Figure 2.6 Percentage of labials in babbling, first words, and the forms predicted by the native tongues of French, American, Swedish, and Japanese children. This bar graph shows the early influence of the native language on the production of labials (from Boysson-Bardies and Vihman, 1991).
The distribution of consonants does not change significantly between babbling and first words. One finds some differences, however. The repertoire of first words is simpler and includes a still higher percentage of occlusives and labials. In fact, the production of a word demands the elaboration of a specific motor program, which imposes constraints on the forms and order of articulatory sequences. If these constraints are too strong, children return to simpler forms and try to adapt them to the word they wish to produce. This leads to different strategies in different children.
The syllable, as a unit of production, grounds the organization of speech. But in young children, motor limitations impose constraints on the structure of syllables. The predominant structures in babbling are simple: consonant-vowel for the most part, sometimes consonant-vowel-consonant. Vowel-consonant-vowel and consonant-vowel-consonant-vowel series are frequently found.
Biomechanical approaches make strong predictions about the consonant-vowel co-occurrences that ought to be found in babbling as well as in the basic syllable of languages. A maximal synergy among the articulators and minimal changes in articulation are assumed in order to predict privileged forms of consonant-vowel association. According to these same approaches, the consonant-vowel structure is strongly privileged, and three associations—labial consonant + central vowel, dental consonant + front vowel, and velar consonant + back vowel—are universally favored (MacNeilage and Davis, 1991).
If one believes that the influence of the local language matters, predictions are a function of particular languages. Thus, Nigerian babies speaking Yoruba, a language in which most of the words begin with a vowel, ought very early to produce more vowel-consonant-vowel forms than French children.1 And, indeed, Yoruba children do merrily mock mechanical constraints and produce many more vowel-consonant-vowel forms than consonant-vow el forms. After all, they have to learn to speak Yoruba. Between 65 and 75 percent of the disyllables produced by the French, English, and Swedish children we studied were of the form consonant-vowel-consonant-vowel, as against 38 percent for Yoruba children, the balance of whose disyllabic productions (62 percent) were of the form vowel-consonant-vowel (Boysson-Bardies, 1993).
Common tendencies are found, of course, but the consonant-vowel associations at ten to twelve months show that the associations predicted by the biomechanical approach and its principle of minimal articulation are found in babbling only when they are also predicted by the structure of the native language. Good learners that babies are, they tend to prefer the most frequent associations in the common words of their native language. Thus, the Yoruba-speaking children whom we called on to help check the data showed a predilection for the forms /ki/ and /ke/, forms that ought to be relatively rare according to motor-based predictions.
This last study further emphasizes how early the selections by the child are made. Between nine and ten months, the interaction between perception and motor performances permits children to organize their babbling. This reveals an agreement between the reorganization of perception and the first productions. The experiments on sound-discrimination capacities suggest that, toward ten to twelve months, babies lose interest in sounds that do not belong to the phonological system of their language. At the same moment, children's performances become oriented toward the production of the vowels, consonants, and syllables privileged in their own language.
Of course, the differences between modes of acquisition and the variability in forms that one finds among children of a given linguistic group indicate the need for caution in generalizing about all the children belonging to it. However, the fact that individual variations do not manage to mask the specific tendencies of different linguistic groups reinforces the view that selective processes organize a phonetic level in production at nine to ten months.
They Begin Speaking Their Language Without an Accent
As we have seen, French adults are able to distinguish the vocalizations of French children of eight months from vocalizations produced by foreign babies. When asked to explain the reasons for their decisions, they point to signs such as intonation and rhythm. But are these kinds of coloring, which are given very early to vocalizations, related to characteristics of prosodic organization specific to French? Prosodic cues make it easier to break utterances into segments, and at two months children are already sensitive to them. But do they use them before their productions come to be organized into words and sentences?
Andrea Levitt and Qi Wang (1991) have analyzed forms of babbling consisting of repeated syllables in French and American children aged seven to eleven months. They were able to identify patterns of pitch and temporal organization of the terminal syllable for these ages that tend to draw it closer to the characteristic form of the terminal syllable of the ambiant language. Thus, lengthened duration of the terminal syllable and a rise in its fundamental frequency (FO) contours were observed much more often in French children than in English-speaking children (54 percent versus 24 percent). Pierre Hallé, B. de Boysson-Bardies, and M. Vihman (1991) have compared temporal organization and FO contours in the disyl-lables of French and Japanese children of eighteen months. Rising voice and terminal lengthening are usual in French children, whereas in the Japanese a falling contour and the absence of terminal lengthening characterize the final syllables of disyllabic productions. These data agree with the characteristics of prosody in French and in Japanese. In Japanese, unlike in French, terminal syllables are not lengthened (see figures 2.7 and 2.8).
One of the most interesting aspects of how children acquire language that remains to be studied is surely the organization of intonation and rhythm at five to six months—before the first productions of babbling occur but after the infant has mastered phona-tion. If prosody has the role attributed to it in the segmentation of continuous speech, the possibility of checking its contribution to the perceptual processing of organized forms through the productions of children seems extremely important. Factors such as oral posture, overall laryngeal position in relation to the modes of production characteristic of certain languages, and organization of intonation contours resembling those of the native language may manifest themselves before six months. This would confirm a parallel evolution, albeit moved forward, of perceptual processing and organization of productions with respect to prosody.
As we have seen, children do not begin by speaking with the rhythm and intonation of a universal Esperanto. The voices of
Was this article helpful?