A distributed attractor network is trained on an abstract version of the task of deriving the meanings of written words. When processing a word, the network starts from the final activity pattern of the previous word. Two words are semantically related if they overlap in their semantic features, whereas they are associatively related if one word follows the other frequently during training. After training, the network exhibits two empirical effects that have posed problems for distributed network theories: much stronger associative priming than semantic priming, and significant associative priming across an intervening unrelated item. It also reproduces the empirical findings of greater priming for low-frequency targets, degraded targets, and high-dominance category exemplars.

We present a new distributed connectionist model of the perception of spoken words. The model employs a representatio n of speech that combines lexical information with abstract phonological information, with lexical access modelled as a direct mapping onto this single distributed representatio n. We �rst examine the integration of partial cues to phonological identity, showing that the model provides a sound basis for simulating phonetic and lexical decision data from Marslen-Wilson and Warren (1994). We then investigate the time course of lexical access, and argue that the process of competition between word candidates during lexical access can be interpreted in terms of interference between distributed lexical representatio ns. The relation between our model and other models of spoken word recognition is discussed.

In the novel noun generalization task, 2 1/2-year-old children display generalized expectations about how solid and nonsolid things are named, extending names for never-before-encountered solids by shape and for never-before-encountered nonsolids by material.This distinction between solids and nonsolids has been interpreted in terms of an ontological distinction between objects and substances.Nine simulations and behavioral experiments tested the hypothesis that these expectations arise from the correlations characterizing early learned noun categories.In the simulation studies, connectionist networks were trained on noun vocabularies modeled after those of children.These networks formed generalized expectations about solids and nonsolids that match children’s performances in the novel noun generalization task in the very different languages of English and Japanese.The simulations also generate new predictions supported by new experiments with children.Implications are discussed in terms of children’s development of distinctions between kinds of categories and in terms of the nature of this knowledge. Concepts are hypothetical constructs, theoretical devices hypothesized to explain data, what people do, and what people say. The question of whether a particular theory can explain children’s concepts is therefore semantically strange because strictly speaking this question asks about an explanation of an explanation.We begin with this reminder because the goal of the research reported here is to understand the role of associative processes in children’s systematic attention to the shape of solid things and to the material of nonsolid things in the task of forming new lexical categories. These attentional biases have been interpreted in terms of children’s concepts about the ontological kinds of object and substance

In everyday tasks, selecting actions in the proper sequence requires a continuously updated representation of temporal context. Many existing models address this problem by positing a hierarchy of processing units, mirroring the roughly hierarchical structure of naturalistic tasks themselves. Such an approach has led to a number of difficulties, including a reliance on overly rigid sequencing mechanisms, an inability to account for context sensitivity in behavior, and a failure to address learning. We consider here an alternative framework, according to which the representation of temporal context is facilitated by recurrent connections within a network mapping from environmental inputs to actions. Applying this approach to a specific, and in many ways prototypical, everyday task (coffee-making), we examine its ability to account for several central characteristics of normal and impaired human performance. The model we consider learns to deal flexibly with a complex set of sequencing constraints, encoding contextual information at multiple time-scales within a single, distributed internal representation. Mildly degrading this context representation leads

the properties of distributed network models, and support this account by demonstrating that an implemented simulation closely approximates the empirical findings despite the absence of expectancy-based processes and postlexical semantic matching. The results suggest that distributed network models can provide a viable single-mechanism account of lexical processing. Introduction It is well-established that people are faster and more accurate to read a word (e.g., BUTTER) when it is preceded by a related word (e.g., BREAD) compared with when it is preceded by an unrelated word (e.g., DOCTOR; The research was supported by an NIMH FIRST award (MH55628) to the first author and by NIMH Training Grant 5T32MH19102 and NICHD Grant 80258. The computational simulation was run using customized software written within the Xerion simulator (version 3.1) developed by Drew van Camp, Tony Plate, and Geoff Hinton at the Univers

Using an auditory semantic priming paradigm, Moss, Ostrin, Tyler and Marslen-Wilson (1995, Experiment 2) demonstrated facilitation for category coordinates and functionally-related stimuli both with and without the additive effect of normative association strength. In this paper we replicate these results computationally using a corpus-derived Contextual Similarity measure. In Experiment 1 we consider the adequacy of the Contextual Similarity measure in accounting for Moss et al.'s results, and discuss how functional and categorical semantic relations are represented in corpus-based approaches to lexical semantics. We also offer an explanation for how the Contextual Similarity measure succeeds in replicating the additive effect of association strength on semantic priming without postulating a qualitatively different mechanism for associative priming. We then investigate why previous corpus-based approaches (Lund, Burgess & Atchley, 1995) have failed to produce similar results. We argue...

A central concern of psycholinguistic research is explaining the relative ease or difficulty involved in processing words. In this thesis, we explore the connection between lexical processing effort and measurable properties of the linguistic environment. Distributional information (information about a word's contexts of use) is easily extracted from large language corpora in the form of co-occurrence statistics. We claim that such simple distributional statistics can form the basis of a parsimonious model of lexical processing effort.

Several cases of category-specific semantic deficits for living things have now been reported, most of whom have a greater difficulty with knowledge of visual properties of living things than with non-visual properties. This has motivated two different kinds of account of category-specific deficits for living things. One account is that theimpairment forliving things arises as a result of general damage to visual properties, and so it predicts that impairment for visual properties of living things will be accompanied by some degree of impairment for visual properties of other categories. The second account posits explicit category-based organisation of semantic memory, such that visual properties of living and non-living things are stored separately, and so can be independently impaired. We investigated the semantic impairment of a patient SE, who had received a diagnosis of herpes simplex encephalitis several years previously, resulting in a mild category-specific semantic deficit for living things. Both off-line and on-line tasks revealed that SE had a highly selective impairment for the visual properties of living things, with no sign of problems for the visual properties of non-living things. This raises a challenge for models in which category-specific deficits emerge as a result of a general deficit determined by some other factor. We suggest one way in which these models may account for the data by encoding inter-relations between the function and form of objects in semantic memory. Requests for reprints should be addressed to Dr Helen Moss, Department of Psychology,

Four lexical decision experiments were conducted to examine under which conditions automatic semantic priming effects can be obtained. Experiments 1 and 2 analyzed associative/semantic effects at several very short stimulus-onset asynchronies (SOAs), whereas Experiments 3-4 used a single-presentation paradigm at two response-stimulus intervals (RSIs). Experiment 1 tested associatively related pairs from three semantic categories (synonyms, antonyms, and category coordinates). The results showed reliable associative priming effects at all SOAs. In addition, the correlation between associative strength and magnitude of priming was significant only at the shortest SOA (66 ms). When prime-target pairs were semantically but not associatively related (Experiment 2), reliable priming effects were obtained at SOAs of 83 ms and longer. Using the single-presentation paradigm with a short RSI (200 ms, Experiment 3), the priming effect was equal in size for associative+semantic and for semantic-on...

It has been claimedthat concepts in differentsemantic domains vary in the extent to which their meaning is comprised of different kinds of semantic information. Discussion has mainly focused around two kinds of concepts—living things and man-made objects—arguing that functional information is central to the meaning of artefacts whereas perceptual information is more important for the meaning of living things. This distinction has been important in accounting for patterns of semantic impairments following brain injury (Warrington & Shallice, 1984). We suggest that functional information may be especially salient in the semantic representations of both living and nonliving things. Our evidence for this claim comes from priming studies with normal subjects, and data from brain-damaged patients that supports the claim that functional information is relatively spared following brain damage. We explore further implications of the role of functional properties in semantic representations,considering distinctions between different types of functional information in the representationof living things. We focus on the developmental claim that biological functional information,