An empty room with a doorway

The human ways of sharing post left a child in the middle of a world already talking, going and getting language for itself. It is worth asking why it took so long to develop, and why the schools that dominated the study of the mind for most of a century see it differently.

The first leaves the inside empty. Behaviourism’s subject is an automaton: stimulus in, response out, reinforce and repeat. Between the two there is, by doctrine, nothing worth discussing — a black box, and the point is that you needn’t open it. Watson and Skinner made a virtue of the closure. But the active subject the developmental picture describes is precisely what lives in that box. Everything the child does — the reaching, the trying, the building — is the middle term behaviourism throws out on principle.

The second gives the inside a basement. Psychoanalysis has a vivid interior — drives, fantasy, buried history, a whole theatre running its own drama out of reach and often at odds with the plain world outside. But reality comes in two, an inner and an outer, and the inner one holds the reins. The developmental picture declines the second reality. A child builds in private, but it builds one reality — the same one it construes and shares in.

The third is the interesting one, and it is still ongoing.

It begins with a genuinely great argument. Noam Chomsky, in 1959, took Skinner’s account of language apart — the child learns far too much, far too fast, from far too little, for conditioning to be the story. Nobody supplies a list of the sentences that are impossible, correction is rare and ignored, and yet every normal child lands a grammar in a few years. Something must be coming from the child’s side. Behaviourism could not say what, because it had declared the question closed.

If the child couldn’t have got it from experience, says nativism, then it was already there. The inside is not empty and not split — it comes pre-filled. Grammar arrives installed; exposure sets the switches. The child doesn’t build the grammar. It finds it.

What happens over the next sixty years is the part worth watching. The endowment starts out rich — whole systems of innate rules. By the 1980s it is universal principles plus a set of binary switches. By the 1990s the Minimalist Program has stripped it toward a single operation, Merge, that joins two things into one and can run on its own output. And the field is explicit about why it shrank: universal grammar has to be rich enough to explain what the child gets from thin evidence, and small enough to have plausibly evolved. Those two pull opposite ways. The theory retreated under the second, not because the first found new evidence.

Meanwhile the question of where the pieces come from — the things Merge combines — stays open. Chomsky’s answer is that lexical concepts are too complex to have been extracted from experience, so they must already be there too, triggered rather than learned; he has said this holds even for carburetor. Jerry Fodor took it to the end of the line: concepts can’t be learned at all, only triggered, so car is innate. Putnam could hardly conceal his astonishment. Dennett observed that on this view Aristotle had the concept of an aeroplane in his brain. The field’s own verdict is that Fodor went too far — but nobody leaves the position that pushed him there. And in Why Only Us, Berwick and Chomsky write that the origin of those atoms “remains for us a mystery.”

What the sixty years show is a stance taken early that then had to be held. This is science fighting a war in the trenches, actively non-pluralistic, each holding one position. The war was never about the findings — it was only about the tribe.

This post is part of the seed series. See also the reference entry on nativism.

Photo: Iñaki del Olmo / Unsplash