If human speech and language had been an experiment, the numbers and the time scale would make this the largest experiment in the history of science. Most children learn to talk successfully and without help. But while the entire modern apparatus is shared by the whole of the modern human population, it is still developmentally vulnerable. And such is the complexity of the human speech and language system and the uniqueness of the individual, every case of disorder has to be treated as similarly unique.

But there are patterns in developmental disorders.

By the proposal here, most developmental issues are either by non-activation or by activation at the wrong point or in the wrong way. It would thus appear that the greatest developmental vulnerabilities are with respect to the most recently evolved characters, on which fixation is likely to be least robust.

This vulnerability is quite different to the sort of damage which can occur if a previously healthy brain is deprived of oxygen or if the brain as a whole is affected by some trauma, as in an accident or act of violence.

If there was, as I contend here, an extended process by which the modern human faculty for speech and language evolved, it is not unreasonable to suppose that the emergence of this faculty in childhood is itself highly structured. This was the firm conclusion of Roger Brown (1973). That conclusion has not been significantly disputed, even though the stages posited by Brown are quite different from those posited today.

A child once asked me why he was coming to see me every week. He had a characteristically very specific disorder affecting his speech and nothing else. The next time we met I said that talking was like soldiers marching. Sometimes a soldier decides to have some fun and make another soldier fall over. My answer seemed to work for him.