Microanalyses of unstructured videotaped interactions of three mother-infant dyads revealed dramatic shifts in dyadic vocalization patterns from primarily overlapping to primarily alternating. Maximal overlapping vocalization appeared between 7 and 13 weeks for the different dyads, and subsequent alternation predominance peaked between 12 and 18 weeks. The findings are compared with prior work which had not found alternation predominance and reasons for the difference are suggested. The early emergence of a sequence of predominantly overlapping vocalization followed by predominantly alternating vocalization may be linked to the concomitant development of multimodal sensory-sensory and sensory-motor integration. If so, then the emergent patterns reflect increased potential for co-ordination within the dyad, which may be capitalized upon by the mother and by the infant.

Maternal speech to 3-month-old infants was compared for American and Japanese mother-infant dyads. Utterances were analysed at the levels of function, form and referent, and in relation to infant gaze direction. Loglinear categorical analysis revealed that infant gaze affected the intended functions of maternal speech differently for the two cultural groups. Cultural differences were also seen in the nature of function-form, and function-referent relationships. The differences seem to be consistent with culture-characteristic profiles of maternal behaviour towards older infants and children. It is suggested that transmission of culture through maternal speech starts at the earliest stage of infant development.

Three factors have been hypothesized to play an important role in the reduction of children's initial overextensions: spontaneous adult use of the correct label, correction of the child's errors, and demonstration of the important attributes that make an object a member of its adult category. The role of these factors was examined in relation to data collected from a longitudinal study of early lexical development. This study used an observational methodology combined with systematic comprehension and production testing. Results indicated that demonstrations were the most important factor in inducing toddlers to assign an object to its adult category. The question of why purely linguistic input initially plays a minor role in changing children's categories is discussed.

This article uses longitudinal data from four Mexican-American children to explore two aspects of the acquisition of Spanish word stress that precede and accompany learning of the stress system itself. First, contrary to Allen & Hawkins' proposed universal ‘trochaic bias’ (Allen 1982, Allen & Hawkins 1977, 1979, 1980), it is shown that children have a ‘neutral start’ in stress learning: they approach the task of stress learning unbiased towards any particular stress type. Secondly, several examples are found in which children's attention to phonetic or semantic aspects of normatively unstressed syllables leads them to shift stress to that syllable.

This paper explores children's early talk about specific distant past events and its development into conversational stories of personal experience, with special focus on its evaluative function. The data consist of longitudinal home observations of five working-class children and their mothers from ages 2; 0 to 2; 6. Results indicate that during this period the children talked primarily about negative past events, especially events of physical harm; the rate of talk about specific past events doubled; temporally-ordered sequences increased dramatically; and the children became better able to accomplish such talk independently. In addition, one of the most striking findings is the extent to which these 2–year-olds were able to communicate their attitude toward the past event, with speech about the past containing five times more evaluative devices than other speech. These results suggest that (1) by age 2; 6 stories of personal experience emerged in incipient form, and (2) the roots of this genre lie not only in cognitive skills and social interactional support but in the emotional significance of the depicted event.

The Principle of Contrast, that different words have different meanings, holds for adult language use. But at what age do children assume Contrast ? Do they rely on it from the start, or do they assume that new words may have the same meaning (the Null Hypothesis) until they discover otherwise ? Both the Null Hypothesis and Contrast have certain consequences. The Null Hypothesis places a heavy burden on languagelearners, whereas children could discover Contrast as part of their experience of rational behaviour. Examples that have been claimed to go counter to Contrast fall into two groups. Those in the first do not in fact violate Contrast at all. Those in the second rely on sameness of extension instead of sameness of meaning, and so are indeterminate as counter evidence. Usage consistent with Contrast, on the other hand, is pervasive in children's speech from an early age.

Recent linguistic studies have proposed that verb morphology has evolved to serve particular functions inherent to discourse. In this paper, an attempt is made to apply this proposal to the development of verb morphology in the spontaneous speech of a child from 1;10. 16 to 2;0.2. The question asked was whether there were any aspects of the child's discourse which may have been in part responsible for her differential use of verb morphology. To answer this question, a distributional analysis is presented of the child's speech in two different speech contexts: dialogue vs. crib-monologue. Despite the exiguous amount of data in the corpus, the analysis yields striking patterns of co-occurrences involving verb morphology and forms of self-reference. The patterns were interpreted as provisional evidence that the child may have been sensitive to discourse factors in her selective use of verb morphology. The data is analysed into four developmental phases. It is suggested that the semantic-level meanings of Phase IV are due in part to the type of discourse use to which the morphology is put in the earlier phases.

A new corpus of spontaneous conversations between adults and children is examined for evidence that adults simplify their vocabulary choices when speaking with young children. If these simplifications are found to be age-dependent, then they would broaden the pattern of simplifications characteristic of ‘motherese’ to include lexical choice as well. For the age-range newborns to 12 years, the results are both consistent with and contrary to the attested set of grammatical simplifications. In this corpus, MLU and TTR are strongly age-dependent, but adults do not choose their words from the 10,000 most common word-types in English in an age-dependent manner. Rather, the additional types for school-aged children come from the same part of the vocabulary and share the same-shaped distributions as in adult speech with preschool children and infants. This absence of an age-dependent accommodation in word choice has implications for models of child lexical acquisition which assume adult language accommodation.

This study determined at what age and in what form children (aged 4;5 to 11;11) produce speech acts which commit them to some future action, by having them speak a doll's part in dialogues with the experimenter, who spoke for another doll. All children produced directive speech acts, indicating that the task was not too difficult for them, but only the commissive speech acts were analysed in detail. A majority of children at all ages produced at least one commissive speech act, but only the older children used the explicit performative verb promise to reassure the hearer of their commitment.

Little is known about children's acquisition of modality. It seems clear that at the age of five, when they are said to be linguistically competent, children have not mastered the adult system of modal meaning. This paper describes research which tests children's understanding of modal meaning at the ages of eight and twelve. The results of these tests are compared with the results of the same test on adult informants. The research uses the card-sorting method devised by Miller (1971) to investigate semantic similarities and dissimilarities; cluster analysis of the data reveals the underlying patterns. My tests revealed that eight-year-old children have only a rudimentary system of modal meaning, and even by the age of twelve a child's system will not be isomorphic with the adult system.

The present investigation examined the relationship between attribute saliency and metaphor interpretation in school children. Ortony's theory that metaphors entail salience imbalance prompted the generation of two types of metaphors, PREDICATE-PROMOTING (PP) and PREDICATE-INTRODUCING (PI) metaphors. An adult sample was used to select metaphors of each type which were then presented to 24 children in each of grades three, five, and seven (mean ages 8; 5, 10; 6 and 12; 8). The children were asked to generate attribute lists for TOPIC and VEHICLE terms and, six weeks later, to verbally explain metaphors containing those terms. Older children correctly interpreted more metaphors than younger children, and at each grade level no difference was observed between the number of correct interpretations of PP and PI metaphors. In addition, at all grade levels the incorrect ground of each child's interpretation errors most often had been listed in his/her attribute lists. This suggests that attribute saliency for the individual perceiving the metaphor plays a key role in the interpretation process.