Infants' sensitivity to word units in fluent speech was examined by inserting 1 sec pauses either at boundaries between successive words (Coincident versions) or between syllables within words (Noncoincident versions). In Experiment 1, 24 11-month-olds listened significantly longer to the Coincident versions. In Experiment 2, 24 four-and-a-half-and 24 nine-month-olds did not exhibit the preference for the Coincident versions that the 11-month-olds showed. When the stimuli were low-pass filtered in Experiment 3, 24 11-month-olds showed no preference for the Coincident versions, suggesting they rely on more than prosodic cues. New stimulus materials in Experiment 4 indicated that responses by 24 11-month-olds to the Coincident and Noncoindent versions did not depend solely on prior familiarity with the targets. Two groups of 30 11-month-olds tested in Experiment 5 were as sensitive to boundaries for Strong/Weak words as for Weak/Strong words. Taken together, the results suggest that, by 11 months, infants are sensitive to word boundaries in fluent speech, and that this sensitivity depends on more than just prosodic information or prior knowledge of the words.

The production of target consonant clusters at early stages of acquisition is analysed from a phonological representational perspective. The data stem from five normal monolingual German and four normal monolingual Spanish children at ages from 0;9 to 2;1, observed in naturalistic settings. At the beginning stages, target clusters are reduced to a single consonantal position, due to lack of branching of the syllabic constituents. This finding coincides with other results in the literature, which have in general been explained by means of universal principles. Nevertheless, there is an essential difference between the German and the Spanish data: German children tend to prefer the first consonant and Spanish children the second one. This difference can only be explained in terms of parameterization of syllabification, which in German takes place from left to right and in Spanish from right to left. At later stages, when clusters begin to be produced with two consonantal positions, they offer evidence for the beginning of branching of syllabic constituents, due to parameterization, and for the chronological order of the setting of the subsyllabic parameters. Our data offer evidence in favour of the following acquisitional hierarchy: CV > CVC > CVCC > CCVCC.

Several competing proposals for the (under)specification of phonological representations are evaluated against the facts of phonemic acquisition. Longitudinal evidence relating to the emergence of a voice contrast in the well-documented study of Amahl (from age 2;2 to 3;11) is reconsidered. Neither contrastive specification nor context-free radical underspecification is capable of accounting for the facts. The problem is in the characterization of the change in the status of a feature from being noncontrastive and conditioned by context at one stage to being contrastive with phonetic effects that diffuse gradually through the lexicon. Both frameworks must treat as accidental the persistence of the early substitution pattern and require the postulation of wholesale changes in underlying representations, where these changes do not accord well with the observed phonetic changes or with the facts available to the learner. Context-sensitive radical underspecification provides a plausible account of each stage and the transition between stages with minimal grammar change.

Laryngeal-supralaryngeal cyclicity has been put forth as an acquisition principle that describes expansion of the phonemic inventory as a bivalent cycle with alternations between the elaboration of laryngeal and supralaryngeal properties of the system (Gierut, 1994). This study evaluated experimentally the claims of the principle by manipulating the domain of the cycle and the phase relationship of the cycle as independent variables, and by monitoring longitudinally the order of emergent phonemic distinctions in the sound systems of seven children with phonological delays (aged 3;4 to 5;8) as the dependent variable. Three general findings emerged: (1) the course of acquisition for all children was consistent with cyclicity; (2) there were no differences in phonemic acquisition between children taught a laryngeal as opposed to a supralaryngeal distinction; and (3) children taught out-of-phase with the cycle evidenced greater phonemic expansion than those taught inphase with the cycle.

In order to separate competence and performance factors in acquisition of knowledge of syntactic subjects, we audiotaped and analysed the spontaneous speech of 20 Portuguese-speaking two-year-olds in natural conversation with Portuguese-speaking adults. We separated the children into three groups based on Mean Length of Utterance in Words: 1·5–1·99; 2·0–2·99; 3·0–4·99. Our cross-sectional data demonstrated that Portuguese-speaking children increased their use of subjects from 28% in the lowest-MLUW group to 57% in the highest-MLUW group. The children in the highest-MLUW group almost perfectly matched the adult speakers in the study on every measure. The increase in children's use of subjects was primarily due to an increase in the use of pronominal subjects. A comparison between Portuguese- and English-speaking children suggests that adult competence about the status of subjects is present at the onset of combinatorial speech, as shown by differential production of subjects. Each group also experiences performance limitations, as shown by the increase in subject use as development proceeds.

Passive structures are typically assumed to be one of the later acquired constructions in child language. English-speaking children have been shown to produce and comprehend their first simple passive structures productively by about age four and to master more complex structures by about age nine. Recent crosslinguistic data have shown that this pattern may not hold across languages of varying structures. This paper presents data from four Inuit children aged 2;0 to 3;6 that shows relatively early acquisition of both simple and complex forms of the passive. Within this age range children are productively producing truncated, full, action and experiential passives. Some possible reasons for this precociousness are explored including adult input and language structure.

Previous studies have demonstrated that children aged 2;0 can learn new words in a variety of non-ostensive contexts. The current two studies were aimed at seeing if this was also true of children just beginning to learn words at 1;6. In the first study an adult interacted with 48 children. She used a nonce word to announce her intention to find an object (‘Let's find the gazzer’), picked up and rejected an object with obvious disappointment, and then gleefully found the target object (using no language). Children learned the new word as well in this condition as in a condition in which the adult found the object immediately. In the second study the adult first played several rounds of a finding game with each of 60 children, in which it was first established that one of several novel objects was always in a very distinctive hiding place (a toy barn). The adult then used a nonce word to announce her intention to find an object (‘Let's find the toma’) and then proceeded to the barn. In the key condition the barn was mysteriously ‘locked’; the child thus never saw the target object after the nonce word was introduced. Children learned the new word as well in this condition as in a condition in which the adult found the object immediately. The results of these two studies suggest that from very early in language acquisition children learn words not through passive, associative pro- cesses, but rather through active attempts to understand adult behaviour in a variety of action and discourse contexts.

It is commonly believed that colour terms are acquired unusually slowly. However, several recent studies suggest that preschoolers have more colour term knowledge than previously believed. In two studies, we investigated the colour term knowledge of 49 two-year-olds and the influence of schooling on their colour term acquisition. Both studies revealed considerable knowledge in such young children. A longitudinal comparison of children with and without school experience in Study 2 indicated that the higher performance of the school group could not be accounted for by differences predating that group's experience. In reviewing the evidence from the children's performance, we conclude that contemporary children's early and frequent experience with colour terms may account for this cohort's improved performance over previous generations of children.

Global tendencies for the relative absence of covocalization (simul- taneous talk) have been identified in both conversations between adult partners and conversations between mothers and their infants; in each case, the alternating mode in which one partner speaks at a time is predominant. The present investigation examined the timing of the partners' talk in mother-infant engagements over infant age to determine whether: (a) variations occur in the incidence of the alternating mode; and (b) variations occur in the extent to which the alternating mode predominates. Conversations involving a total of 48 mothers and their infants aged from 0;3 to 2;0 were investigated at each of eight infant ages (0;3, 0;6, 0;9, 1;0, 153, 1;6, 1;9 and 2;0). The results indicated that, within a global tendency for the relative absence of covocalization, there was: (a) a curvilinear tendency for the incidence of covocalization to decrease over the first 18 months, and then to increase; and (b) a linear tendency for the extent to which the alternating mode predominates to increase over age. These changes are interpreted as reflecting the facilitative effects of covocalization in the case of young preverbal infants, and the need for the alternating, turn-taking pattern to predominate as mutual comprehension becomes possible in conversations between mothers and their older infants.

Research on early mother-child interaction has documented the crucial role affect plays in the content and modulation of early interactions. For hearing mothers, voice quality is considered to be the single most informative channel for affective expression. For deaf caregivers who use American Sign Language (ASL), the vocal channel is unavailable, and facial expression is critically important. Not only do facial behaviours signal affective and communicative information, but specific facial behaviours also function as obligatory grammatical markers. This multifunctionality of facial expression presents a dilemma for deaf parents signing to their toddlers as these two systems potentially compete for expression on the face. This study addresses how affective facial expression interacts with the linguistic forms in ASL motherese. To address this issue, we present data from both cross-sectional and longitudinal videotaped interaction from a total of 15 deaf mothers signing with their deaf toddlers (ages 0;9–2;8). Using Ekman & Friesen's Facial Action Coding System (FACS) (1978) we analysed child-directed maternal wh- questions. Because they are frequent in early discourse, AND they require furrowed brows which also signal anger and puzzlement, wh- questions represent an ideal context to address the potential conflict of grammatical and affective facial expression in ASL motherese. Our studies indicate a shift from affect to grammar at about the child's second birthday. These findings shed new insight on the nature and possible role of input on the language acquisition process.

We reply to Mervis & Bertrand's report of three children (aged 1; 6–1; 8 at the start of the study) who evidenced a ‘late’ vocabulary spurt. Differences in assessing productive vocabulary, and the questionable inference that size of the lexicon is a reliable indicator of the vocabulary spurt, make it inappropriate to compare these children to previous studies that directly measure change in rate of word learning. Further work using continuous records of lexical development and controls for repeated cognitive assessments is needed to test hypotheses about the spurt and related cognitive and linguistic achievements.