Two major strategies for linguistically encoding an idea or proposition are suggested. The first strategy involves encoding an idea in the space of a single utterance, while the second strategy conveys the proposition through a sequence of two or more utterances. The tendency has been to focus on discourse as a composite of sentences (the first strategy) rather than on its role as an alternative to the sentence. The extent to which speakers rely on one strategy or the other is both situation-specific and language-specific. It is proposed that languages may differ in the extent to which they use discourse for conveying a single proposition and that this is linked to communicative strategies employed by the language-learning child. The discussion focuses on the sequential strategy for encoding a proposition, from a functional perspective. First, the pragmatic functions performed by each utterance are evaluated, i.e. its role as an attention-getting device, an acknowledgment, etc. Second, the logical function of each utterance within the sequence is considered, i.e. its role as argument, predicate, modifier, etc. The findings demonstrate the importance of considering the proposition rather than the utterance as a natural unit and they provide an alternate measure for assessing the child's level of linguistic competence. (Author/NCR)