As a goal in treatment
Take hippopotamus as one of the longest words that small children are likely to hear and want to say. it has two feet, HIPPO, and POTA, the second also a separate word, and a final unstressed syllable, MUS, discounted for the purpose of right-to-left stress computation in the framework here, from the 1987 work of Morris Halle and Jean Louis Vergnaud and from the 1991 work of Bill Idsardi. Both feet have a stress on the first syllable. In the second foot, this becomes a primary word stress.
Many children say hippopotamus as HITTAPOTAMUS, seeming to copy the T from its position after the primary stressed syllable to a position after the secondary stressed syllable, replacing one feature of the P, the position at which the air stream is closed. P does this with the lips, and T with the tongue tip. HITTAPOTAMUS is by far the commonest child error in this word. It mimics one of the ways in which the structure of the word can be changed – in a way easily recognisable as such by normally developing children of seven or so. WECKIMUDDINIT changes every sound in the structure. But this can be done one sound at a time in any logically possible combination,.
The uneven distribution
There are nowhere near as many actual words as there are possible words. English speech sounds are distributed quite unevenly. Despite the large number of possible syllables in English, many either occur only in one syllable words, or are rare and unlikely to be encountered by children other than in special cases. Soldier has a structure which seems to be unique in English, with L in the coda of the stressed syllable and the affricate J in the onset of the final syllable. Many normally developing children say it like shoulder, moving a property from the right edge of what sounds like a J to the initial S, leaving the D without its right edge. But the uniqueness of soldier in the lexicon makes this hard or impossible to analyse.
A child I once treated for a severe disorder said the word as HOWOOV where the OO is as in Southern British foot. Here the child correctly analysed the lip rounding of the vowel in the oh sound and the almost complete closure of the mouth at the beginning of the second syllable, writtten as DI, pronounced as a J, but pronounced what should have been a J sound by a W, the S as an H, and mirrored the roundness of the two vowels as pronounced in a final lip articulation in V. Another child with the same sort of ‘process’ was reported by Pamela Grunwell (1987).
Some children with speech disorders say finger as TIN-NER, with none of the consonants correct, and with the N, not pronounced as an N by a correct pronunciation, but doubled in the middle of the word in a way not allowed in English. Here the tongue tip property of N is pulled left and right, overwhelming the G and pronounced as a tongue tip T in the leftmost consonant.
Most unstressed syllables are simple. So spaghetti may be the only word likely to be heard and known by children with a syllable beginning with SP before the stressed syllable. This uneven distribution is easily misconstrued by children. They may, for instance, assume that if a particular sort of structure sounds wrong, it is wrong, when actually it isn’t wrong at all, but just unusual. Some errors in child speech may be the result of this sort of misanalysis.
Many children spaghetti as BASKETI. But a few say it as PSKETI. They may be assuming either that English allows a PSK cluster at the beginning of the word or that this is a complex phoneme. Of course, such mistaken assumptions are not conscious, but strictly unconscious. Either of them, unfortunately, leads to a mispronunciation, in the case of PSKETI, not a possible word in English.
On all of these points, there is the possibility of the child learner either misconstruing or misanalysing the evidence of what he or she hears, or deciding that the evidence is not clear enough for a firm decision. The particular combination of complexities in English metricality, syllable structure, and the phonemic inventory, is especially problematic for some children. Some of these complexities are beyond the scope of any plausible instructions. The speech won’t come out right unless the analysis on all of these points is correct.
In a more complex way, some children with clinical speech problems say cardigan as KARDINTON. Here the G sound has turned into a D. And then that has turned into a T. And the final N has got copied into the previous syllable. That is seemingly by three steps, each one clear and simple in itself, but in something which looks like a sequence because it seems that no child says anything like KARDIKAN or KARDINGAN with the steps happening with the sequence varied.
In a similarly complex way, I once heard a child of four and a half saying cardigan as KARDALUNG, with some of the G sound moving right, some of the N sound moving left, and then turning into a sort of L, in other words becoming more different from the target sound structure than it should have been. The speech could hardly be understood at all.
With all the complexity that English allows in the phonemic, syllabic, and stress systems, there is also significant variation in how these things interact from one variety of English to another and probably in the speech of individuals. Consider an apparently simple word like little. When it is said as one of the first words, say between 18 months and 2 years, very few children say it in a way that seems correct in any of the varieties they are likely to hear around them. The first L is what is often described as ‘clear’, articulated with a brief contact with the tip of the tongue just behind the upper front teeth, letting the airflow pass smoothly past each side. The stressed I is relatively unproblematic, articulated with the tongue high and forward in the ‘vowel space’, but not at the extremity of the space, and maintained only very briefly. The T in the middle of the word is appropriately written with two letter Ts, because it has separate roles in two syllables, ending one and beginning another. In its role as role at the beginning of a syllable it may be not released as a stop. The fact that it contrasts with the D in middle is marked by a brief pause in the vibration of the vocal folds. The second L represents what may be a plain tongue tip ‘lateral’ in the underlying lexical representation. The opening to the airstream by the two sides of the tongue marks the release of the T. Without any features defining it as a vowel, it is not allowed to carry any stress. It is pronounced very ‘darkly’, in a way more like the ER schwa in dinner. But in a way characteristic of small English children’s speech the T DISSIMILATES to K in little and G in middle – contrasting more sharply with the tongue-tip L segment in the rime of the second syllable before it becomes a short OO vowel. Thus characteristically, children increase the vocalisation of the L, losing all trace of any underlying representation with the tongue tip, and at the same time replacing the tongue tip articulation of T or D by a back of the tongue articulation, giving something like LIKU or MIGU.
What children are likely and unlikely to hear
Children can hear different varieties of English from their parents, releasing the T and vocalising the L to different degrees.
In another area of the system, one that small children are much less likely to hear, in the words huntsman, appointment and ointment, in what may be all varieties of the language, the T between the N and the M is replaced by a complete stopping action by the vocal cords, a ‘glottal stop’, and voicing is mostly or completely lost in the final sequence of nasals. But despite the rarity of these words in everyday conversation for most people, the process seem to be reliably learnable. So stipulation does not cut the mustard here. Learners must be able to deduce this special case from others more reliably available to common experience.