Charles-Luce & Luce (1990) found smaller phonological similarity neighbourhoods in five- and seven-year-old children's expressive lexicons than in an adult receptive lexicon, a finding they interpreted as evidence that children need not employ fine-grained auditory perceptual analyses in lexical processing. In the present investigation, neighbourhood sizes were calculated for an expressive lexicon derived from two vocabulary lists representative of children aged 1;0 to 3;0 (Rescorla, 1989; Reznick & Goldsmith, 1989). Over 80% of the words in these early lexicons had at least one phonological neighbour; nearly 20% had six or more phonological neighbours. Very young children must have access to reasonably detailed phonological information in order to create and distinguish among such phonologically similar lexical entries.

The acquisition of language in the preschool years by multiple-birth children is often reported to be atypical. Some researchers have claimed that they use ‘twin language’, i.e. an autonomous language specific to a multiple-birth set. This claim was investigated by describing and comparing the phonological characteristics of the speech of 19 sets of two- to four-year-old multiple-birth children, and by measuring multiple-birth children's understanding of their twins' or triplets' context-free speech. The results indicated that multiple-birth children are prone to phonological disorder and consequently their speech is often unintelligible. Siblings' phonologies evidenced some similarities, although they were not identical. Multiple-birth children were better able to understand their siblings' mispronunciations than were other children of the same age, although that understanding was dependent on how closely the error form resembled the adult form. While these three factors conspire to give the impression that ‘twin language’ is common, none of these findings provided support for the claim that multiple-birth children use an autonomous language.

The phonemic inventories of 30 children (aged 3;4–5;7) with phonological delays were examined in terms of featural distinctions in order to address universal vs. individual accounts of acquisition. Phonetic inventories of these same children were also identified for comparison purposes. Across children, four hierarchical and implicationally related types of phonemic inventory were identified. The typology uniquely captured common distinctions maintained by all children, and at the same time, allowed for individual differences in the specific phonemic composition of each system. These cross-sectional results have theoretical implications for the longitudinal course of phonemic acquisition. In particular, children appear to have a number of linguistic choices that relate to the course, the specifics, and the mechanism of change in acquisition.

A blind child and her sighted twin sister were recorded at home once a month from 2;5 to 3;5. Their imitations, repetitions and routines (IRR speech) were analysed with respect to three dimensions (type, faithfulness to the model, and function). MLU for IRR speech was greater than that for productive utterances, which supports the idea that these productions facilitate language development. The blind girl used routines and modified imitations and repetitions more frequently than her sighted sister, and her use of modified repetitions increased during the study. She also seemed to use IRR speech for a longer period in her development. Both children frequently used these forms as a means of maintaining conversation and to fulfil their intentions. By introducing modifications (varying some element, or expanding the model) both children, especially the blind girl, analysed language and learned to use these utterances in the appropriate circumstances with a clear pragmatic function. Such behaviour is typical of holistic language processing and acquisition.

Several recent studies have demonstrated strong relationships between lexical acquisition and subsequent developments within the domain of morphosyntax. A connectionist model of the acquisition of a morphological System analogous to that of the English past tense (Plunkett & Marchman, 1993) suggests that growth in vocabulary size may relate to the onset of over-regularization errors. However, this model suggests that the relationships between vocabulary size and morphosyntactic development are non-linear. Incremental increases in the number of verbs to be learned result in qualitative shifts in the treatment of both previously learned and novel forms, but only after the size of the lexicon exceeds a particular level (i.e. reaches a ‘critical mass’). In this paper we present parental report data from an extensive study of English-speaking children aged 1;4 to 2;6 using the MacArthur Communicative Development Inventory: Toddler form (N = 1130). These data corroborate several findings from previous studies, including the early usage of unmarked verb stems and the correct production of irregular past tense forms. Further, we demonstrate support for the ‘critical mass’ view of the onset of over-regularization errors, focusing on continuity among lexical and morphological developments. In our view, these data suggest that these linguistic milestones may be paced by similar, if not identical mechanisms.

Word learning by young children is viewed as a problem of deriving meaning from the use of forms in discourse contexts. Uses of causal and temporal terms in private speech by a child studied longitudinally from 1;9 to 3;0 are analysed from this perspective. Evidence is presented that words are first constrained to uses in specific discourse contexts, and later used more flexibly and with greater control over the semantics of the terms. Derivation of meaning from discourse is described as a dialectical process, and as such it is claimed to be more consistent with the full range of observational data, and with theories of word learning applicable to older children and adults, than other current theories of lexical acquisition in early childhood.

Recent research suggests that young children assume, as a default, that a word applied ostensively to an unfamiliar object (e.g. a man riding in a car) is a basic-level count noun (e.g. person) rather than a situation-restricted count noun (e.g. passenger). In this experiment, 14 mothers and their children (mean age 3;7; range 3;0 to 4;6) participated in a story-book reading session in which the mothers taught their children both a basic-level count noun and a situation-restricted count noun for a series of object drawings. An analysis of mothers' spontaneous teaching strategies revealed that, for a given object, they typically taught a basic-level count noun before a situation-restricted count noun. Furthermore, they tended to teach basic-level count nouns exclusively through estensive definition, especially when the objects were unfamiliar. Moreover, they were (1) more likely to use estensive definitions and (2) less likely to provide additional information concerning the application of the word, when teaching basic-level count nouns than when teaching situation-restricted count nouns. The results provide new evidence that word teachers, like young word learners, assume that a basic-level count noun is the psychologically-privileged candidate for an ostensive definition involving an (unfamiliar) object.

An experimental study examined the comprehension of sentences containing concessive connectives, considered from an argumentative-conclusive point of view, in eight- and ten-year-old French children (24 subjects in each age group). Two tasks were used; subjects had to choose between (1) opposite preceding contexts of sentences (context choice task), and (2) conclusions that could be drawn from the same sentences (conclusion choice task). Results indicated a clear developmental increase in performance. For all concessive items, except mais (‘but’), results of the context choice task were better than results of the conclusion choice task. Older children gave significantly better responses for mais in the conclusion choice task. The ambivalence of the connective mais is discussed; comparisons with production and crosslinguistic data are suggested.

In this paper we investigate the development of interactive communication skills in 170 children aged seven to thirteen. Using a communication task that allows extended dialogues between pairs of young speakers, we are able to assess both the overall communicative success achieved by any pair; and from an analysis of the dialogues we can identify several interactive strategies that characterize older and more successful communicators. Successful communication involves the active involvement of both participants: asking and answering questions, volunteering information and responding sensitively to contributions from their partners. In contrast to the process of language acquisition, the development of these interactive skills takes place over an extended time period and subjects vary greatly in their ability to communicate effectively. For many analyses differences in the communicative success achieved by subjects differ more within than between age groups. For example, a substantial minority of our oldest subjects communicate no better than children six years younger

Twenty-four infants were followed longitudinally from 1;2 to 1;10. Parents maintained diaries of the child's spoken words and at two-month intervals completed a representative checklist of words produced. There was good agreement across the two instruments and robust month-to-month correlations for both. However, the overall pattern of results suggests that the diary method is more effective during the early emergence of language and the representative checklist method is more effective late in the second year when vocabulary size becomes relatively large. Accurate longitudinal assessment of vocabulary may require a combination of the diary and checklist approaches.

This paper constitutes a response to Lust & Mazuka's (1989) defence of the Principal Branching Direction parameter and their critique of O'Grady, Suzuki-Wei & Cho's (1986) experiment, which purported to show that even children learning left-branching languages exhibit a preference for forward patterns of anaphora. Results of new experimental work with children learning Japanese are reported and shown to support the claim that there is a universal preference for forward patterns of anaphora in the early stages of language acquisition.