Learning in Trackton In the black community of Trackton, babies are not spoken to, but are held facing outwards, so that they can feel the person who is holding them, and watch the bodies of those who are conversing with the caretakers :

Encapsulated in an almost total human world, Trackton babies are in the midst of nearly constant human communication, verbal and nonverbal. They literally feel the body signals of shifts in emotion of those who hold them; they are never excluded from verbal interactions. They are listeners and observers in a stream of communication which flows about them, but is not especially channeled or modified for them. Everyone talks about the baby, but rarely to the baby. Childtenders direct their talk about the baby to others : 'Dis young 'un wet his britches moren' any young 'un I know.' ... Only occasionally do older children talk to the baby while nudging, cuddling or fondling him.


Heath, p. 75.

Coaching is useless

... parents who consciously try to 'coach' their children by simplifying and repeating may be actually interfering with their progress. It does not pay to talk to children as if one was telling a foreign tourist how to get to the zoo. Language that is impoverished is harder to learn, not simpler. Children appear to be naturally 'set' to extract a grammar for themselves, provided they have sufficient data at their disposal. Direct teaching is irrelevant, and those who get on best are those who are exposed to a rich variety of language - in other words, those whose parents talk to them in the normal way.

Aitchison, p. 72.

Chomsky on Language Faculty From a formal point of view, the grammar that is internalized by every normal human being can be described as a theory of his language, a theory of a highly intricate and abstract form that determines, ultimately, a connection between sound and meaning by generating structural descriptions of sentences ... each with its phonetic, semantic, and syntactic aspects. From this point of view, one can describe the child's acquisition of knowledge of language as a kind of theory construction. Presented with highly restricted data, he constructs a theory of language of which this data is a sample (and, in fact, a highly degenerate sample, in the sense that much of it must be excluded as irrelevant and incorrect - thus the child learns rules of grammar that identify much of what he has heard as ill-formed, inaccurate, and inappropriate). The child's ultimate knowledge of language obviously extends far beyond the data presented to him. In other words, the theory he has in some way developed has a predictive scope of which the data on which it is based constitute a negligeable part. The normal use of language characteristically involves new sentences, sentences that bear no point-by-point resemblance or analogy to those in the child's experience. Furthermore, the task of constructing this system is carried out in a remarkably similar way by all normal language learners, despite wide differences in experience and ability.

Chomsky, pp. 170/1.

Hagège's objection to the LAD La transmission héréditaire d'une capacité d'apprendre à parler, ou même d'un schéma fixe régulateur de la langue, ne peut suffire à rendre compte de l'apprentissage tel qu'on le voit se déployer. Certes, la faculté de langage ne saurait être elle-même apprise. Mais comment pourrait-elle à elle seule rendre compte de l'acquisition de la langue entre vingt-deux mois et 3-4 ans, si l'imitation des adultes, elle-même articulée sur la capacité d'intégrer ce qui est imité, ne jouait un rôle essentiel?

Hagège, 31-2

Baby is listening Les nouveaux-nés de quelques jours prérèrent écouter la voix de leur mère quand cell-ci est présentée en concurrence avec celle d'une autre mre parlant à son bébé. Mais il faut que l'intonation de la mère soit naturelle. Si l'on joue la bande à l'envers, la préférence de l'enfant ne se voit plus. Cette préférence est liée aux aspects dynamique de la parole maternelle, telle l'intonation, et non à des aspects statiques des sons puisque ceux-ci sont préservés lorsqu'on fait passer la bande à l'envers ...
Préférence pour la voix de la mère et préférence aussi pour la langue maternelle. Lorsque des séquences de français succèdent à des séquences de russe, les nourrissons français de quatre jours montrent une reprise plus importante de la succion que lorsque les séquences sont présentées dans l'ordre inverse ... Cette "préférence" se maintient lorsque les séquences ont été filtrées de façon à enlever la majeure partie de l'information phonétique tout en laissant intact la prosodie. Les différences prosodiques entre la langue maternelle et la langue étrangère sont donc suffisantes pour susciter une réaction plus vive lors de la présentation de la langue maternelle.

Boysson-Bardies, pp. 34-5