A set of hypotheses is formulated for a connectionist approach to cognitive modeling. These hypotheses are shown to be incompatible with the hypotheses underlying traditional cognitive models. The connectionist models considered are massively parallel numerical computational systems that are a kind of continuous dynamical system. The numerical variables in the system correspond semantically to fine-grained features below the level of the concepts consciously used to describe the task domain. The level of analysis is intermediate between those of symbolic cognitive models (...) and neural models. The explanations of behavior provided are like those traditional in the physical sciences, unlike the explanations provided by symbolic models. (shrink)

Human agents draw a variety of inferences effortlessly, spontaneously, and with remarkable efficiency – as though these inferences were a reflexive response of their cognitive apparatus. Furthermore, these inferences are drawn with reference to a large body of background knowledge. This remarkable human ability seems paradoxical given the complexity of reasoning reported by researchers in artificial intelligence. It also poses a challenge for cognitive science and computational neuroscience: How can a system of simple and slow neuronlike elements represent a large (...) body of systemic knowledge and perform a range of inferences with such speed? We describe a computational model that takes a step toward addressing the cognitive science challenge and resolving the artificial intelligence paradox. We show how a connectionist network can encode millions of facts and rules involving n-ary predicates and variables and perform a class of inferences in a few hundred milliseconds. Efficient reasoning requires the rapid representation and propagation of dynamic bindings. Our model (which we refer to as SHRUTI) achieves this by representing (1) dynamic bindings as the synchronous firing of appropriate nodes, (2) rules as interconnection patterns that direct the propagation of rhythmic activity, and (3) long-term facts as temporal pattern-matching subnetworks. The model is consistent with recent neurophysiological evidence that synchronous activity occurs in the brain and may play a representational role in neural information processing. The model also makes specific psychologically significant predictions about the nature of reflexive reasoning. It identifies constraints on the form of rules that may participate in such reasoning and relates the capacity of the working memory underlying reflexive reasoning to biological parameters such as the lowest frequency at which nodes can sustain synchronous oscillations and the coarseness of synchronization. (shrink)

When cognitive neuropsychologists make inferences about the functional architecture of the normal mind from selective cognitive impairments they generally assume that the effects of brain damage are local, that is, that the nondamaged components of the architecture continue to function as they did before the damage. This assumption follows from the view that the components of the functional architecture are modular, in the sense of being informationally encapsulated. In this target article it is argued that this “locality” assumption is probably (...) not correct in general. Inferences about the functional architecture can nevertheless be made from neuropsychological data with an alternative set of assumptions, according to which human information processing is graded, distributed, and interactive. These claims are supported by three examples of neuropsychological dissociations and a comparison of the inferences obtained from these impairments with and without the locality assumption. The three dissociations are: selective impairments in knowledge of living things, disengagment of visual attention, and overt face recognition. In all three cases, the neuropsychological phenomena lead to more plausible inferences about the normal functional architecture when the locality assumption is abandoned. Also discussed are the relations between the locality assumption in neuropsychology and broader issues, including Fodor's modularity hypothesis and the choice between top-down and bottom-up research approaches. (shrink)

Connectionist models provide a promising alternative to the traditional computational approach that has for several decades dominated cognitive science and artificial intelligence, although the nature of connectionist models and their relation to symbol processing remains controversial. Connectionist models can be characterized by three general computational features: distinct layers of interconnected units, recursive rules for updating the strengths of the connections during learning, and “simple” homogeneous computing elements. Using just these three features one can construct surprisingly elegant and powerful models of (...) memory, perception, motor control, categorization, and reasoning. What makes the connectionist approach unique is not its variety of representational possibilities or its departure from explicit rule-based models, or even its preoccupation with the brain metaphor. Rather, it is that connectionist models can be used to explore systematically the complex interaction between learning and representation, as we try to demonstrate through the analysis of several large networks. (shrink)

Converging data from different disciplines are showing the role of classical conditioning processes in the elaboration of human and animal behavior to be larger than previously supposed. Restricted views of classically conditioned responses as merely secretory, reflexive, or emotional are giving way to a broader conception that includes problem-solving, and other rule-governed behavior thought to be the exclusive province of either operant conditiońing or cognitive psychology. These new views have been accompanied by changes in the way conditioning is conducted and (...) evaluated. Data from a number of seemingly unrelated phenomena such as relapse to drug abuse by postaddicts, the placebo effect, and the immune response appear to involve classical conditioning processes. Classical conditioning, moreover, has been found to occur in simpler and simpler organisms and recently even demonstrated in brain slices and in utero. This target article will integrate the several research areas that have used the classical conditioning process as an explanatory model; it will challenge teleological interpretations of the classically conditioned CR and offer some basic principles for testing conditioning in diverse areas. (shrink)

This paper argues for a noncognitiveist computationalism in the philosophy of mind. It further argues that both humans and computers have intentionality, that is, their mental states are semantical -- they are about things in their worlds.

Mental imagery (varieties of which are sometimes colloquially refered to as “visualizing,” “seeing in the mind's eye,” “hearing in the head,” “imagining the feel of,” etc.) is quasi-perceptual experience; it resembles perceptual experience, but occurs in the absence of the appropriate external stimuli. It is also generally understood to bear intentionality (i.e., mental images are always images of something or other), and thereby to function as a form of mental representation. Traditionally, visual mental imagery, the most discussed variety, was thought (...) to be caused by the presence of picturelike representations (mental images) in the mind, soul, or brain, but this is no longer universally accepted. (shrink)

A fundamental claim associated with parallel distributed processing theories of cognition is that knowledge is coded in a distributed manner in mind and brain. This approach rejects the claim that knowledge is coded in a localist fashion, with words, objects, and simple concepts, that is, coded with their own dedicated representations. One of the putative advantages of this approach is that the theories are biologically plausible. Indeed, advocates of the PDP approach often highlight the close parallels between distributed representations learned (...) in connectionist models and neural coding in brain and often dismiss localist theories as biologically implausible. The author reviews a range a data that strongly challenge this claim and shows that localist models provide a better account of single-cell recording studies. The author also contrast local and alternative distributed coding schemes and argues that common rejection of grandmother cell theories in neuroscience is due to a misunderstanding about how localist models behave. The author concludes that the localist representations embedded in theories of perception and cognition are consistent with neuroscience; biology only calls into question the distributed representations often learned in PDP models. (shrink)

The neurophysiological evidence from the Miyashita group's experiments on monkeys as well as cognitive experience common to us all suggests that local neuronal spike rate distributions might persist in the absence of their eliciting stimulus. In Hebb's cell-assembly theory, learning dynamics stabilize such self-maintaining reverberations. Quasi-quantitive modeling of the experimental data on internal representations in association-cortex modules identifies the reverberations (delay spike activity) as the internal code (representation). This leads to cognitive and neurophysiological predictions, many following directly from the language (...) used to describe the activity in the experimental delay period, others from the details of how the model captures the properties of the internal representations. (shrink)

The study of memory is witnessing a spirited clash between proponents of traditional laboratory research and those advocating a more naturalistic approach to the study of “real-life” or “everyday” memory. The debate has generally centered on the “what” (content), “where” (context), and “how” (methods) of memory research. In this target article, we argue that the controversy discloses a further, more fundamental breach between two underlying memory metaphors, each having distinct implications for memory theory and assessment: Whereas traditional memory research has (...) been dominated by thestorehousemetaphor, leading to a focus on thenumberof items remaining in store and accessible to memory, the recent wave of everyday memory research has shifted toward acorrespondencemetaphor, focusing on theaccuracyof memory in representing past events. The correspondence metaphor calls for a research approach that differs from the traditional one in important respects: in emphasizing the intentional –representational function of memory, in addressing the wholistic and graded aspects of memory correspondence, in taking an output-bound assessment perspective, and in allowing more room for the operation of subject-controlled metamemory processes and motivational factors. This analysis can help tie together soine of the what, where, and how aspects of the “real-life/laboratory” controversy. More important, however, by explicating the unique metatheoretical foundation of the accuracy-oriented approach to memory we aim to promote a more effective exploitation of the correspondence metaphor inbothnaturalistic and laboratory research contexts. (shrink)

Thought experiments have been influential in philosophy at least since Plato, and they have contributed to science at least since Galileo. Some of this influence is appropriate, because thought experiments can have legitimate roles in generating and clarifying hypotheses, as well as in identifying problems in competing hypotheses. I will argue, however, that philosophers have often overestimated the significance of thought experiments by supposing that they can provide evidence that supports the acceptance of beliefs. Accepting hypotheses merely on the basis (...) of thinking about them constitutes a kind of epistemic hubris with many negative consequences, including the acquisition of false beliefs and the blocking of .. (shrink)

Previous neural network models of concept learning were mainly implemented with supervised learning schemes. However, studies of human conceptual memory have shown that concepts may be learned without a teacher who provides the category name to associate with exemplars. A modular neural network architecture that realizes concept acquisition through two functionally distinct operations, categorizing and naming, is proposed as an alternative. An unsupervised algorithm realizes the categorizing module by constructing representations of categories compatible with prototype theory. The naming module associates (...) category names to the output of the categorizing module in a supervised mode. In such a modular architecture, the interface between the modules can be conceived of as an “information relay” that encodes, constrains, and propagates important information. Five experiments were conducted to analyze the relationships among internal conceptual codes and simple conceptual and lexical development. The first two experiments show a prototype effect and illustrate some basic characteristics of the system. The third experiment presents a bottom‐up model of the narrowing down of children's early lexical categories that honors mutual exclusivity. The fourth experiment introduces top‐down constraints on conceptual coding. The fifth experiment exhibits how hierarchical relationships between concepts are learned by the architecture, and also demonstrates how a spectrum of conceptual expertise may gradually emerge as a consequence of experiencing more with certain categories than with others. (shrink)

A set of hypotheses is formulated for a connectionist approach to cognitive modeling. These hypotheses are shown to be incompatible with the hypotheses underlying traditional cognitive models. The connectionist models considered are massively parallel numerical computational systems that are a kind of continuous dynamical system. The numerical variables in the system correspond semantically to fine-grained features below the level of the concepts consciously used to describe the task domain. The level of analysis is intermediate between those of symbolic cognitive models (...) and neural models. The explanations of behavior provided are like those traditional in the physical sciences, unlike the explanations provided by symbolic models.Higher-level analyses of these connectionist models reveal subtle relations to symbolic models. Parallel connectionist memory and linguistic processes are hypothesized to give rise to processes that are describable at a higher level as sequential rule application. At the lower level, computation has the character of massively parallel satisfaction of soft numerical constraints; at the higher level, this can lead to competence characterizable by hard rules. Performance will typically deviate from this competence since behavior is achieved not by interpreting hard rules but by satisfying soft constraints. The result is a picture in which traditional and connectionist theoretical constructs collaborate intimately to provide an understanding of cognition. (shrink)

How people learn chunks or associations between adjacent items in sequences was modelled. Two previously successful models of how people learn artificial grammars were contrasted: the CCN, a network version of the competitive chunker of Servan‐Schreiber and Anderson [J. Exp. Psychol.: Learn. Mem. Cogn. 16 (1990) 592], which produces local and compositionally‐structured chunk representations acquired incrementally; and the simple recurrent network (SRN) of Elman [Cogn. Sci. 14 (1990) 179], which acquires distributed representations through error correction. The models' susceptibility to two (...) types of interference was determined: prediction conflicts, in which a given letter can predict two other letters that appear next with an unequal frequency; and retroactive interference, in which the prediction made by a letter changes in the second half of training. The predictions of the models were determined by exploring parameter space and seeing howdensely different regions of the space of possible experimental outcomes were populated by model outcomes. For both types of interference, human data fell squarely in regions characteristic of CCN performance but not characteristic of SRN performance. (shrink)

A technique for the bilateral activation of neural nets that leads to a functional asymmetry of two simulated ''cerebral hemispheres'' is described. The simulation is designed to perform object recognition, while exhibiting characteristics typical of human consciousness-specifically, the unitary nature of conscious attention, together with a dual awareness corresponding to the ''nucleus'' and ''fringe'' described by William James (1890). Sensory neural nets self-organize on the basis of five sensory features. The system is then taught arbitrary symbolic labels for a small (...) number of similar stimuli. Finally, the trained network is exposed to nonverbal stimuli for object recognition, leading to Gaussian activation of the ''sensory'' maps-with a peak at the location most closely related to the features of the external stimulus. ''Verbal'' maps are activated most strongly at the labeled location that lies closest to the peak on homologous sensory maps. On the verbal maps activation is characterized by both excitatory and inhibitory Gaussians (a Mexican hat), the parameters of which are determined by the relative locations of the verbal labels. Mutual homotopic inhibition across the ''corpus callosum'' then produces functional cerebral asymmetries, i.e., complementary activation of homologous ''association'' and ''frontal'' maps within a common focus of attention-a nucleus in the left hemisphere and a fringe in the right hemisphere. An object is recognized as corresponding to a known label when the total activation of both hemispheres (nucleus plus fringe) is strongest for that label. The functional dualities of the cerebral hemispheres are discussed in light of the nucleus/fringe asymmetry. (shrink)

In the first section of the article, we examine some recent criticisms of the connectionist enterprise: first, that connectionist models are fundamentally behaviorist in nature (and, therefore, non-cognitive), and second that connectionist models are fundamentally associationist in nature (and, therefore, cognitively weak). We argue that, for a limited class of connectionist models (feed-forward, pattern-associator models), the first criticism is unavoidable. With respect to the second criticism, we propose that connectionist modelsare fundamentally associationist but that this is appropriate for building models (...) of human cognition. However, we do accept the point that there are cognitive capacities for which any purely associative model cannot provide a satisfactory account. The implication that we draw from is this is not that associationist models and mechanisms should be scrapped, but rather that they should be enhanced.In the next section of the article, we identify a set of connectionist approaches which are characterized by “active symbols” — recurrent circuits which are the basis of knowledge representation. We claim that such approaches avoid criticisms of behaviorism and are, in principle, capable of supporting full cognition. In the final section of the article, we speculate at some length about what we believe would be the characteristics of a fully realized active symbol system. This includes both potential problems and possible solutions (for example, mechanisms needed to control activity in a complex recurrent network) as well as the promise of such systems (in particular, the emergence of knowledge structures which would constitute genuine internal models). (shrink)

Koriat & Goldsmith are right in their claim that the “ecological” and “traditional” approaches to memory rely on different metaphors. But the underlying ecological metaphor is notcorrespondence(which in any case is not a metaphorical notion): it isaction. Remembering is a kind of doing; like most other forms of action it is purposive, personal, and particular.