Kim Plunkett

U-shaped learning and frequency effects in a multi-layered perception: Implications for child language acquisition

A three-layer back-propagation network is used to implement a pattern association task in which four types of mapping are learned. These mappings, which are considered analogous to those which characterize the relationship between the stem and past tense forms of English verbs, include arbitrary mappings, identity mappings, vowel changes, and additions of a suffix. The degree of correspondence between parallel distributed processing (PDP) models which learn mappings of this sort (e.g., Rumelhart & McClelland, 1986, 1987) and childrens acquisition of inflectional morphology has recently been at issue in discussions of the applicability of PDP models to the study of human cognition and language (Pinker & Mehler, 1989; Bever, in press). In this paper, we explore the capacity of a network to learn these types of mappings, focusing on three major issues. First, we compare the performance of a single-layered perceptron similar to the one used by Rumelhart and McClelland with a multi-layered perceptron. The results suggest that it is unlikely that a single-layered perceptron is capable of finding an adequate solution to the problem of mapping stems and past tense forms in input configurations that are sufficiently analogous to English. Second, we explore the input conditions which determine learning in these networks. Several factors that characterize linguistic input are investigated: (a) the nature of the mapping performed by the network (arbitrary, suffixation, identity, and vowel change); (b) the competition effects that arise when the task demands simultaneous learning of distinct mapping types; (c) the role of the type and token frequency of verb stems; and (d) the influence of phonological subregularities in the irregular verbs. Each of these factors is shown to have selective consequences on both successful and erroneous performance in the network. Third, we outline several types of systems which could result in U-shaped acquisition, and discuss the ways in which learning in multi-layered networks can be seen to capture several characteristics of U-shaped learning in children. In general, these models provide information about the role of input in determining the kinds of errors that a network will produce, including the conditions under which rule-like behavior and U-shaped learning will and will not emerge. The results from all simulations are discussed in light of behavioral data on childrens acquisition of the past tense and the validity of drawing conclusions about the acquisition of language from models of this sort.

From rote learning to system building: acquiring verb morphology in children and connectionist nets

The traditional account of the acquisition of English verb morphology supposes that a dual architecture underlies the transition from early rote-learning processes (in which past tense forms of verbs are correctly produced) to the systematic treatment of verbs (in which irregular verbs are prone to error). A connectionist account supposes that this transition can occur in a single mechanism (in the form of a neural network) driven by gradual quantitative changes in the size of the training set to which the network is exposed. In this paper, a series of simulations is reported in which a multi-layered perceptron learns to map verb stems to past tense forms analogous to the mappings found in the English past tense system. By expanding the training set in a gradual, incremental fashion and evaluating network performance on both trained and novel verbs at successive points in learning, it is demonstrated that the network undergoes reorganizations that result in a shift from a mode of rote learning to a systematic treatment of verbs. Furthermore, we show that this reorganizational transition is dependent upon the number of regular and irregular verbs in the training set and is sensitive to the phonological sub-regularities characterizing the irregular verbs. The pattern of errors observed is compared to that of children acquiring the English past tense, as well as childrens performance on experimental studies with nonsense verbs. It is concluded that a connectionist approach offers a viable alternative account of the acquisition of English verb morphology, given the current state of empirical evidence relating to processes of acquisition in young children.

Infant vocabulary development assessed with a British communicative development inventory

Communicative Development Inventories (CDIs) were collected from 669 British children aged between 1;0 and 2;1. Comprehension and production scores in each age group are calculated. This provides norming data for the British infant population. The influence of socioeconomic group on vocabulary scores is considered and shown not to have a significant effect. The data from British infants is compared to data from American infants (Fenson, Dale, Reznick, Bates, Thal & Pethick, 1994). It is found that British infants have lower scores on both comprehension and production than American infants of the same age.

Symbol Grounding or the Emergence of Symbols? Vocabulary Growth in Children and a Connectionist Net

Abstract The Symbolic Grounding Problem is viewed as a by-product of the classical cognitivist approach to studying the mind. In contrast, an epigenetic interpretation of connectionist approaches to studying the mind is shown to offer an account of symbolic skills as an emergent, developmental phenomenon. We describe a connectionist model of concept formation and vocabulary growth that auto-associates image representations and their associated labels. The image representations consist of clusters of random dot figures, generated by distorting prototypes. Any given label is associated with a cluster of random dot figures. The network model is tested on its ability to reproduce image representations given input labels alone (comprehension) and to identify labels given input images alone (production). The model implements several well-documented findings in the literature on early semantic development; the occurrence of over- and under-extension errors; a vocabulary spurt; a comprehension/production asymmetr...

Phonological specificity in early words

Young children often fail to distinguish words differing by a single phoneme. It has been suggested that the phonological representations associated with early words are under-specified, with detail being added either as required to distinguish more words as a child’s vocabulary increases, or as a function of the amount of exposure a child has had to a particular word. We report on an inter-modal preferential looking task in which 48, 18- and 24-month-olds heard either accurate pronunciations of picture names or slight (1- or 2-feature) mispronunciations. Each child was tested on four words that they learned at a very young age, and four words they had learned only recently relative to the test session. Across all children, 30 different words were tested. Analysis of children’s looking behavior found differences between their responses to accurate and inaccurate pronunciations. These results show that toddlers’ representations for words have enough detail to detect even slight mispronunciations. This ability was unaffected by their age, the size of their receptive vocabulary, the neighborhood density of individual test words, or the recency with which words had been learned. The results suggest that the specificity of lexical representations may be unrelated to any of these factors, at least within the range of ages and word familiarity examined in this study.

Labels can override perceptual categories in early infancy

An extensive body of research claims that labels facilitate categorisation, highlight the commonalities between objects and act as invitations to form categories for young infants before their first birthday. While this may indeed be a reasonable claim, we argue that it is not justified by the experiments described in the research. We report on a series of experiments that demonstrate that labels can play a causal role in category formation during infancy. Ten-month-old infants were taught to group computer-displayed, novel cartoon drawings into two categories under tightly controlled experimental conditions. Infants were given the opportunity to learn the two categories under four conditions: Without any labels, with two labels that correlated with category membership, with two labels assigned randomly to objects, and with one label assigned to all objects. Category formation was assessed identically in all conditions using a novelty preference procedure conducted in the absence of any labels. The labelling condition had a decisive impact on the way infants formed categories: When two labels correlated with the visual category information, infants learned two categories, just as if there had been no labels presented. However, uncorrelated labels completely disrupted the formation of any categories. Finally, consistent use of a single label across objects led infants to learn one broad category that included all the objects. These findings demonstrate that even before infants start to produce their first words, the labels they hear can override the manner in which they categorise objects.

The Shape of Words in the Brain.

The principle of arbitrariness in language assumes that there is no intrinsic relationship between linguistic signs and their referents. However, a growing body of sound-symbolism research suggests the existence of some naturally-biased mappings between phonological properties of labels and perceptual properties of their referents (Maurer, Pathman, & Mondloch, 2006). We present new behavioural and neurophysiological evidence for the psychological reality of sound-symbolism. In a categorisation task that captures the processes involved in natural language interpretation, participants were faster to identify novel objects when label-object mappings were sound-symbolic than when they were not. Moreover, early negative EEG-waveforms indicated a sensitivity to sound-symbolic label-object associations (within 200ms of object presentation), highlighting the non-arbitrary relation between the objects and the labels used to name them. This sensitivity to sound-symbolic label-object associations may reflect a more general process of auditory-visual feature integration where properties of auditory stimuli facilitate a mapping to specific visual features.

A neurocomputational account of taxonomic responding and fast mapping in early word learning

We present a neurocomputational model with self-organizing maps that accounts for the emergence of taxonomic responding and fast mapping in early word learning, as well as a rapid increase in the rate of acquisition of words observed in late infancy. The quality and efficiency of generalization of word-object associations is directly related to the quality of prelexical, categorical representations in the model. We show how synaptogenesis supports coherent generalization of word-object associations and show that later synaptic pruning minimizes metabolic costs without being detrimental to word learning. The role played by joint-attentional activities is identified in the model, both at the level of selecting efficient cross-modal synapses and at the behavioral level, by accelerating and refining overall vocabulary acquisition. The model can account for the qualitative shift in the way infants use words, from an associative to a referential-like use, for the pattern of overextension errors in production and comprehension observed during early childhood and typicality effects observed in lexical development. Interesting by-products of the model include a potential explanation of the shift from prototype to exemplar-based effects reported for adult category formation, an account of mispronunciation effects in early lexical development, and extendability to include accounts of individual differences in lexical development and specific disorders such as Williams syndrome. The model demonstrates how an established constraint on lexical learning, which has often been regarded as domain-specific, can emerge from domain-general learning principles that are simultaneously biologically, psychologically, and socially plausible.

Lexical segmentation and vocabulary growth in early language acquisition.

The identification of appropriate lexical segmentations of the speech signal constitutes a problem for the language learner and the child language researcher alike. Articulatory precision and fluency criteria for identifying formulaic expressions, sub-lexical forms and target lexemes in linguistic productions are defined and applied to the analysis of two Danish childrens language development between the ages of 1;0 and 2;0. The results of this analysis are compared to the results of applying standard distributional and frequency criteria in the tabulation of mean length of utterance and vocabulary profiles for both standard and non-standard lexical segmentations. It is argued that although the two methods yield converging profiles of development during the latter part of the period studied, articulatory precision and fluency criteria offer a more powerful tool for identifying alternative segmentation strategies in early language acquisition. Profiles of vocabulary development for these two children suggest that the solution to the segmentation problem may be an important trigger for their vocabulary spurts.

A computational and neuropsychological account of object‐oriented behaviours in infancy

Infants under 7 months of age fail to reach behind an occluding screen to retrieve a desired toy even though they possess sufficient motor skills to do so. However, even by 3.5 months of age they show surprise if the solidity of the hidden toy is violated, suggesting that they know that the hidden toy still exists. We describe a connectionist model that learns to predict the position of objects and to initiate a response towards these objects. The model embodies the dual-route principle of object information processing characteristic of the cortex. One route develops a spatially invariant surface feature representation of the object whereas the other route develops a feature blind spatial–temporal representation of the object. The model provides an account of the developmental lag between infants’ knowledge of hidden objects and their ability to demonstrate that knowledge in an active retrieval task, in terms of the need to integrate information across multiple object representations using (associative) connectionist learning algorithms. Finally, the model predicts the presence of an early dissociation between infants’ ability to use surface features (e.g. colour) and spatial–temporal features (e.g. position) when reasoning about hidden objects. Evidence supporting this prediction has now been reported.