Journal of Pragmatics 9 (1985) 527-546 North-Holland 527 zyxwvutsrqponm THE MEANING OF CHILDREN’S FIRST WORDS Evidence from the Input Anat NINIO * This study is an investigation into the meaning of the first words children acquire, through an examination of the input to the semantic learning device. An analysis of the way word meanings might be learned shows that (a) the set of words potentially learnable at the very onset of language acquisition is the set of words addressed to the child as single-word utterances, and (b) that the most likely description of the meaning children attribute to one-word utterances is in terms of the intentional communicative acts speakers perform in uttering these utterances. One-word utterances of 8 mothers to preverbal infants were analyzed for their communicative function. It was found that words appearing as one-word utterances in the input possess a high degree of function- specificity and even unifunctionality. The results imply that the first word meanings children learn are probably rules for the lexicalization of specific communicative acts. 1. Introduction The subject-matter of this investigation is the question: what is the meaning of the first words children acquire? Meaning is a mental phenomenon, and since young children cannot introspect for us, it is unamenable to direct investiga- tion: The usuai strategy for attacking this question has been to examine the output of the ‘black box’, namely, children’s production of words and their comprehension responses (cf. Bowerman (1976)). The study of children’s early productions has resulted in a multitude of conflicting opinions, not only about what word meanings are for children but also about what children mean by their one-word utterances (cf. De Laguna (1927), Braine (1974) Greenfield and Smith (1976) Bloom (1973) Bloom and Lahey (1978) Halliday (1975) Schle- singer (1982) Dore (1975) Bates (1976) Bruner (1975)). Far from being an easy way into the child’s unobservable semantic system, research into the meaning of early utterances complicates the issue, because it is not at all clear whether it is necessary to distinguish between what words mean to the child (i.e., their sense) and what children mean by them (cf. Greenfield and Smith (1976)). The basic problem with this approach is that a correct interpretation * Author’s address: A. Ninio, Dept. of Psychology, The Hebrew University of Jerusalem, Jerusa- lem 91905, Israel. 0378-2166/85/$3.30 0 1985, Elsevier Science Publishers B.V. (North-Holland)