Two-year-olds will name artifacts by their functions.

Do young children take functional information into account in naming artifacts? In three studies of lexical categorization, 112 children 2 years of age learned new names for novel artifacts with novel functions and then extended the names to new objects. The objects were designed to have functions that were causally related in simple and compelling ways to perceptible aspects of their physical structure. Despite only minimal opportunity to familiarize themselves with the objects, children generalized the names in accordance with the objects' functions. This result obtained even when children had to discover the functions of the named objects on their own (Experiment 2) and when all the test objects had some discernible function (Experiment 3). Two-year-olds name by function when they can make sense of the relation between the appearances and the functions of artifacts.

Young children's use of functional information to categorize artifacts: three factors that matter.

Three experiments addressed factors that might influence whether or not young children take into account function, as opposed to overall appearance or shape, when they extend the names of novel artifacts. Experiment 1 showed that 4-year-olds more often extend a name on the basis of a demonstrated function when that function provides a plausible causal account of perceptible object structure. Experiment 2 showed that they more often extend a name by function when they respond slowly, and hence thoughtfully. Experiment 3 demonstrated that they are more likely to take function into account when they extend names than when they judge similarities. Comparisons of lexical and non-lexical conditions in younger children failed to show any differences. Overall, the findings suggest that by 4 years of age, children may learn names as labels for novel artifact kinds rather than perceptual classes, and that the processes by which they categorize may be mindful and reflective, as in adults.

Infants' sensitivity to word boundaries in fluent speech.

Infants' sensitivity to word units in fluent speech was examined by inserting I 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 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.

When experimental findings conflict with everyday observations: reflections on children's category learning.

Attempts to reconcile the ease with which young children naturally learn everyday categories with their frequent difficulty in acquiring artificial categories in the laboratory have taken different forms. Kemler Nelson suggested that one reason for the discrepancy may be that many everyday object categories have a family-resemblance structure that can be learned by means of a holistic mode of processing. While Ward et al. have recently questioned this account of why children learn family-resemblance categories easily, conclusions based on their laboratory data fail to provide a good explanation of the real-world case. Accordingly, it is suggested that the laboratory family-resemblance task used by these previous investigators may be unrepresentative and may fail to mimic crucial aspects of the everyday category-learning context. It is also suggested that aspects of Ward et al.'s methodology may lead them to underestimate holistic (or nonselective) processing.

How the prosodic cues in motherese might assist language learning.

The function of motherese has become a pivotal issue in the language-learning literature. The current research takes the approach of asking whether the prosodic characteristics that are distinctive to motherese could play a special role in facilitating the acquisition of syntax. Hirsh-Pasek, Kemler Nelson, Jusczyk, Cassidy, Druss & Kennedy (1987) showed that infants aged 0;7-0;10 are sensitive to prosodic cues that would help them segment the speech stream into perceptual units that correspond to clauses. The present study shows that infants' sensitivity to segment-marking cues in ongoing speech holds for motherese but not for adult-directed speech. The finding is that, for motherese only, infants orient longer to speech that has been interrupted at clausal boundaries than to matched speech that has been interrupted at within-clause locations. This selective preference indicates that the prosodic qualities of motherese provide infants with cues to units of speech that correspond to grammatical units of language-a potentially fundamental contribution of motherese to the learning of syntax.

The distinction between integral and separable dimensions: evidence for the integrality of pitch and loudness.

Six experiments are reported that investigated the reality and generality of dimensional integrality (Garner, 1974a) by evaluating whether the auditory dimensions of pitch and loudness are psychologically privileged and whether they combine in an integral fashion. In Experiment 1 we psychophysically scaled the dimensions to ensure that, within the stimulus range used, the perceived value on each dimension would remain constant in the face of variation on the other dimension. In Experiments 2 through 4 we assessed performance by using the converging operations by which Garner defined integrality and separability. Experiment 2 showed that in speeded classification, pitch and loudness lead to facilitation with redundant variation and interference with orthogonal variation. Experiment 3 showed that unspeeded classifications are guided predominantly by overall similarity. Experiment 4 established that the better-fitting metric by which multidimensional similarity is appreciated is Euclidean rather than city block. These results suggest that the dimensions of pitch and loudness combine in an integral fashion. In Experiments 5 and 6 we investigated whether the dimensions of pitch and loudness have a privileged status by assessing the impact of rotating the dimensional axes on performance in a speeded sorting task. Experiment 5 looked at six alternative dimensional orientations to pitch and loudness. If anything, rotating the dimensional axes increased the amount of interference in filtering. In Experiment 6 we assessed an alternative dimensional description of the stimuli based on the dimensions of volume and brightness. We found greater interference when the stimuli varied along the dimensions of volume and brightness than when they varied along the dimensions of pitch and loudness. The fact that the least interference is observed when the stimuli vary along the dimensions of pitch and loudness suggests that these dimensions are the more psychologically valid ones. These findings indicate that integrality is not a "myth," that is, merely a case of psychophysical mismatch. Instead, dimensions that are psychologically real are sometimes processed in a unitary fashion.