As far as I am concerned children learn best by the sensation that they are succeeding, on the basis of positive experience, rather than by a focus on the points of maximum difficulty and thus giving the child repeated experiences of failure. By focusing on success rather than failure, it is possible to proceed from one positive experience to another. This lowers the visibility of the intervention, and mirrors the natural process of learning to speak. The only difference between this and normal experience is that there is a shaper focus on particular points of need.
Guiding the growth of understanding
Everyone who knows how to read and write knows that speech can be represented by letters arranged one after another. But that does not mean that that is how things work in the brain. It seems almost certain to me that what is stored in the brain is in fact quite different, more like a highly-compressed file of some computer software. So the child has to learn how the compression and de-compression work.
The natural process of learning to talk normally takes around ten years. But not always. And that that is what this website is all about.
Whichever language a child is learning, he or she has to work out what sort of a language it is, what it allows and doesn’t allow, and so on. And the evidence of actual speech can be quite confusing. But there are three vital things to work out: How the system of speech sounds or ‘phonemes’ is organised; How the phonemes combine in what is known as the ‘phonotactics’; And how the rhythm is expressed in what is known as the ‘metricality’ or ‘prosody’.
Most of the keys to the natural learning process are in children’s brains, not in their mouths. Such keys are not easily turned by exercises of the sort found useful in training for sport or playing music. By the natural process, the child has to mentally go over what he or she has heard, and turn this into speech and language. This process can be guided. But such guidance is quite unlike what the gardener does with the shoots of plants or vines.
Ways of cutting to the chase
Children find various ways of cutting to the chase. One (wrong) way – for children learning English – is to pronounce the syllable carrying primary stress, and either ignore the rest or reduce it to a very unclear series of gestures. So potato becomes something like (X) TAY (X), where the Xs represent possible gestures which can’t be easily or accurately represented. Another wrong way is to deduce from words like dinner and table that stress falls on the first syllable. That may work for many words, but not for most words of more than two syllables. By a better approach, what is known as the scansion starts correctly from the right, but then incorrectly stops as soon as the stressed vowel is pronounced. So everything before the stressed vowel goes unpronounced. So the word monopoly is mispronounced as OPOLY.
The only way to find out what is really happening in the child’s mind is to analyse what the child can say, what he or she gets wrong and how extensively, and what he or she refuses to try to say.
It is sometimes thought that for the sake of getting a numerical count of a child’s progress in speech it is possible to score the Percentage of Consonants Correct or the PCC. But the PCC fails to take account of the effects of word stress.
By what I believe is a better clinical approach, it is important to take account of all aspects of acquisition, not just the phonemes and the phonotactics, but the stress and the way these things are integrated into conversational speech, as expressions of what the child knows. This is not behaviour, but knowledge of a special sort, knowledge which can be investigated.
Suppose the child is shown a picture of an magnet, told what it is (if this is necessary, and it isn’t always), and then asked to tell someone else, if there is anyone else to tell. (As long as the child has not reached a certain level of sophistication, this is an acceptable game.) And the child then names the object as a MAGNIK. In this case a feature of the G sound, the fact that it is articulated with the back of the tongue, is copied into the final sound. But only this feature. What makes the G sound different from the K, the feature which defines the relation between two timings, the release of the closure in the mouth, and the opening of the vocal cords in the voice box, is realised correctly, as is the stress. The error is in just one feature which is incorrectly copied from right edge of the stressed syllable to the right edge of the final, unstressed syllable.
Clinical linguistics as treatment
The clinical linguistic task is to find what can be said correctly or the closest possible approach to this. This is not necessarily a real word. There is a very large number of ways of fractionally changing the sound structure to create a possible word which the child can hopefully say correctly. The M is said with the lips and the N with the tongue tip. Either of those could be changed into the other. The G could be changed minimally into a D or a B, or more radically into a K, T, or P. The same changes could be run in the final T. The stressed vowel could be changed. Or either the G or the T or both could be deleted. And bigger changes could be made in either or both of the vowels. Or a vowel could be added in the middle or at the end, with various possibilities for the metrical structure. Multiplying these out, the total number of possible forms is 368,640, obviously far more than any child could ever be asked to try to say. But with a child saying magnet as MAGNIK, I would hope to enable the child to say the word correctly in something between 5o and 500 trials. After any given trial, the child may be aware that the word did not come out quite right. But whether the child is aware of the error or not, in my experience there is no need to point the error out, unless the child asks to be told, and some children do. To avoid confusing the child’s system, it is usually necessary to make two changes from one trial to the next. I might go from say NOGNI, to DAGNI, to DEGNER, hopefully getting all of these correct, before eventually coming back to magnet, and hopefully getting that right too for the first time. Sometimes at the end of a session working on a particular structure, it still isn’t right. But a few days or a week later it is right. And usually once the child has it right, it stays right.
With most small children it is often important to preserve the game environment, and thus to create an image or a model for every trial. Felt tip pens and paper, plasticine, sticks of carrot, or pipe cleaners all come into their own here,
At a later point in development a child might say archaeopteryx as ARTIOPTERYX, reversing the same two features, but in the opposite direction. The child can say the K sound as the first part of the final X, but not as the first consonant in the second syllable. Its formation does not proceed in this environment, at the beginning of a secondary stressed syllable where there is another syllable with a similar secondary stress, two syllables to the right. Here, with five syllables rather than two, the combinatorix of possible changes is exponentially greater. But the same principles apply.
The sense of possible words
The critical factor here is not the amount of practice, or the experience of correction, but the child’s sense of the ‘space’ of possible words. (I call this a space because it has a number of dimensions, according to the features and variations on the metricality and phonotactics.) This sense is obviously both subtle and unconscious. It normally develops over eight or nine years. But it can change overnight.
It seems useful to me to regard the very complex and variable ways that children say particular words as failures to complete the sound structures, rather than as well-defined processes. These failures can occur at any of a number of points. The set of all the possible points of failure is vastly larger than any published list of processes.