Language knowledge is at first acquired in an item-specific manner and constructed with isolated words and formulaic expressions on the basis of children's elaboration of the surrounding speech they interact with. But language is organized as a dynamic system with its inner coherence. Such systems usually undergo great reorganizations which can be either progressive or of a catastrophic nature. The emergence of grammatical categories in children's language leads to a reorganization of the whole system (Veneziano 1999, 2001, 2003) in which syntax and phonology are intertwined and cannot be broken up into smaller parts. The system grows into another stage, which we will try to tackle in this paper.
In a longitudinal case-study of a French child recorded at home once a month for an hour, we focused on a) preverbal and pronominal markers: filler syllables, determiners, auxiliaries, modal verbs, pronouns; b) verbal morphology and its variation; c) use of the most frequent phonemes in nominal and verbal morphology marking.
Our aim was to explore whether the development of these linguistic features underwent catastrophic changes.
a) As far as the differentiation between pronominal and preverbal markers is concerned, until 2;0, the child used the same filler syllables. At 2;1, some differenciation appeared. As of 2;2, all preverbal and pronominal markers were clearly differenciated.
b) The child's verbal morphology underwent an abrupt change. Until 2;0, only one verb per recording was used with two different forms (aller and manger). As of 2;1, five to eight verbs were used in two different forms per session.
c) The child's phonological development seemed to change in parallel. The phoneme [a] was very frequently used from the beginning and did not change, but infrequent phonemes such as [r] and [ə] develop mostly was of 2;1. For example, the phoneme [r] took off from 1% before 2;1 to 4% after 2;1 and [ə] from 1,5 to 3%.
This leads us to two conclusions:
- There is a risk of overlooking these phenomena, which give us finer insights into language development, if we only take patent cues on language development such as the MLU or the size of the lexicon into account.
- We observed a great correlation between the three types of markers studied in this paper, indicating a sudden change in the child's system. This correlation seems to indicate that language does not develop out of separate and autonomous components, be they phonological or morpho-syntactical. When children learn to use the right phoneme to mark an accomplished state, they probably extend that marker to the whole system and it enables them to handle accomplished states in a more fluid manner. Therefore, when one type of process emerges, it might have consequences on the child's entire linguistic system.