There is currently a debate over whether cognitive architecture is classical or connectionist in nature. One finds the following three comparisons between classical architecture and connectionist architecture made in the pro-connectionist literature in this debate: (1) connectionist architecture is neurally plausible and classical architecture is not; (2) connectionist architecture is far better suited to model pattern recognition capacities than is classical architecture; and (3) connectionist architecture is far better suited to model the acquisition of pattern recognition capacities by learning than (...) is classical architecture. If true, (1)–(3) would yield a compelling case against the view that cognitive architecture is classical, and would offer some reason to think that cognitive architecture may be connectionist. We first present the case for (1)–(3) in the very words of connectionist enthusiasts. We then argue that the currently available evidence fails to support any of (1)–(3). (shrink)

There is currently a debate over whether cognitive architecture is classical or connectionist in nature. One finds the following three comparisons between classical architecture and connectionist architecture made in the pro-connectionist literature in this debate: (1) connectionist architecture is neurally plausible and classical architecture is not; (2) connectionist architecture is far better suited to model pattern recognition capacities than is classical architecture; and (3) connectionist architecture is far better suited to model the acquisition of pattern recognition capacities by learning than (...) is classical architecture. If true, (1)-(3) would yield a compelling case against the view that cognitive architecture is classical, and would offer some reason to think that cognitive architecture may be connectionist. We first present the case for (1)-(3) in the very words of connectionist enthusiasts. We then argue that the currently available evidence fails to support any of (1)-(3). (shrink)

One of the main challenges that Jerry Fodor and Zenon Pylyshyn (Cognition 28:3–71, 1988) posed for any connectionist theory of cognitive architecture is to explain the systematicity of thought without implementing a Language of Thought (LOT) architecture. The systematicity challenge presents a dilemma: if connectionism cannot explain the systematicity of thought, then it fails to offer an adequate theory of cognitive architecture; and if it explains the systematicity of thought by implementing a LOT architecture, then it fails to offer an (...) alternative to the LOT hypothesis. Given that thought is systematic, connectionism can offer an adequate alternative to the LOT hypothesis only if it can meet the challenge. Although some critics tried to meet the challenge, others argued that it need not be met since thought is not in fact systematic; and some claimed not to even understand the claim that thought is systematic. I do not here examine attempts to answer the challenge. Instead, I defend the challenge itself by explicating the notion of systematicity in a way that I hope makes clear that thought is indeed systematic, and so that to offer an adequate alternative to the LOT hypothesis, connectionism must meet the challenge. (shrink)

Smolensky's (1995) proposal for a connectionist explanation of systematicity doesn't work.

I consider the role of cognitive modeling in cognitive science. Modeling, and the computers that enable it, are central to the field, but the role of modeling is often misunderstood. Models are not intended to capture fully the processes they attempt to elucidate. Rather, they are explorations of ideas about the nature of cognitive processes. In these explorations, simplification is essential—through simplification, the implications of the central ideas become more transparent. This is not to say that simplification has no downsides; (...) it does, and these are discussed. I then consider several contemporary frameworks for cognitive modeling, stressing the idea that each framework is useful in its own particular ways. Increases in computer power (by a factor of about 4 million) since 1958 have enabled new modeling paradigms to emerge, but these also depend on new ways of thinking. Will new paradigms emerge again with the next 1,000‐fold increase? (shrink)

The study of human intelligence was once dominated by symbolic approaches, but over the last 30 years an alternative approach has arisen. Symbols and processes that operate on them are often seen today as approximate characterizations of the emergent consequences of sub- or nonsymbolic processes, and a wide range of constructs in cognitive science can be understood as emergents. These include representational constructs (units, structures, rules), architectural constructs (central executive, declarative memory), and developmental processes and outcomes (stages, sensitive periods, neurocognitive (...) modules, developmental disorders). The greatest achievements of human cognition may be largely emergent phenomena. It remains a challenge for the future to learn more about how these greatest achievements arise and to emulate them in artificial systems. (shrink)

Taking inspiration from developments in neurocomputational modeling, Paul Church-land develops his positions in the philosophy of mind and the philosophy of science. Concerning the former, Churchland relaxes his eliminativism at various points and seems to endorse a traditional identity account of sensory qualia. Although he remains unsympathetic to folk psychology, he no longer seeks the elimination of normative epistemology, but rather its transformation to a philosophical enterprise informed by current developments in the relevant sciences. Churchland supplies suggestive discussions of the (...) character of knowledge, simplicity, explanation, theory, and conceptual change. Many of his treatments turn on his prototype activation model of neural representation, which looks to the notion of a 'prototype' as it is employed in the psychological literature on concept representation, however, this and other features of Churchland's neurocomputational program do not square well with some of his views about cross-scientific relations. (shrink)

Fodor and Pylyshyn (1988), Fodor and McLaughlin (1990) and McLaughlin (1993) challenge connectionists to explain systematicity without simply implementing a classical architecture. In this paper I argue that what makes the challenge difficult for connectionists to meet has less to do with what is to be explained than with what is to count as an explanation. Fodor et al. are prepared to admit as explanatory, accounts of a sort that only classical models can provide. If connectionists are to meet the (...) challenge, they are going to have to insist on the propriety of changing what counts as an explanation of systematicity. Once that is done, there would seem to be as yet no reason to suppose that connectionists are unable to explain systematicity. (shrink)

Classicists and connectionists alike claim to be able to explain systematicity. The proposed classicist explanation, I argue, is little more than a promissory note, one that classicists have no idea how to redeem. Smolensky's (1995) proposed connectionist explanation fares little better: it is not vulnerable to recent classicist objections, but it nonetheless fails, particularly if one requires, as some classicists do, that explanations of systematicity take the form of a‘functional analysis’. Nonetheless, there are, I argue, reasons for cautious optimism about (...) the prospects of a connectionist explanation. (shrink)

Classicists and connectionists alike claim to be able to explain systematicity. The proposed classicist explanation, I argue, is little more than a promissory note, one that classicists have no idea how to redeem. Smolensky's (1995) proposed connectionist explanation fares little better: it is not vulnerable to recent classicist objections, but it nonetheless fails, particularly if one requires, as some classicists do, that explanations of systematicity take the form of a‘functional analysis’. Nonetheless, there are, I argue, reasons for cautious optimism about (...) the prospects of a connectionist explanation. (shrink)

Some theorists hold that the human perceptual system has a component that receives input only from units lower in the perceptual hierarchy. This thesis, that we shall here refer to as the encapsulation thesis, has been at the center of a continuing debate for the past few decades. Those who deny the encapsulation thesis often rely on the large body of psychological findings that allegedly suggest that perception is influenced by factors such as the beliefs, desires, goals, and the expectations (...) of the perceiver. Proponents of the encapsulation thesis, however, often argue that, when correctly interpreted, these psychological findings are compatible with the thesis. In our view, the debate over the significance and the correct interpretation of these psychological findings has reached an impasse. We hold that this impasse is due to the methodological limitations over psychophysical experiments, and it is very unlikely that such experiments, on their own, could yield results that would settle the debate. After defending this claim, we argue that integrating data from cognitive neuroscience resolves the debate in favor of those who deny the encapsulation thesis. (shrink)

The systematicity argument only challenges connectionism if systematicity is a general property of cognition. I examine this thesis in terms of properties of concepts. First, I propose that Evans's Generality Constraint only applies to attributions of belief. Then I defend a variety of conceptual pluralism, arguing that concepts share two fundamental properties related to centrality and belief-attribution, and contending that there are two kinds of concepts that differ in their compositional properties. Finally, I rely on Dual Systems Theory and on (...) differences between animal and human cognition to suggest a scenario of two processing systems that work on different kinds of concepts, with only one of them supporting full systematicity. I sketch a non-classical systematicity argument that rules out classicism as the basis of one of those systems given that it would wrongly entail that both systems are fully systematic. (shrink)

The article criticises the attempt to establish connectionism as an alternative theory of human cognitive architecture through the introduction of thesymbolic/subsymbolic distinction (Smolensky, 1988). The reasons for the introduction of this distinction are discussed and found to be unconvincing. It is shown that thebrittleness problem has been solved for a large class ofsymbolic learning systems, e.g. the class oftop-down induction of decision-trees (TDIDT) learning systems. Also, the process of articulating expert knowledge in rules seems quite practical for many important domains, (...) including common sense knowledge.The article discusses several experimental comparisons betweenTDIDT systems and artificial neural networks using the error backpropagation algorithm (ANNs usingBP). The properties of one of theTDIDT systemsID3 (Quinlan, 1986a) are examined in detail. It is argued that the differences in performance betweenANNs usingBP andTDIDT systems reflect slightly different inductive biases but are not systematic; these differences do not support the view that symbolic and subsymbolic systems are fundamentally incompatible. It is concluded, that thesymbolic/subsymbolic distinction is spurious. It cannot establish connectionism as an alternative cognitive architecture. (shrink)

Cognitive scientists must understand not just what the mind does, but how it does what it does. In this paper, I consider four aspects of cognitive architecture: how the mind develops, the extent to which it is or is not modular, the extent to which it is or is not optimal, and the extent to which it should or should not be considered a symbol‐manipulating device (as opposed to, say, an eliminative connectionist network). In each case, I argue that insights (...) from developmental and evolutionary biology can lead to substantive and important compromises in historically vexed debates. (shrink)

Explicitness has usually been approached from two points of view, labelled by Kirsh the structural and the process view, that hold opposite assumptions to determine when information is explicit. In this paper, we offer an intermediate view that retains intuitions from both of them. We establish three conditions for explicit information that preserve a structural requirement, and a notion of explicitness as a continuous dimension. A problem with the former accounts was their disconnection with psychological work on the issue. We (...) review studies by Karmiloff-Smith, and Shanks and St. John to show that the proposed conditions have psychological grounds. Finally, we examine the problem of explicit rules in connectionist systems in the light of our framework. (shrink)

Explicitness has usually been approached from two points of view, labelled by Kirsh the structural and the process view, that hold opposite assumptions to determine when information is explicit. In this paper, we offer an intermediate view that retains intuitions from both of them. We establish three conditions for explicit information that preserve a structural requirement, and a notion of explicitness as a continuous dimension. A problem with the former accounts was their disconnection with psychological work on the issue. We (...) review studies by Karmiloff-Smith, and Shanks and St. John to show that the proposed conditions have psychological grounds. Finally, we examine the problem of explicit rules in connectionist systems in the light of our framework. (shrink)

La cognición incorporada es antipresencialista y por lo mismo no sucumbe al externalismo objetual. Su naturaleza es archeológica y teleológica: en cuanto a la primera, todo conocer está encadenado al desarrollo de la especie, por lo mismo no se la entiende desde la sola dimensión antropológica. En cuanto a la segunda, es finalizada en comportamientos, representaciones, objetivaciones-cosistas. El conocer es incorporado porque es filogenético, entonces perteneciente a la especie; pero también es ontogenético, entonces perteneciente al individuo. El conocimiento del sujeto (...) individual es una consecuencia de la especie, en un ambiente particular, por ello es ontogenético. El eje articulante de toda la cognición nueva es, sobre todo, la neo-corteza, la cual es muy sensible a la periferia. (shrink)

According to radical versions of embodied cognition, human cognition and agency should be explained without the ascription of representational mental states. According to a standard reply, accounts of embodied cognition can explain only instances of cognition and agency that are not “representation-hungry”. Two main types of such representation-hungry phenomena have been discussed: cognition about “the absent” and about “the abstract”. Proponents of representationalism have maintained that a satisfactory account of such phenomena requires the ascription of mental representations. Opponents have denied (...) this. I will argue that there is another important representation-hungry phenomenon that has been overlooked in this debate: temporally extended planning agency. In particular, I will argue that it is very difficult to see how planning agency can be explained without the ascription of mental representations, even if we grant, for the sake of argument, that cognition about the absent and abstract can. We will see that this is a serious challenge for the radical as well as the more modest anti-representationalist versions of embodied cognition, and we will see that modest anti-representationalism is an unstable position. (shrink)

A standard view within psychology is that there have been two important shifts in the study of concepts and that each has led to some improvements. The first shift was from the classical theory of concepts to probabilistic theories, including the prototype theory. The second shift was from probabilistic theories to theory-based theories. In this article, I critically evaluate the view that the first shift was a major advance and argue that the prototype theory suffers some of the same problems (...) that have been thought to challenge the classical theory. (shrink)

O'Brien & Opie's (O&O's) theory demands a view of unconscious processing that is incompatible with virtually all current PDP models of neural activity. Relative to the alternatives, the theory is closer to an AI than a parallel distributed processing (PDP) perspective, and its treatment of phenomenology is ad hoc. It raises at least one important question: Could features of network relaxation be the “switch” that turns an unconscious into a conscious network?

We suggest a common ground for alternative proposals In different domains of cognitive science which have previously seemed to have little in common. The underlying common theme is associated with a redefinition of the basic unit of analysis in each domain of thought. Our framework suggests a definition of unity which is based not on inherent properties of the elements constituting the unit, but rather on dynamic patterns of correlation across the elements. We introduce a set of features that characterize (...) the new dynamic units, and distinguish them from traditional units of analysis. A conceptual connection is made with the identification of elementary units in modern physics theories, as well as with concepts and structures in the study of complex dynamical systems and connectionism. The paper analyses the evolution of the concept of unit in different domains of thought in cognitive science, and examines the proposed framework of “dynamic unity” with regard to various theoretical problems within each domain. The particular cognitive science issues discussed in the paper include: (1) the binding problem in the brain; (2) the mental unit of conceptual organization; (3) defining ‘wordhood’ in linguistics; and (4) identifying the unit of cognition in the natural environment. (shrink)

The Representational Theory of the Mind allows for psychological explanations couched in terms of the contents of propositional attitudes. Propositional attitudes themselves are taken to be relations to mental representations. These representations (partially) determine the contents of the attitudes in which they figure. Thus, Representationalism owes an explanation of the contents of mental representations. This essay constitutes an atomistic theory of the content of formally or syntactically simple mental representation, proposing that the content of such a representation is determined by (...) the intersection of the representation's correlational and control properties. The theory is distinguished from standard information-based accounts of mental content in allowing that the relevant correlations be contingent while insisting on an efferent aspect to mental content. The theory on offer allows for a natural explanation of misrepresentation, finds a niche for the notion of narrow content, welcomes radical first person fallibility with respect to questions of content, admits of mental ambiguity and recognizes that the future of a psychological agent is a factor in determining the content of the agent's present psychological states. (shrink)

Philosophers of science today by and large reject the cataclysmic and irrational interpretation of the scientific enterprise claimed by Kuhn. Many computational models have been implemented to rationally study the conceptual change in science. In this recent tradition a key role is played by the concept of abduction as a mechanism by which new explanatory hypotheses are introduced. Nevertheless some problems in describing the most interesting abductive issues rise from the classical computational approach. It describes a cognitive process (and so (...) abduction) by the manipulation of internal symbolic representations of external world. This view assumes a discrete set of representations fixed in discrete time jumps, and cannot adequately account for the issue of anticipation and causation of a new hypothesis. An integration of the traditional computational view with some ideas developed inside the so-called dynamical approach can suggest some important insights. The concept of attractor is very significant. It permits a description of the abductive generation of new hypotheses in terms of a catastrophic rearrangement of the parameters responsible for the behavior of the system. (shrink)

The question, ‘Is cognition linguistic?' divides recent cognitive theories into two antagonistic groups. Sententialists claim that we think in some language, while advocates of non linguistic views of cognition deny this claim. The Introspective Argument for Sententialism is one of the most appealing arguments for sententialism. In substance, it claims that the introspective fact of inner speech provides strong evidence that our thoughts are linguistic. This article challenges this argument. I claim that the Introspective Argument for Sententialism confuses the content (...) of our thoughts with their vehicles: while sententialism is a thesis about the vehicles of our thoughts, inner speech sentences are the content of auditory or articulatory images. The rebuttal of the introspective argument for sententialism is shown to have a general significance in cognitive science: Introspection does not tell us how we think. (shrink)

This commentary challenges the authors to use their computational modeling techniques to support one of their central claims: that schizophrenic deficits in bottom-up (Gestalt-type tasks) and top-down (cognitive control tasks) context processing tasks arise from the same dysfunction. Further clarification about the limits of cognitive coordination would also strengthen the hypothesis.

Perceptual representations are sometimes said to be iconic, or picture-like. But what does this mean, and is it true? I suggest that the most fruitful way to understand iconicity is in terms of similarity, but there are three importantly different grades of similarity that that might hold between perceptual representations and their objects, and these should be distinguished. It is implausible that all perceptual representations achieve even the weakest grade of iconicity, but I speculatively suggest a “Kantian” view, whereby all (...) perceptual representation involves fillers placed in a “strongly iconic” spatial framework. Those fillers will likely be hybrid structures, with varying grades of iconicity combined with non-iconic elements. (shrink)

The traditional understanding of analyticity in terms of concept containment is revisited, but with a concept explicitly understood as a certain kind of mental representation and containment being read correspondingly literally. The resulting conception of analyticity avoids much of the vagueness associated with attempts to explicate analyticity in terms of synonymy by moving the locus of discussion from the philosophy of language to the philosophy of mind. The account provided here illustrates some interesting features of representations and explains, at least (...) in part, the special epistemic status of analytic judgments. (shrink)

The paper offers a solution to the generality problem for a reliabilist epistemology, by developing an “algorithm and parameters” scheme for type-individuating cognitive processes. Algorithms are detailed procedures for mapping inputs to outputs. Parameters are psychological variables that systematically affect processing. The relevant process type for a given token is given by the complete algorithmic characterization of the token, along with the values of all the causally relevant parameters. The typing that results is far removed from the typings of folk (...) psychology, and from much of the epistemology literature. But it is principled and empirically grounded, and shows good prospects for yielding the desired epistemological verdicts. The paper articulates and elaborates the theory, drawing out some of its consequences. Toward the end, the fleshed-out theory is applied to two important case studies: hallucination and cognitive penetration of perception. (shrink)

One holistic thesis about symbols is that a symbol cannot exist singly, but only as apart of a symbol system. There is also the plausible view that symbol systems emerge gradually in an individual, in a group, and in a species. The problem is that symbol holism makes it hard to see how a symbol system can emerge gradually, at least if we are considering the emergence of a first symbol system. The only way it seems possible is if being (...) a symbol can be a matter of degree, which is initially problematic. This article explains how being a cognitive symbol can be a matter of degree after all. The contrary intuition arises from the way a process of interpretation forces an all-or-nothing character on symbols, leaving room for underlying material to realize symbols to different degrees in a way that Daniel Dennett’s work can help illuminate. Holism applies to symbols as interpreted, while gradualism applies to how the underlying material realizes symbols.Selon une thèse holistique sur les symboles, un symbole ne peut exister isolément mais doit faire partie d’un systéme symbolique. Une opinion, elle aussi plausible, veut que les systèmes symboliques émergent graduellement chez un individu, un groupe ou une espèce. Le problème c’est qu’on voit mal, si le holisme des systèmes symboliques tient, comment un système symbolique peut émerger graduellement, du moins pour la première fois. Ce n’est possible, semble-t-il, que si être un symbole est affaire de degré, thèse au départ problématique. Cet article explique comment être un symbole cognitif peut après tout être affaire de degré. L’intuition contraire vient de ce que le processus d’interprétation nous force au tout ou rien, ce qui laisse un jeu dans la façon dont le matériel sous-jacent réalise les symboles à des degrés divers. Les travaux de Daniel Dennett sont à cet égard éclairants. Le holisme vaut pour les symboles tels qu’ils sont interprétés, tandis que le gradualisme vaut pour la façon dont le matériel sous-jacent réalise les symboles. (shrink)