We begin by distinguishing computationalism from a number of other theses that are sometimes conflated with it. We also distinguish between several important kinds of computation: computation in a generic sense, digital computation, and analog computation. Then, we defend a weak version of computationalism—neural processes are computations in the generic sense. After that, we reject on empirical grounds the common assimilation of neural computation to either analog or digital computation, concluding that neural computation is sui generis. Analog computation requires continuous (...) signals; digital computation requires strings of digits. But current neuroscientific evidence indicates that typical neural signals, such as spike trains, are graded like continuous signals but are constituted by discrete functional elements (spikes); thus, typical neural signals are neither continuous signals nor strings of digits. It follows that neural computation is sui generis. Finally, we highlight three important consequences of a proper understanding of neural computation for the theory of cognition. First, understanding neural computation requires a specially designed mathematical theory (or theories) rather than the mathematical theories of analog or digital computation. Second, several popular views about neural computation turn out to be incorrect. Third, computational theories of cognition that rely on non-neural notions of computation ought to be replaced or reinterpreted in terms of neural computation. (shrink)

This paper proposes how various disciplinary theories of cognition may be combined into a unifying, sub-symbolic, computational theory of cognition. The following theories are considered for integration: psychological theories, including the theory of event coding, event segmentation theory, the theory of anticipatory behavioral control, and concept development; artificial intelligence and machine learning theories, including reinforcement learning and generative artificial neural networks; and theories from theoretical and computational neuroscience, including predictive coding and free (...) energy-based inference. In the light of such a potential unification, it is discussed how abstract cognitive, conceptualized knowledge and understanding may be learned from actively gathered sensorimotor experiences. The unification rests on the free energy-based inference principle, which essentially implies that the brain builds a predictive, generative model of its environment. Neural activity-oriented inference causes the continuous adaptation of the currently active predictive encodings. Neural structure-oriented inference causes the longer term adaptation of the developing generative model as a whole. Finally, active inference strives for maintaining internal homeostasis, causing goal-directed motor behavior. To learn abstract, hierarchical encodings, however, it is proposed that free energy-based inference needs to be enhanced with structural priors, which bias cognitive development towards the formation of particular, behaviorally suitable encoding structures. As a result, it is hypothesized how abstract concepts can develop from, and thus how they are structured by and grounded in, sensorimotor experiences. Moreover, it is sketched-out how symbol-like thought can be generated by a temporarily active set of predictive encodings, which constitute a distributed neural attractor in the form of an interactive free-energy minimum. The activated, interactive network attractor essentially characterizes the semantics of a concept or a concept composition, such as an actual or imagined situation in our environment. Temporal successions of attractors then encode unfolding semantics, which may be generated by a behavioral or mental interaction with an actual or imagined situation in our environment. Implications, further predictions, possible verification and falsifications, as well as potential enhancements into a fully spelled-out unified theory of cognition are discussed at the end of the paper. (shrink)

We don't yet have adequate theories of what the human mind is representing when it represents a social group. Worse still, many people think we do. This mistaken belief is a consequence of the state of play: Until now, researchers have relied on their own intuitions to link up the concept social group on the one hand and the results of particular studies or models on the other. While necessary, this reliance on intuition has been purchased at a considerable cost. (...) When looked at soberly, existing theories of social groups are either literal, but not remotely adequate, or simply metaphorical. Intuition is filling in the gaps of an explicit theory. This paper presents a computational theory of what, literally, a group representation is in the context of conflict: It is the assignment of agents to specific roles within a small number of triadic interaction types. This “mental definition” of a group paves the way for a computational theory of social groups – in that it provides a theory of what exactly the information-processing problem of representing and reasoning about a group is. For psychologists, this paper offers a different way to conceptualize and study groups, and suggests that a non-tautological definition of a social group is possible. For cognitive scientists, this paper provides a computational benchmark against which natural and artificial intelligences can be held. (shrink)

Few areas of study have led to such close and intense interactions among computer scientists, psychologists, and philosophers as the area now referred to as cognitive science. Within this discipline, few problems have inspired as much debate as the use of notions such as meaning, intentionality, or the semantic content of mental states in explaining human behavior. The set of problems surrounding these notions have been viewed by some observers as threatening the foundations of cognitive science as currently conceived, and (...) by others as providing a new and scientifically sound formulation of certain classical problems in the philosophy of mind. The chapters in this volume help bridge the gap among contributing disciplines-computer science, philosophy, psychology, neuroscience-and discuss the problems posed from various perspectives. (shrink)

Computational modeling is now ubiquitous in psychology, and researchers who are not modelers may find it increasingly difficult to follow the theoretical developments in their field. This book presents an integrated framework for the development and application of models in psychology and related disciplines. Researchers and students are given the knowledge and tools to interpret models published in their area, as well as to develop, fit, and test their own models. Both the development of models and key features of (...) any model are covered, as are the applications of models in a variety of domains across the behavioural sciences. A number of chapters are devoted to fitting models using maximum likelihood and Bayesian estimation, including fitting hierarchical and mixture models. Model comparison is described as a core philosophy of scientific inference, and the use of models to understand theories and advance scientific discourse is explained. (shrink)

The Computational Theory of Mind (CTM) holds that the mind is a computer and that cognition is the manipulation of representations. CTM is commonly viewed as the main hypothesis in cognitive science, with classical CTM (related to the Language of Thought Hypothesis) being the most popular variant. However, other computational accounts of the mind either reject LOTH or do not subscribe to RTM. CTM proponents argue that it clarifies how thought and content are causally relevant in (...) the physical world, and they frame this argument in terms of the physical symbol hypothesis, formalism, or syntactic engines. This article focuses on specific problems with CTM, with four main sections covering the introduction of the three main variants of CTM, the most important conceptions of computational explanation in cognitive science, skeptical arguments against CTM raised by Hilary Putnam, and common objections to CTM. (shrink)

The book brings into relief the variety of approaches and disciplines that have informed the quest for a theory of cognition. The center of interest are the historical, geographical, and theoretical peripheries of classic AI's mainstream research program. The twelve chapters bring back into focus the variety of strategies and theoretical questions that researchers explored while working toward a scientific theory of cognition and pre-cognition. The volume is organized in four parts, each one including three (...) essays. The first one deals with cybernetics, the approach that may be considered as the most important periphery of classic AI research. The second part focuses on the geographical periphery of AI research. It examines how the theories and techniques developed on AI's home ground were translated into countries with different cultures and traditions: Italy, France, and the Soviet Union. The third part focuses on AI's periphery understood in the cultural and historical meaning of the term. It contains essays that locate some of the central concepts of AI, like representation and computability, within a broader philosophical (Descartes, Aristotle, Leibniz) and technical background (programming theory and practice). The fourth and final part of the volume is focused directly on the limitation of Turing's classic computability theory and its possible alternatives, some of which were studied in the early years of AI's research (e.g. Ashby's re-entrant information model), while others have been intensely studied in recent times (quantum automata). (shrink)

Turing's analysis of computation is a fundamental part of the background of cognitive science. In this paper it is argued that a re‐interpretation of Turing's work is required to underpin theorizing about cognitive architecture. It is claimed that the symbol systems view of the mind, which is the conventional way of understanding how Turing's work impacts on cognitive science, is deeply flawed. There is an alternative interpretation that is more faithful to Turing's original insights, avoids the criticisms made of the (...) symbol systems approach and is compatible with the growing interest in agent‐environment interaction. It is argued that this interpretation should form the basis for theories of cognitive architecture. (shrink)

Moods have global and profound effects on our thoughts, motivations and behavior. To understand human behavior and cognition fully, we must understand moods. In this paper I critically examine and reject the methodology of conventional ?cognitive theories? of affect. I lay the foundations of a new theory of moods that identifies them with processes of our cognitive functional architecture. Moods differ fundamentally from some of our other affective states and hence require distinct explanatory tools. The computational (...) class='Hi'>theory of mood I propose places them within the context of other mental phenomena and is consistent with the empirical data on moods. (shrink)

In this chapter, I argue that some aspects of cognitive phenomena cannot be explained computationally. In the first part, I sketch a mechanistic account of computational explanation that spans multiple levels of organization of cognitive systems. In the second part, I turn my attention to what cannot be explained about cognitive systems in this way. I argue that information-processing mechanisms are indispensable in explanations of cognitive phenomena, and this vindicates the computational explanation of cognition. At the same (...) time, it has to be supplemented with other explanations to make the mechanistic explanation complete, and that naturally leads to explanatory pluralism in cognitive science. The price to pay for pluralism, however, is the abandonment of the traditional autonomy thesis asserting that cognition is independent of implementation details. (shrink)

The book presents the case that cognitive science should turn its attention to developing theories of human cognition that cover the full range of human perceptual, cognitive, and action phenomena. Cognitive science has now produced a massive number of high-quality regularities with many microtheories that reveal important mechanisms. The need for integration is pressing and will continue to increase. Equally important, cognitive science now has the theoretical concepts and tools to support serious attempts at unified theories. The argument is (...) made entirely by presenting an exemplar unified theory of cognition both to show what a real unified theory would be like and to provide convincing evidence that such theories are feasible. The exemplar is SOAR, a cognitive architecture, which is realized as a software system. After a detailed discussion of the architecture and its properties, with its relation to the constraints on cognition in the real world and to existing ideas in cognitive science, SOAR is used as theory for a wide range of cognitive phenomena: immediate responses ; discrete motor skills ; memory and learning ; problem solving ; language ; and development. The treatments vary in depth and adequacy, but they clearly reveal a single, highly specific, operational theory that works over the entire range of human cognition, SOAR is presented as an exemplar unified theory, not as the sole candidate. Cognitive science is not ready yet for a single theory – there must be multiple attempts. But cognitive science must begin to work toward such unified theories. (shrink)

In this paper we start from the assumption that in a metaphor, or an analogy, some terms belonging to one domain (source domain) are used to refer to objects other than their conventional referents belonging to a possibly different domain (target domain). We describe a formalism, which is based on the First Order Predicate Calculus, for representing the knowledge structure associated with a domain and then develop a theory of Constrained Semantic Transference [CST] which allows the terms and the (...) structural relationships of the source domain to be transferred coherently across to the target domain. We show how metaphors and analogies can be characterized in CST in such a way that many of their cognitive properties con be explained.We then propose a theory of Approximate Semantic Transference [AST] which is a computational version of CST and is derived from it by replacing the coherency requirement with approximate coherency. We show how AST can be used as a basis for designing models of cognitive processes involved in comprehending metaphors and analogies. (shrink)

Contemporary philosophy and theoretical psychology are dominated by an acceptance of content-externalism: the view that the contents of one's mental states are constitutively, as opposed to causally, dependent on facts about the external world. In the present work, it is shown that content-externalism involves a failure to distinguish between semantics and pre-semantics---between, on the one hand, the literal meanings of expressions and, on the other hand, the information that one must exploit in order to ascertain their literal meanings. It is (...) further shown that, given the falsity of content-externalism, the falsity of the Computational Theory of Mind (CTM) follows. It is also shown that CTM involves a misunderstanding of terms such as "computation," "syntax," "algorithm," and "formal truth." Novel analyses of the concepts expressed by these terms are put forth. These analyses yield clear, intuition-friendly, and extensionally correct answers to the questions "what are propositions?, "what is it for a proposition to be true?", and "what are the logical and psychological differences between conceptual (propositional) and non-conceptual (non-propositional) content?" Naively taking literal meaning to be in lockstep with cognitive content, Burge, Salmon, Falvey, and other semantic externalists have wrongly taken Kripke's correct semantic views to justify drastic and otherwise contraindicated revisions of commonsense. (Salmon: What is non-existent exists; at a given time, one can rationally accept a proposition and its negation. Burge: Somebody who is having a thought may be psychologically indistinguishable from somebody who is thinking nothing. Falvey: somebody who rightly believes himself to be thinking about water is psychologically indistinguishable from somebody who wrongly thinks himself to be doing so and who, indeed, isn't thinking about anything.) Given a few truisms concerning the differences between thought-borne and sentence-borne information, the data is easily modeled without conceding any legitimacy to any one of these rationality-dismantling atrocities. (It thus turns out, ironically, that no one has done more to undermine Kripke's correct semantic points than Kripke's own followers!). (shrink)

One of the most debated questions in psychology and cognitive science is the nature and the functioning of the mental processes involved in deductive reasoning. However, all existing theories refer to a specific deductive domain, like syllogistic, propositional or relational reasoning.Our goal is to unify the main types of deductive reasoning into a single set of basic procedures. In particular, we bring together the microtheories developed from a mental models perspective in a single theory, for which we provide a (...) formal foundation. We validate the theory through a computational model (UNICORE) which allows fine‐grained predictions of subjects' performance in different reasoning domains.The performance of the model is tested against the performance of experimental subjects—as reported in the relevant literature—in the three areas of syllogistic, relational and propositional reasoning. The computational model proves to be a satisfactory artificial subject, reproducing both correct and erroneous performance of the human subjects. Moreover, we introduce a developmental trend in the program, in order to simulate the performance of subjects of different ages, ranging from children (3–6) to adolescents (8–12) to adults (>21). The simulation model performs similarly to the subjects of different ages.Our conclusion is that the validity of the mental model approach is confirmed for the deductive reasoning domain, and that it is possible to devise a unique mechanism able to deal with the specific subareas. The proposed computational model (UNICORE) represents such a unifying structure. (shrink)

summaryThis paper argues that an epistemological duality between mind/brain and an external world is an uncritically held working assumption in recent computational models of cognition. In fact, epistemological dualism largely drives computational models of mentality and representation: An assumption regarding an external world of perceptual objects and distal stimuli requires the sort of mind/brain capable of representing and inferring true accounts of such objects. Hence we have two distinct ontologies, one denoting external world objects, the other cognitive (...) events and neural transformations. A basic question is then raised and explored: If the elements of these two different ontologies are so radically different, how can the one access the other?The last part of the paper sketches an alternative to ED and the associated two ontologies account. Instead of external world objects being construed as independent givens, initiating perceptual processes, they are rather interpreted as ‘computational constructs’. It is argued at length that this account is far more consistent with recent computational models of mind/brain than the two ontologies model associated with ED. It is further claimed that an ontology of cognitive and neural states, as well as an external world ontology, have no other epistemic status that as contingent working hypotheses. As such, both ontologies are in principle amenable to revision as evidence, explanatory models, and scientific practices mutually readjust and redefine one another over time. (shrink)

Empirical analyses have provided some important constraints for computational theories of metaphor. Three such constraints relate to (1) the similar processing time for literal and metaphorical language, (2) the time‐limited processing of many metaphors, and (3) the dissociation of metaphor comprehension and appreciation. Indurkhya's (1986, 1987) model is discussed with respect to these issues.

_Abstract_: This article discusses the unity of cognitive science that seemed to emerge in the 1950s, based on the computational view of cognition. This unity would entail that there is a single set of mechanisms (i.e. algorithms) for all cognitive behavior, in particular at the level of productive human cognition as exemplified in language and reasoning. In turn, this would imply that theories in psychology, and cognitive science in general, would consist of algorithms based on symbol manipulation (...) as found in digital computing. However, a number of developments in recent decades cast doubt on this unity of cognitive science. Also, there are fundamental problems with the claim that cognitive theories are just algorithms. This article discusses some of these problems and suggests that, instead of unified theories of cognition, specific mechanisms for cognitive behavior in specific cognitive domains could be needed, with architectures that are tailor-made for specific forms of implementation. A sketch of such an architecture for language is presented, based on modifiable connection paths in small-world like network structures. _Keywords_: Connection Paths; Control of Activation; Small-world Networks; Symbol Manipulation; Unity of Cognition _Dalle teoria unificate della cognizione a quelle specifiche_ _Riassunto_: Questo articolo discute l’unità della scienza cognitiva che sembrava emergere negli Anni ’50 e che era basata su una concezione computazionale della cognizione. Questa unità prevedeva l’esistenza di un singolo insieme di meccanismi (algoritmi) per tutti i comportamenti cognitivi, in particolare al livello della cognizione umana produttiva come, per esempio, linguaggio e ragionamento. A sua volta ciò implicava che le teorie psicologiche e, più in generale della scienza cognitiva, prevedessero algoritmi basati sulla manipolazione di simboli come nella computazione digitale. E, tuttavia, diversi sviluppi degli ultimi decenni hanno messo in dubbio questa unità della scienza cognitiva. Affermare che le teorie cognitive sarebbero solo algoritmi presenta problemi di fondo. Questo articolo discute alcuni di questi problemi, suggerendo che, invece di teorie della cognizione unificata, si potrebbe aver bisogno di meccanismi specifici per il comportamento cognitivo in specifici domini cognitivi, con architetture ritagliate per specifiche forme di implementazione. Questo articolo presenta uno schizzo di una simile architettura per il linguaggio, basata su vie di connessione modificabili in piccoli mondi come le strutture di reti. _Parole chiave_: Vie di connessione; Controllo dell’attivazione; Reti di piccoli mondi; Manipolazione di simboli; Unità della cognizione. (shrink)

What are the conditions under which groups of agents will perform well across multiple tasks? The author establishes a set of “alignment conditions” that enforce identity of beliefs and desires across agents. These conditions are necessary and sufficient for ensuring that a multiagent system behaves as if controlled by a rational centralized controller. Several widely observed social phenomena can be understood as facilitating the alignment conditions.

summaryThis paper argues that an epistemological duality between mind/brain and an external world is an uncritically held working assumption in recent computational models of cognition. In fact, epistemological dualism largely drives computational models of mentality and representation: An assumption regarding an external world of perceptual objects and distal stimuli requires the sort of mind/brain capable of representing and inferring true accounts of such objects. Hence we have two distinct ontologies, one denoting external world objects, the other cognitive (...) events and neural transformations. A basic question is then raised and explored: If the elements of these two different ontologies are so radically different, how can the one access the other?The last part of the paper sketches an alternative to ED and the associated two ontologies account. Instead of external world objects being construed as independent givens, initiating perceptual processes, they are rather interpreted as ‘computational constructs’. It is argued at length that this account is far more consistent with recent computational models of mind/brain than the two ontologies model associated with ED. It is further claimed that an ontology of cognitive and neural states, as well as an external world ontology, have no other epistemic status that as contingent working hypotheses. As such, both ontologies are in principle amenable to revision as evidence, explanatory models, and scientific practices mutually readjust and redefine one another over time. (shrink)

Computation is central to the foundations of modern cognitive science, but its role is controversial. Questions about computation abound: What is it for a physical system to implement a computation? Is computation sufficient for thought? What is the role of computation in a theory of cognition? What is the relation between different sorts of computational theory, such as connectionism and symbolic computation? In this paper I develop a systematic framework that addresses all of these questions. Justifying (...) the role of computation requires analysis of implementation, the nexus between abstract computations and concrete physical systems. I give such an analysis, based on the idea that a system implements a computation if the causal structure of the system mirrors the formal structure of the computation. This account can be used to justify the central commitments of artificial intelligence and computational cognitive science: the thesis of computational sufficiency, which holds that the right kind of computational structure suffices for the possession of a mind, and the thesis of computational explanation, which holds that computation provides a general framework for the explanation of cognitive processes. The theses are consequences of the facts that (a) computation can specify general patterns of causal organization, and (b) mentality is an organizational invariant, rooted in such patterns. Along the way I answer various challenges to the computationalist position, such as those put forward by Searle. I close by advocating a kind of minimal computationalism, compatible with a very wide variety of empirical approaches to the mind. This allows computation to serve as a true foundation for cognitive science. (shrink)

Marr's computational theory of stereopsis is shown to imply that human vision employs a system of representation which has all the properties of a number system. Claims for an internal number system and for neural computation should be taken literally. I show how these ideas withstand various skeptical attacks, and analyze the requirements for describing neural operations as computations. Neural encoding of numerals is shown to be distinct from our ability to measure visual physiology. The constructs in Marr's (...) theory are neither propositional nor pictorial, and provide a counter example to many commonly held dichotomies concerning mental representation. (shrink)

Upshot: Written by recognized experts in their fields, the book is a set of essays that deals with the influences of early cybernetics, computational theory, artificial intelligence, and connectionist networks on the historical development of computational-representational theories of cognition. In this review, I question the relevance of computability arguments and Jonasian phenomenology, which has been extensively invoked in recent discussions of autopoiesis and Ashby’s homeostats. Although the book deals only indirectly with constructivist approaches to cognition, (...) it is useful reading for those interested in machine-based models of mind. (shrink)

In this paper I review some leading developments in the empirical theory of affect. I argue that (1) affect is a distinct perceptual representation governed system, and (2) that there are significant modular factors in affect. The paper concludes with the observation thatfeeler (affective perceptual system) may be a natural kind within cognitive science. The main purpose of the paper is to explore some hitherto unappreciated connections between the theory of affect and the computational theory of (...) mind. (shrink)

An important, but relatively neglected, aspect of human theory of mind is emotion inference: understanding how and why a person feels a certain why is central to reasoning about their beliefs, desires and plans. The authors review recent work that has begun to unveil the structure and determinants of emotion inference, organizing them within a unified probabilistic framework.

In The Mind Doesn’t Work that Way, Jerry Fodor argues that mental representations have context sensitive features relevant to cognition, and that, therefore, the Classical Computational Theory of Mind (CTM) is mistaken. We call this the Globality Argument. This is an in principle argument against CTM. We argue that it is self-defeating. We consider an alternative argument constructed from materials in the discussion, which avoids the pitfalls of the official argument. We argue that it is also unsound (...) and that, while it is an empirical issue whether context sensitive features of mental representations are relevant to cognition, it is empirically implausible. (shrink)

Hubert Dreyfus once noted that it would be difficult to ascertain whether Edmund Husserl had a computational theory of mind. I provide evidence that he had one. Both Steven Pinker and Steven Horst think that the computational theory of mind must have two components: a representational-symbolic component and a causal component. Bearing this in mind, we proceed to a close-reading of the sections of “On the Logic of Signs” wherein Husserl presents, if I’m correct, his (...) class='Hi'>computational theory of mind embedded in a language of thought. My argument goes like this: the computational theory of mind is the idea, following Haugeland, that the mind comes prepackaged as, or is endogenously constrained to be, an automatic formal system; this explains, according to Husserl, why automatic trains of thought without logical intent resemble arguments exhibiting deductive structure with logical intent. In general, an automatic formal system yields true results provided that the syntactic symbols with which they compute are univocal and are semantically evaluable, and the mechanized inferences they perform are valid and preserve truth. These two conditions describe a computational cognitive process: the first condition connects representations to syntax, and the second condition uses the syntax, in inauthentic judging, to arrive at true conclusions through blind causality. Now, in point of textual fact, these are the conditions which Husserl attributes to our “natural psychological mechanism of symbolic inference” which typically yields true results. Since a formal system attributed to the “internal structure” of the mind, and guided by blind causality, just is the computational theory of mind, it follows, I think, that Husserl had a computational theory of mind. This computational theory is, moreover, embedded in a language of thought, since Husserl attributes a language-like form to our thoughts so that they may be mechanically processed. I conclude with a discussion of my results. (shrink)

A major challenge for research in artificial intelligence is to develop systems that can infer the goals, beliefs, and intentions of others (i.e., systems that have theory of mind, ToM). In this research, we propose a cognitive ToM framework that uses a well-known theory of decisions from experience to construct a computational representation of ToM. Instance-based learning theory (IBLT) is used to construct a cognitive model that generates ToM from the observation of other agents' behavior. The (...) IBL model of the observer distinguishes itself from previous models of ToM that make unreasonable assumptions about human cognition, are hand-crafted for particular settings, complex, or unable to explain a cognitive development of ToM compared to human's ToM. The IBL model learns from the observation of goal-directed agents' behavior in a gridworld navigation task, and it infers and predicts the behaviors of the agents in new gridworlds across different degrees of decision complexity in similar ways to the way human observers do. We provide evidence for the alignment of the IBL observer's predictions under various levels of decision complexity. We also advance the demonstration of the IBL predictions using a classic test of false beliefs (the Sally–Anne test), which is commonly used to test ToM in humans. We discuss our results and the potential of the IBL observer model to improve human–machine interactions. (shrink)

Those who endorse the free energy principle as a theory of cognition are committed to three propositions that are jointly incompatible but which will cohere if one of them is denied. The first of these is that the free energy principle gives us a self-sufficient explanation of what all cognitive systems consist in: a specific computational architecture. The second is that all adaptive behavior is driven by the free energy principle and the process of model-based inference it (...) entails. The third is that cognition is not ubiquitous. These three incompatible propositions together comprise a problem of scope for the free energy principle as a theory of cognition. The prospects for rejecting each of these propositions are considered. To drop either the first or the second would limit the explanatory success of the principle. However, there are plausible ways to bite the bullet on denial of the third proposition. In particular, I argue that it is possible f... (shrink)

Over the past several decades, the philosophical community has witnessed the emergence of an important new paradigm for understanding the mind.1 The paradigm is that of machine computation, and its influence has been felt not only in philosophy, but also in all of the empirical disciplines devoted to the study of cognition. Of the several strategies for applying the resources provided by computer and cognitive science to the philosophy of mind, the one that has gained the most attention from (...) philosophers has been the Computational Theory of Mind (CTM). CTM was first articulated by Hilary Putnam (1960, 1961), but finds perhaps its most consistent and enduring advocate in Jerry Fodor (1975, 1980, 1981, 1987, 1990, 1994). It is this theory, and not any broader interpretations of what it would be for the mind to be a computer, that I wish to address in this paper. What I shall argue here is that the notion of symbolic representation employed by CTM is fundamentally unsuited to providing an explanation of the intentionality of mental states (a major goal of CTM), and that this result undercuts a second major goal of CTM, sometimes refered to as the vindication of intentional psychology. This line of argument is related to the discussions of derived intentionality by Searle (1980, 1983, 1984) and Sayre (1986, 1987). But whereas those discussions seem to be concerned with the causal dependence of familiar sorts of symbolic representation upon meaning-bestowing acts, my claim is rather that there is not one but several notions of meaning to be had, and that the notions that are applicable to symbols are conceptually dependent upon the notion that is applicable to mental states in the fashion that Aristotle refered to as paronymy. That is, an analysis of the notions of meaning applicable to symbols reveals that they contain presuppositions about meaningful mental states, much as Aristotle's analysis of the sense of healthy that is applied to foods reveals that it means conducive to having a healthy body, and hence any attempt to explain mental semantics in terms of the semantics of symbols is doomed to circularity and regress. I shall argue, however, that this does not have the consequence that computationalism is bankrupt as a paradigm for cognitive science, as it is possible to reconstruct CTM in a fashion that avoids these difficulties and makes it a viable research framework for psychology, albeit at the cost of losing its claims to explain intentionality and to vindicate intentional psychology. I have argued elsewhere (Horst, 1996) that local special sciences such as psychology do not require vindication in the form of demonstrating their reducibility to more fundamental theories, and hence failure to make good on these philosophical promises need not compromise the broad range of work in empirical cognitive science motivated by the computer paradigm in ways that do not depend on these problematic treatments of symbols. (shrink)

In this comment on Joshua Greene's essay, The Secret Joke of Kant's Soul, I argue that a notable weakness of Greene's approach to moral psychology is its neglect of computational theory. A central problem moral cognition must solve is to recognize (i.e., compute representations of) the deontic status of human acts and omissions. How do people actually do this? What is the theory which explains their practice?

In this carefully argued critique, Steven Horst pronounces the theory deficient.

The computational view of mind rests on certain intuitions regarding the fundamental similarity between computation and cognition. We examine some of these intuitions and suggest that they derive from the fact that computers and human organisms are both physical systems whose behavior is correctly described as being governed by rules acting on symbolic representations. Some of the implications of this view are discussed. It is suggested that a fundamental hypothesis of this approach is that there is a natural (...) domain of human functioning that can be addressed exclusively in terms of a formal symbolic or algorithmic vocabulary or level of analysis. Much of the paper elaborates various conditions that need to be met if a literal view of mental activity as computation is to serve as the basis for explanatory theories. The coherence of such a view depends on there being a principled distinction between functions whose explanation requires that we posit internal representations and those that we can appropriately describe as merely instantiating causal physical or biological laws. In this paper the distinction is empirically grounded in a methodological criterion called the " cognitive impenetrability condition." Functions are said to be cognitively impenetrable if they cannot be influenced by such purely cognitive factors as goals, beliefs, inferences, tacit knowledge, and so on. Such a criterion makes it possible to empirically separate the fixed capacities of mind from the particular representations and algorithms used on specific occasions. In order for computational theories to avoid being ad hoc, they must deal effectively with the "degrees of freedom" problem by constraining the extent to which they can be arbitrarily adjusted post hoc to fit some particular set of observations. This in turn requires that the fixed architectural function and the algorithms be independently validated. It is argued that the architectural assumptions implicit in many contemporary models run afoul of the cognitive impenetrability condition, since the required fixed functions are demonstrably sensitive to tacit knowledge and goals. The paper concludes with some tactical suggestions for the development of computational cognitive theories. (shrink)

One of the most lasting contributions of Marr's posthumous book is his articulation of the different “levels of analysis” that are needed to understand vision. Although a variety of work has examined how these different levels are related, there is comparatively little examination of the assumptions on which his proposed levels rest, or the plausibility of the approach Marr articulated given those assumptions. Marr placed particular significance on computational level theory, which specifies the “goal” of a computation, its (...) appropriateness for solving a particular problem, and the logic by which it can be carried out. The structure of computational level theory is inherently teleological: What the brain does is described in terms of its purpose. I argue that computational level theory, and the reverse-engineering approach it inspires, requires understanding the historical trajectory that gave rise to functional capacities that can be meaningfully attributed with some sense of purpose or goal, that is, a reconstruction of the fitness function on which natural selection acted in shaping our visual abilities. I argue that this reconstruction is required to distinguish abilities shaped by natural selection—“natural tasks” —from evolutionary “by-products”, rather than merely demonstrating that computational goals can be embedded in a Bayesian model that renders a particular behavior or process rational. (shrink)

The thesis examines the philosophical implications of the computational theory of early vision developed by Marr. According to Marr, early visual processes consist of sequences of "modular" computational mechanisms. These processes rely on functional relations between rates of change in stimulus magnitudes which result from certain contingent, global properties--natural constraints--of the physical world. ;Marr argues that explanations of early vision must have three distinct levels of description: computational, algorithmic and physical. In Chapter 1 I defend the (...) explanatory significance of this distinction in levels. In fulfilling its role in describing the dependence of visual processes on natural constraints, the computational level forms an autonomous level of description in the sense that it is unaffected by the computational steps at the other levels. ;In Chapter 2 I discuss the implications of natural constraints for the issues of individualism and methodological solipsism. I conclude that Mart's theory is nonindividualistic in the sense that visual content does not supervene on neural properties. However, this merely reflects the fact that different computational theories may be selected for the same system. Importantly, Marr's theory does not violate methodological solipsism since interpretations within theories must supervene on neural properties. ;In Chapter 3 I argue from the results of Chapters 1 and 2 that psychological explanation does not reduce to neurophysiology. This conclusion does not follow from the functionalist argument against physicalism, which is based on an incorrect account of computational theories. Rather the conclusion reflects the explanatory incompleteness of neurophysiological theories given the autonomous role of the computational level. ;In Chapter 4 I look in detail at the arguments for a "language of thought" as they apply to early vision. I distinguish two versions of the language of thought hypothesis, one weaker than the other. I conclude that the stronger version, which claims that a cognitive system is "program-using", is false of early vision because of the role of natural constraints. The weaker claim that cognitive processes employ symbolic transformations is true of the computational-level theory of early vision, but there is insufficient evidence to establish the claim at the algorithmic level. (shrink)

The authors present a neurological theory of how cognitive information and emotional information are integrated in the nucleus accumbens during effective decision making. They describe how the nucleus accumbens acts as a gateway to integrate cognitive information from the ventromedial prefrontal cortex and the hippocampus with emotional information from the amygdala. The authors have modeled this integration by a network of spiking artificial neurons organized into separate areas and used this computational model to simulate 2 kinds of cognitive–affective (...) integration. The model simulates successful performance by people with normal cognitive–affective integration. The model also simulates the historical case of Phineas Gage as well as subsequent patients whose ability to make decisions became.. (shrink)

Some computational phenomena rely essentially on pragmatic considerations, and seem to undermine the independence of the specification from the implementation. These include software development, deviant uses, esoteric languages and recent data-driven applications. To account for them, the interaction between pragmatics, epistemology and ontology in computational artefacts seems essential, indicating the need to recover the role of the language metaphor. We propose a User Levels (ULs) structure as a pragmatic complement to the Levels of Abstraction (LoAs)-based structure defining the (...) ontology and epistemology of computational artefacts. ULs identify a flexible hierarchy in which users bear their own semantic and normative requirements, possibly competing with the logical specification. We formulate a notion of computational act intended in its pragmatic sense, alongside pragmatic versions of implementation and correctness. (shrink)