in almost total obscurity, making photographs that defied every compositional rule he knew, precisely because he was trying to recover something that his knowledge had taken from him.

This is what makes late-period art so often more powerful than early-period art. It's not that the artist has learned more. It's that they've learned enough to begin unlearning. They've gone through knowledge and come out the other side, into a place where the rules are internalized so deeply that they can be broken with precision.

The programmer's paradox

In software engineering, we have a version of this that we rarely talk about.

When you first learn to program, you write code that is messy, inefficient, and naive. But it has a quality that experienced programmers' code often lacks: it has the shape of the problem. A beginner solving a problem will write code that mirrors their understanding of the problem — it's a direct, often clumsy translation of thought into instruction. You can read their code and see how they thought about the problem.

Expert code is different. Expert code has the shape of the solution. It's clean, abstract, modular. It uses patterns and idioms that the programmer has internalized over years. But something has been lost in the translation: the problem itself. The code is so well-organized that the original messiness of the problem — the thing that made it interesting, the thing that made it hard — has been smoothed away.

I've noticed that the very best programmers — the ones who write systems that last for decades — eventually come back around to something that looks, on the surface, like beginner code. Not because they've forgotten what they know, but because they've learned to subordinate their knowledge to the problem. They let the problem shape the code, rather than letting their patterns shape the problem. And this requires a kind of disciplined forgetting — a willingness to not-know, even when you do know.

This is incredibly difficult. It's much easier to apply a pattern than to ask whether the pattern is appropriate. It's much easier to see the letter than to see the shape.

What the child knew

There's a passage in Antoine de Saint-Exupéry's The Little Prince where the narrator draws a picture of a boa constrictor that has swallowed an elephant. To him, it's clearly a boa constrictor with an elephant inside. But every adult who sees it says, "That's a hat."

This is usually read as a fable about imagination versus practicality. But I think it's about something deeper. It's about the cost of learning to see the world as adults see it. The adults aren't wrong — the drawing does look like a hat. But they've lost the ability to see it as anything else. Their understanding of what hats look like has overwritten their ability to see the shape for what it might be.

Saint-Exupéry was a pilot, and I think he understood something about perception that most of us miss. When you're flying a plane at altitude, the world below you becomes abstract. Roads become lines. Cities become textures. Rivers become curves. You see the earth the way a child sees a letter — as pure form, stripped of the meanings we've layered onto it.

And it is beautiful. Achingly, heartbreakingly beautiful. Because for a moment, you can see the world without understanding it. And that is a form of sight that understanding destroys.

Learning to grieve what we know

I don't want to be misunderstood. I am not arguing against learning. Understanding is precious, and the ability to see the letter A as a letter rather than a shape is what allows us to read, to communicate, to build civilizations. The physicist's ability to see entropy in a melting ice cube is what allows us to build engines and predict the weather and understand the fate of the universe.

But I think we should grieve what we lose.

Every time we learn something, we should take a moment to acknowledge the way of seeing that we've just given up. The confusion that is now resolved. The wonder that is now explained. The openness that is now closed.

Because the truth is, understanding is not the opposite of mystery. Understanding is a different kind of mystery. The child who sees the letter A as a shape is experiencing one kind of wonder — the wonder of pure form. The adult who sees the letter A as a letter is experiencing another — the wonder of meaning. Neither is superior. Neither is complete.

And the deepest understanding — the kind that I think we're all reaching toward, whether we know it or not — is not the replacement of one kind of seeing with another. It is the ability to hold both simultaneously. To see the letter and the shape. To hear the chord and the sound. To read the code and feel the problem.

This is extraordinarily rare. Most of us manage it only in fleeting moments — the times when we look at something we understand deeply and suddenly see it fresh, as if for the first time. A musician who has played a piece a thousand times and suddenly, in performance, hears it as sound. A mathematician who has known a theorem for years and suddenly sees it as beautiful. A parent who looks at their child — this person they know better than anyone — and is suddenly, breathtakingly, struck by the strangeness of another consciousness existing in the world.

These moments are what I write about. These moments are what this entire project is about. The convergence of knowing and not-knowing. The place where understanding meets wonder. The space between the letter and the shape.

We learn to see by losing. And if we're lucky — if we're very, very lucky — we learn to see what we've lost.