As is well known the first tentatives in the use of the common speech-forms are very rough. The child in reproducing transforms, and these transformations are often curious and sufficiently puzzling.
The most obvious thing about these first infantile renderings of the adult’s language is that they are a simplification. This applies to all words alike. Monosyllables if involving a complex mass of sound are usually reduced, as when ‘dance’ is shortened to ‘da’. This clearly illustrates the difficulty of certain sound-combinations, a point to be touched on presently. More striking is the habitual reduction of dissyllables and polysyllables. Here we note that the child concentrates his effort on the reproduction of a part only of the syllabic series, which part he may of course give but very imperfectly. The shortening tends to go to the length of reducing to a monosyllable. Thus ‘biscuit’ becomes ‘bik,’ ‘Constance’ ‘tun,’ ‘candle’ ‘ka,’ ‘bread and butter’ ‘bup’ or ‘bŭ’. Polysyllables, though occasionally cut down to monosyllables, as when ‘hippopotamus’ became ‘pots,’ are more frequently reduced to dissyllables, as when ‘periwinkle’ was shortened to ‘pinkle’. Handkerchief is a trying word for the English child, and for obvious reasons has to be learnt. It was reduced by the eldest child of a family to ‘hankish,’ by the two next to ‘hamfisch’ and by the last two to ‘hanky’. The little girl M. also reduced the last two syllables to ‘fish,’ making the sound ‘hanfish’.
There seems to be no simple law governing these reductions of verbal masses. The accentuated syllable, by exciting most attention, is commonly the one reproduced, as when ‘nasturtium’ became ‘turtium’.[[77]] In the case of long words the position of a syllable at the beginning or at the end of the word seems to give an advantage in this competition of sounds, the former by impressing the sound as the first heard (compare the way in which we note and remember the initial sound of a name),[[78]] the latter by impressing it as the last heard, and therefore best retained. The unequal articulatory facility of the several sound-combinations making up the word may also have an influence on this unconscious selection. I think it not unlikely, too, that germs of a kind of æsthetic preference for certain sounds as new, striking or fine, may co-operate here.[[79]]
Such simplification of words is from the first opposed, and tends in time to be counteracted, by the growth of a feeling for their general form as determined by the number of syllables, as well as the distribution of stress and any accompanying alterations of tone or pitch. The infant’s first imitations of the sounds ‘good-bye,’ ‘all gone,’ and so forth, by couples which preserve hardly anything of the articulatory character, though they indicate the syllabic form, position of stress, and rising and falling inflection, illustrate the early development of this feeling. Hence we find in general an attempt to reproduce the number of syllables, and also to give the proper distribution of stress. Thus ‘biscuit’ becomes ‘bítchic,’ ‘cellar’ ‘sítoo,’ ‘umbrella’ ‘nobélla,’ ‘elephant’ ‘étteno,’ or (by a German child) ‘ewebón,’ ‘kangaroo’ ‘kógglegoo,’ ‘hippopotamus’ ‘ippenpótany,’ and so forth.[[80]]
As suggested above there goes from the first with the cutting down of the syllabic series a considerable alteration of the single constituent sounds. The vowel sounds are rarely omitted; yet they may be greatly modified, and these modifications occur regularly enough to suggest that the child finds certain nuances of vowel sounds comparatively hard to reproduce. Thus the short ă in hat, and the long ī (ai), seem to be acquired only after considerable practice.[[81]] But it is among the consonants that most trouble arises. Many of these, as the sibilants or ‘hisses,’ s, sh, the various l and r sounds, the dentals, the “point-teeth-open” th and dh (in ‘thin,’ ‘this’), the back or guttural ‘stops,’ i.e., k and hard g, and others as j or soft g (as in ‘James,’ ‘gem’), appear, often at least, to cause difficulty at the beginning of the speech period. With these must be reckoned such combinations as st, str.
In many cases the difficult sounds are merely dropped. Thus ‘poor’ may become ‘poo,’ ‘look’ ‘ook,’ ‘Schulter’ (German) ‘Ulter’. In the case of awkward combinations this dropping is apt to be confined to the difficult sound, provided, that is to say, the other is manageable alone. Thus ‘dance’ becomes ‘dan,’ ‘trocken’ (German) becomes ‘tokko’. More particularly s and sh are apt to be omitted before other consonants. Thus ‘stair’ becomes ‘tair,’ ‘sneeze’ ‘neeze,’ ‘schneiden’ (German) ‘neida,’ and so forth.
Along with such lame omissions we have the more vigorous procedure of substitutions. In certain cases there seems little if any kinship between the sounds or the articulatory actions by which they are produced. At the early stage more particularly almost any manageable sound seems to do duty as substitute. The early-acquired labials, including the labio-dental f come in as serviceable ‘hacks’ at this stage. What we call lisping is indeed exemplified in this class of infantile substitutions. Children have been observed to say ‘fank’ for ‘thank’ and ‘mouf’ for ‘mouth,’ ‘feepy’ for ‘sleepy,’ ‘poofie’ for ‘pussy,’ ‘wiver’ for ‘river,’ ‘Bampe’ for ‘Lampe’ (German). The dentals, too, d and t, are turned to all kinds of vicarious service. Thus we find ‘ribbon’ rendered by ‘dib,’ ‘gum’ by ‘dam,’ ‘Greete’ (German) by ‘Deete,’ ‘Gummi’ (German) by ‘Dummi,’ ‘cut’ by ‘tut,’ and ‘klopfen’ (German) by ‘topfen’. Similarly ‘gee-gee’ (horse), which oddly enough was first rendered by the child M. as ‘dee-gee,’ is altered to ‘dee-dee’. I find too that new sounds are apt to be put to this miscellaneous use. Thus one child after learning the aspirate (h) at two years not only brought it out with great emphasis in its proper place but began to use it as a substitute for other and unmanageable sounds. Thus he would say, ‘hie down on hofa’ for ‘lie down on sofa’. The aspirate is further used in place of sh, as when ‘shake’ was rendered by ‘hate,’ and of st, as when Preyer’s boy called ‘Stern’ ‘Hern’. In other cases we see that the little linguist is trying to get as near as possible to the sound, and such approximations are an interesting sign of progress. Thus in one case ‘chatterbox’ was rendered by ‘jabberwock,’ in another case ‘dress’ by ‘desh,’ in another (Preyer’s boy), ‘Tisch’ (German) by ‘Tiss’.[[82]]
Besides omissions and substitution of sounds, occasional insertions are said to occur. According to one set of observations r may be inserted after the broad a, as when ‘pocket’ was rendered by ‘barket’. A cockney is apt to do the same, as when he talks of having a ‘barth’ (bath). Yet this observation requires to be verified.
These alterations of articulate sound by the child remind one of the changes which the languages of communities undergo. We know, indeed, that these changes are due to imperfect imitation by succeeding generations of learners.[[83]] Hence we need not be surprised to find now and again analogies between these nursery transformations and those of words in the development of languages. In reproducing the sounds which he hears a child often illustrates a law of adult phonetic change. Thus changes within the same class of sounds, as the frequent alteration of ‘this’ into ‘dis,’ clearly correspond with those modifications recognised in Grimm’s Law. So, too, the common substitution of a dental for a guttural has its parallel in the changes of racial language.[[84]] Nobody again can note the transformation of n into m before f in the form ‘hamfish’ for ‘handkerchief’ without thinking of the Greek change of συν into συμ before β, and like changes. Philologists may probably find many other parallels. One of them tells me that his little girl, on rendering sh by the guttural h, reproduced a change in Spanish pronunciation. M. Egger compares a child’s rendering of ‘trop’ (French) by ‘crop’ with the transformation of the Latin ‘tremere’ into ‘craindre’.
I have assumed here that children’s defective reproduction of our verbal sounds is the result of inability to produce certain sounds and not due to the want of a discrimination of the sounds by the ear. This may seem strange in the light of Preyer’s statement that the earlier impulsive babbling includes most, if not all, of the sounds required later on for articulation. This may turn out to be an exaggeration, yet there is no doubt, I think, that certain sounds, including some as the initial l which are common in the earlier babbling stage, are not produced at the beginning of the articulatory period. As the avoidance of these occurs in all children alike it seems reasonable to infer that they involve difficult muscular combinations in the articulatory organ. At the same time it seems going too far to say, as Schultze does, that the order of acquisition of sounds corresponds with the degree of difficulty. The very variability of this order in the case of different children shows that there is no such simple correspondence as this.[[85]]