My son, eleven, said this in the car after a pickup game.
We had been driving in silence for a few minutes, the way you drive after a kid has played hard and is still cooling down. He had produced something during that game I had not seen him do before — a back-heel flick that lifted the ball over a defender’s leg and turned into a long pass, in stride, under pressure. The kind of action that makes the parents on the sideline turn and look at each other.
I asked him, eventually, what was going through his head when he did that.
He thought about it for a second. Then he said: I was in my Neymar.
Not I did a Neymar move. Not I tried something I saw Neymar do.
I was in my Neymar.
I drove the rest of the way home thinking about the preposition.
A kid who watches highlight clips and tries to copy what he saw says different things.
He says I tried the elastico. He says I was going for the rainbow. He says I almost did the one where Neymar puts it through the guy’s legs. The language is retrieval language. The kid is reaching into a catalog of named moves and pulling one out to attempt.
Most American youth training produces this kid. After enough years of it, his internal list is long — step-over, scissors, Cruyff turn, croqueta, elastico, Maradona spin, rainbow — and on Saturdays he tries to find moments to insert items from the list into the game. Sometimes it works. Often the move he attempts does not fit the situation. He pulls the elastico out of his catalog because he wants to do the elastico, not because the moment called for it. The defender takes the ball. The catalog gets longer. The kid does not get more intelligent.
This is what most of us think development looks like. A growing list of moves the kid can do. Watch the kid play. Count the moves. Mark the ones he executed cleanly. Project a player.
But there is a different kid you can sometimes see, and once you have seen him you cannot unsee him. He never seems to be reaching into a catalog. He never seems to be trying anything in particular. He receives the ball and something happens — a turn, a flick, a pause, a sudden change of pace — and you cannot quite name the move he just did, because the move he just did does not have a name. It was constructed in the moment. It is not on any list.
The first kid is doing what he saw. The second kid is doing what he would do.
Pelé said something about Garrincha that is worth sitting with for a minute.
I don’t know what move he’ll do next, but I know it’ll be his.
Read that twice. The greatest player in the history of the sport, watching his teammate, telling you that he cannot predict the specific action — but he can predict its authorship. He knew the grammar of Garrincha. The output was unforeseen. The signature was not.
That sentence makes no sense if Garrincha was running a catalog. If you knew Garrincha had twelve moves in his repertoire, the next one would be drawn from twelve and you could enumerate the possibilities. But Pelé does not say it’ll be one of his moves. He says it’ll be his. The specific construction is generative. It is built fresh, in real time, from something deeper than a list.
What Pelé is recognizing — without using the word — is that Garrincha had absorbed a grammar of his own play, and was generating from it. The actions were emergent. The list was infinite, because there was no list. There was a way of seeing situations and constructing responses, and the response was new every time.
This is the operation that almost no American youth player is being trained to perform.
Watch a Brazilian kid in a real pickup game. Not academy. Not training. The actual unsupervised hours that built the players we keep importing.
The kid is not doing moves. He is solving problems. The defender is too close, so the foot does something to widen the gap. The ball is bouncing awkwardly, so the chest does something to settle it forward instead of back. A teammate is making a run, so the heel does something to release the ball without the body having to turn. None of these actions are named. None of them were drilled. None of them are reproducible by a coach explaining what just happened.
The kid built them in the moment. He could do it because he had spent thousands of hours next to older kids who were also building things in the moment, watching the way they read situations, absorbing the underlying logic of how problems get solved on a field. He was not learning moves. He was learning a language.
The way you learn a language is not by memorizing sentences. It is by being immersed in fluent speakers long enough that the structure of the language sinks into the part of you that operates faster than conscious thought. After enough immersion, you do not retrieve sentences. You generate them. You say things you have never heard before because the grammar permits them. The grammar is the install. The sentences are the output.
This is what happened to my son.
He did not learn Neymar’s moves. He absorbed the grammar of Neymar’s play. The countless hours of watching the way Neymar reads a defender, the way he uses pace and pause, the way he treats the ball as a tool rather than an obstacle, the specific rhythm and posture and willingness — all of it sank in below the level where you can name it. And then on a pickup field, in a moment that called for it, his body produced a construction Neymar himself may have never made in that exact form, because the grammar of Neymar permitted it.
The output was the kid’s. The state was the install.
I was in my Neymar.
There is a distinction here that is worth naming.
Library is retrieval. Grammar is generation. The first is finite. The kid can only do moves he has seen and practiced. The second is infinite. The kid can construct context-appropriate actions the source player themselves may never have done.
The first feels like trying. The second feels like being.
Neither of them was trained. Both of them learned.
A coach reading this might want to argue. They might say the techniques they teach are useful, that named moves give kids tools, that you cannot generate without first having something to generate from.
They are not wrong about the first part. Techniques are useful. The half-turn is a real skill. The Cruyff turn is a real skill. A player who has never executed a step-over is missing a tool.
What they are wrong about is the order. Techniques installed on top of grammar become tools the grammar can deploy. Techniques installed in place of grammar become a list the kid retrieves from. Same techniques. Different cognitive infrastructure underneath them. One produces a player. The other produces a catalog.
If you want to know whether your kid is running library or grammar, you do not need a test. You need to listen to how he talks about playing.
Library kids talk about moves. I tried this. I was going for that. I almost did the one where. The language is retrieval language. The actions are items pulled from a catalog.
Grammar kids talk about state. I was on. I was feeling it. I was in my [whoever]. The language is being-language. The actions are emergent from a way of seeing the game.
If your kid is the first one, that does not mean he is a bad player. He may be a very good library player. He may execute his catalog at a high level for years. What he will not do, almost ever, is produce a moment of genuine construction — an action no coach taught him, no clip showed him, that fit the moment so well it could not have come from anywhere but inside him.
If your kid is the second one, you are watching something most of the system is not built to recognize. He may not have the cleanest technique. He may not test well at showcases. He may be told he is undisciplined. What he has is the install almost no one else has — the language underneath the moves — and if you protect it long enough, the moves will show up later, sitting on top of something they cannot teach.
What your kid says about his own play tells you which one you are raising.
He was not trying a move.
He was in his Neymar.
That is not a vocabulary you can drill into a child.
That is a child who learned to speak.