The proposal that young children's communicative intentions stem from prelinguistic cognitive abilities is examined in detail. The most developed available formulation of this proposal, that provided by Brown (1973), is evaluated, and the evidence in support of it is found to be insufficient. Three crucial problems, which must be solved before an acceptable version of the proposal can be formulated, are raised. These are: (1) determining prelinguistic cognitive abilities, (2) individuating children's communicative intentions, and (3) finding criteria for determining whether a communicative intention stems from prelinguistic cognitive abilities. Approaches to solving these problems are suggested. (Author)