In artificial intelligence, recent research has demonstrated the remarkable potential of Deep Convolutional Neural Networks (DCNNs), which seem to exceed state-of-the-art performance in new domains weekly, especially on the sorts of very difficult perceptual discrimination tasks that skeptics thought would remain beyond the reach of artificial intelligence. However, it has proven difficult to explain why DCNNs perform so well. In philosophy of mind, empiricists have long suggested that complex cognition is based on information derived from sensory experience, often appealing to (...) a faculty of abstraction. Rationalists have frequently complained, however, that empiricists never adequately explained how this faculty of abstraction actually works. In this paper, I tie these two questions together, to the mutual benefit of both disciplines. I argue that the architectural features that distinguish DCNNs from earlier neural networks allow them to implement a form of hierarchical processing that I call “transformational abstraction”. Transformational abstraction iteratively converts sensory-based representations of category exemplars into new formats that are increasingly tolerant to “nuisance variation” in input. Reflecting upon the way that DCNNs leverage a combination of linear and non-linear processing to efficiently accomplish this feat allows us to understand how the brain is capable of bi-directional travel between exemplars and abstractions, addressing longstanding problems in empiricist philosophy of mind. I end by considering the prospects for future research on DCNNs, arguing that rather than simply implementing 80s connectionism with more brute-force computation, transformational abstraction counts as a qualitatively distinct form of processing ripe with philosophical and psychological significance, because it is significantly better suited to depict the generic mechanism responsible for this important kind of psychological processing in the brain. (shrink)

This metatheoretical paper investigates mind wandering from the perspective of philosophy of mind. It has two central claims. The first is that, on a conceptual level, mind wandering can be fruitfully described as a specific form of mental autonomy loss. The second is that, given empirical constraints, most of what we call “conscious thought” is better analyzed as a subpersonal process that more often than not lacks crucial properties traditionally taken to be the hallmark of personal-level cognition - such as (...) mental agency, explicit, consciously experienced goal-directedness, or availability for veto control. I claim that for roughly two thirds of our conscious life-time we do not possess mental autonomy (M-autonomy) in this sense. Empirical data from research on mind wandering and nocturnal dreaming clearly show that phenomenally represented cognitive processing is mostly an automatic, non-agentive process and that personal-level cognition is an exception rather than the rule. This raises an interesting new version of the mind-body problem: How is subpersonal cognition causally related to personal-level thought? More fine-grained phenomenological descriptions for what we called “conscious thought” in the past are needed, as well as a functional decomposition of umbrella terms like “mind wandering” into different target phenomena and a better understanding of the frequent dynamic transitions between spontaneous, task-unrelated thought and meta-awareness. In an attempt to lay some very first conceptual foundations for the now burgeoning field of research on mind wandering, the third section proposes two new criteria for individuating single episodes of mind-wandering, namely, the “self-representational blink” (SRB) and a sudden shift in the phenomenological “unit of identification” (UI). I close by specifying a list of potentially innovative research goals that could serve to establish a stronger connection between mind wandering research and philosophy of mind. (shrink)

With their “bottom-up” approach, Holk Cruse and Malte Schilling present a highly intriguing perspective on those mental phenomena that have fascinated humankind since ancient times. Among them are those aspects of our inner lives that are at the same time most salient and yet most elusive: we are conscious beings with complex emotions, thinking and acting in pursuit of various goals. Starting with, from a biological point of view, very basic abilities, such as the ability to move and navigate in (...) an unpredictable environment, Cruse & Schilling have developed, step-by-step, a robotic system with the ability to plan future actions and, to a limited extent, to verbally report on its own internal states. The authors then offer a compelling argument that their system exhibits aspects of various higher-level mental phenomena such as emotion, attention, intention, volition, and even consciousness. The scientific investigation of the mind is faced with intricate problems at a very fundamental, methodological level. Not only is there a good deal of conceptual vagueness and uncertainty as to what the explananda precisely are, but it is also unclear what the best strategy might be for addressing the phenomena of interest. Cruse & Schilling’s bio-robotic “bottom-up” approach is designed to provide answers to such questions. In this commentary, I begin, in the first section, by presenting the main ideas behind this approach as I understand them. In the second section, I turn to an examination of its scope and limits. Specifically, I will suggest a set of constraints on good explanations based on the bottom-up approach. What criteria do such explanations have to meet in order to be of real scientific value? I maintain that there are essentially three such criteria: biological plausibility, adequate matching criteria, and transparency. Finally, in the third section, I offer directions for future research, as Cruse & Schilling’s bottom-up approach is well suited to provide new insights in the domain of social cognition and to explain its relation to phenomena such as language, emotion, and self. (shrink)

The target article sketched and explored a mechanism (action-oriented predictive processing) most plausibly associated with core forms of cortical processing. In assessing the attractions and pitfalls of the proposal we should keep that element distinct from larger, though interlocking, issues concerning the nature of adaptive organization in general.

The goal of this paper is to develop a systematic taxonomy of cognitive artifacts, i.e., human-made, physical objects that functionally contribute to performing a cognitive task. First, I identify the target domain by conceptualizing the category of cognitive artifacts as a functional kind: a kind of artifact that is defined purely by its function. Next, on the basis of their informational properties, I develop a set of related subcategories in which cognitive artifacts with similar properties can be grouped. In this (...) taxonomy, I distinguish between three taxa, those of family, genus, and species. The family includes all cognitive artifacts without further specifying their informational properties. Two genera are then distinguished: representational and non-representational (or ecological) cognitive artifacts. These genera are further divided into species. In case of representational artifacts, these species are iconic, indexical, or symbolic. In case of ecological artifacts, these species are spatial or structural. Within species, token artifacts are identified. The proposed taxonomy is an important first step towards a better understanding of the range and variety of cognitive artifacts and is a helpful point of departure, both for conceptualizing how different artifacts augment or impair cognitive performance and how they transform and are integrated into our cognitive system and practices. (shrink)

I argue that examining the explanatory power of the hypothesis of extended cognition (HEC) offers a fruitful approach to the problem of cognitive system demarcation. Although in the discussions on HEC it has become common to refer to considerations of explanatory power as a means for assessing the plausibility of the extended cognition approach, to date no satisfying account of explanatory power has been presented in the literature. I suggest that the currently most prominent theory of explanation in the special (...) sciences, James Woodward’s contrastive-counterfactual theory, and an account of explanatory virtues building on that theory can be used to develop a systematic picture of cognitive system demarcation in the psychological sciences. A major difference between my differential influence (DI) account and most other theories of cognitive extension is the cognitive systems pluralism implied by my approach. By examining the explanatory power of competing traditions in psychological memory research, I conclude that internalist and externalist classificatory strategies are characterized by different profiles of explanatory virtues and should often be considered as complementary rather than competing approaches. This suggests a deflationary interpretation of HEC. (shrink)

This paper introduces a new, expanded range of relevant cognitive psychological research on collaborative recall and social memory to the philosophical debate on extended and distributed cognition. We start by examining the case for extended cognition based on the complementarity of inner and outer resources, by which neural, bodily, social, and environmental resources with disparate but complementary properties are integrated into hybrid cognitive systems, transforming or augmenting the nature of remembering or decision-making. Adams and Aizawa, noting this distinctive complementarity argument, (...) say that they agree with it completely: but they describe it as “a non-revolutionary approach” which leaves “the cognitive psychology of memory as the study of processes that take place, essentially without exception, within nervous systems.” In response, we carve out, on distinct conceptual and empirical grounds, a rich middle ground between internalist forms of cognitivism and radical anti-cognitivism. Drawing both on extended cognition literature and on Sterelny’s account of the “scaffolded mind” (this issue), we develop a multidimensional framework for understanding varying relations between agents and external resources, both technological and social. On this basis we argue that, independent of any more “revolutionary” metaphysical claims about the partial constitution of cognitive processes by external resources, a thesis of scaffolded or distributed cognition can substantially influence or transform explanatory practice in cognitive science. Critics also cite various empirical results as evidence against the idea that remembering can extend beyond skull and skin. We respond with a more principled, representative survey of the scientific psychology of memory, focussing in particular on robust recent empirical traditions for the study of collaborative recall and transactive social memory. We describe our own empirical research on socially distributed remembering, aimed at identifying conditions for mnemonic emergence in collaborative groups. Philosophical debates about extended, embedded, and distributed cognition can thus make richer, mutually beneficial contact with independently motivated research programs in the cognitive psychology of memory. (shrink)

Prominent defenders of the extended cognition thesis have looked to evolutionary theory for support. Roughly, the idea is that natural selection leads one to expect that cognitive strategies should exploit the environment, and exploitation of the right sort results in a cognitive system that extends beyond the head of the organism. I argue that proper appreciation of evolutionary theory should create no such expectation. This leaves open whether cognitive systems might in fact bear a relationship to the environment that leads (...) to their extension. *Received July 2009; revised January 2010. †To contact the author, please write to: Department of Philosophy, 5185 Helen C. White Hall, 600 North Park Street, University of Wisconsin, Madison, WI 53706; e‐mail: [email protected]. (shrink)

Vision constitutes an interesting domain, or range of domains, for debate over the extended mind thesis, the idea that minds physically extend beyond the boundaries of the body. In part this is because vision and visual experience more particularly are sometimes presented as a kind of line in the sand for what we might call externalist creep about the mind: once all reasonable concessions have been made to externalists about the mind, visual experience marks a line beyond which lies a (...) safe haven for individualists. Here I want to put a little more pressure on such a view of visual experience, as well as to offer a more constructive, positive argument in defense of the idea of extended vision. (shrink)

When cognitive scientists apply computational theory to the problem of phenomenal consciousness, as many of them have been doing recently, there are two fundamentally distinct approaches available. Either consciousness is to be explained in terms of the nature of the representational vehicles the brain deploys; or it is to be explained in terms of the computational processes defined over these vehicles. We call versions of these two approaches _vehicle_ and _process_ theories of consciousness, respectively. However, while there may be space (...) for vehicle theories of consciousness in cognitive science, they are relatively rare. This is because of the influence exerted, on the one hand, by a large body of research which purports to show that the explicit representation of information in the brain and conscious experience are _dissociable_, and on the other, by the _classical_ computational theory of mind – the theory that takes human cognition to be a species of symbol manipulation. But two recent developments in cognitive science combine to suggest that a reappraisal of this situation is in order. First, a number of theorists have recently been highly critical of the experimental methodologies employed in the dissociation studies – so critical, in fact, it’s no longer reasonable to assume that the dissociability of conscious experience and explicit representation has been adequately demonstrated. Second, classicism, as a theory of human cognition, is no longer as dominant in cognitive science as it once was. It now has a lively competitor in the form of _connectionism; _and connectionism, unlike classicism, does have the computational resources to support a robust vehicle theory of consciousness. In this paper we develop and defend this connectionist vehicle theory of consciousness. It takes the form of the following simple empirical hypothesis: _phenomenal experience consists in the explicit_ _representation of information in neurally realized PDP networks_.. (shrink)

How does language (spoken or written) impact thought? One useful way to approach this important but elusive question may be to consider language itself as a cognition-enhancing animal-built structure. To take this perspective is to view language as a kind of self-constructed cognitive niche. These self-constructed cognitive niches play, I suggest, three distinct but deeply interlocking roles in human thought and reason. Working together, these three interlocking routines radically transform the human mind, and mark a genuine discontinuity in the space (...) of anitnal minds. (shrink)

1. The Situation in Cognition 2. Situated Cognition: A Potted Recent History 3. Extensions in Biology, Computation, and Cognition 4. Articulating the Idea of Cognitive Extension 5. Are Some Resources Intrinsically Non-Cognitive? 6. Is Cognition Extended or Only Embedded? 7. Letting Nature Take Its Course.

A traditional view of perception and action makestwo assumptions: that the causal flow betweenperception and action is primarily linear or one-way,and that they are merely instrumentally related toeach other, so that each is a means to the other.Either or both of these assumptions can be rejected. Behaviorism rejects the instrumental but not theone-way aspect of the traditional view, thus leavingitself open to charges of verificationism. Ecologicalviews reject the one-way aspect but not theinstrumental aspect of the traditional view, so thatperception and (...) action are seen as instrumentallyinterdependent. It is argued here that a betteralternative is to reject both assumptions, resultingin a two-level interdependence view in whichperception and action co-depend on dynamicallycircular subpersonal relations and as a result may bemore than merely instrumentally interdependent. Thisis illustrated by reference to motor theories ofperception and control theories of action. (shrink)

Many philosophers think that being in an intentional state is a matter of being related to a sentence in a mental language-a 'Language of Thought' (see especially Fodor 1975, 1987 Appendix; Field 1978). According to this view-which I shall call 'the LT hypothesis'-when anyone has a belief or a desire or a hope with a certain content, they have a sentence of this language, with that content, 'written' in their heads. The claim is meant quite literally: the mental representations that (...) make up the items of this language have semantic and syntactic properties. This is why, according to this view, cognitive psychology does, and should, treat that part of the mind which deals with intentional states as a semantic and a syntactic 'engine'. (shrink)

In this paper, I critically assess the enactive account of visual perception recently defended by Alva Noë (2004). I argue inter alia that the enactive account falsely identifies an object’s apparent shape with its 2D perspectival shape; that it mistakenly assimilates visual shape perception and volumetric object recognition; and that it seriously misrepresents the constitutive role of bodily action in visual awareness. I argue further that noticing an object’s perspectival shape involves a hybrid experience combining both perceptual and imaginative elements—an (...) act of what I call ‘make‐perceive’. (shrink)

In approaching the question of whether groups of people can have cognitive capacities that are fundamentally different than the cognitive capacities of the individuals within the group, we lay out a Multiple, Interactive Levels of Cognitive Systems (MILCS) framework. The goal of MILCS is to explain the kinds of cognitive processes typically studied by cognitive scientists, such as perception, attention, memory, categorization, decision making, problem solving, and judgment. Rather than focusing on high-level constructs such as modules in an information processing (...) flow diagram or internal representations, MILCS focuses on mechanisms that allow systems to engage in flexible, adaptive behavior. Examples of these mechanisms are network structure-process pairs that: combine data from perception with top-down theories, achieve consistency and consensus through information exchange, selectively attend to information globally assessed as relevant, and develop specialized units from originally homogeneous networks through competition among the units. Two such systems are considered in some detail -- lateral inhibition within a network for selecting from a candidate set the option that is most attractive, and a diffusion process for accumulating evidence to reach a generally rapid and accurate decision. These system descriptions are aptly applied at multiple levels, including within and across people. These systems provide accounts that unify cognitive processes across multiple levels, can be expressed in a common vocabulary provided by network science, are inductively powerful yet appropriately constrained, and are applicable to a large number of superficially diverse cognitive systems. Given that people are typically highly motivated to participate in strong groups, cognitively resourceful people will tend to form groups that effectively employ cognitive systems at higher levels than the individuals. The impressive cognitive capacities of individual people do not eliminate the need to talk about group cognition. Instead, smart people generally form smart groups. (shrink)

Dead reckoning is a feature of the navigation behaviour shown by several creatures, including the desert ant. Recent work by C. Randy Gallistel shows that some connectionist models of dead reckoning face important challenges. These challenges are thought to arise from essential features of the connectionist approach, and have therefore been taken to show that connectionist models are unable to explain even the most primitive of psychological phenomena. I show that Gallistel’s challenges are successfully met by one recent connectionist model, (...) proposed by Ulysses Bernardet, Sergi Bermúdez i Badia, and Paul F.M.J. Verschure. The success of this model suggests that there are ways to implement dead reckoning with neural circuits that fall within the bounds of what many people regard as neurobiologically plausible, and so that the wholesale dismissal of the connectionist modelling project remains premature. (shrink)

In a recent contribution to this journal, Andrew Fenton and Sheri Alpert have argued that the so-called “extended mind hypothesis” allows us to understand why Brain Computer Interfaces (BCIs) have the potential to change the self of patients suffering from Locked-in syndrome (LIS) by extending their minds beyond their bodies. I deny that this can shed any light on the theoretical, or philosophical, underpinnings of BCIs as a tool for enabling communication with, or bodily action by, patients with LIS: BCIs (...) are not a case of cognitive extension. I argue that Fenton and Alpert’s claim to the contrary is the result of a widespread confusion about some related, but significantly different, approaches to cognition that all fall under the heading of “situated cognition.” I first provide a short taxonomy of various situated approaches to cognition, highlighting (some of) their important commonalities and differences, which should dissolve some of the confusions surrounding them. Then I show why the extended mind hypothesis is unsuitable as a model of BCI enhancements of LIS patients’ capacity to interact with their surroundings, and I argue that the situated approach with obvious bearings on the sort of questions that were driving Fenton and Alpert is not the idea that cognition is extended , but the idea that cognition is enacted. (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?

A suitable project for the new Millenium is to radically reconfigure our image of human rationality. Such a project is already underway, within the Cognitive Sciences, under the umbrellas of work in Situated Cognition, Distributed and De-centralized Cogition, Real-world Robotics and Artificial Life1. Such approaches, however, are often criticized for giving certain aspects of rationality too wide a berth. They focus their attention on on such superficially poor cousins as.

Before one can even begin to model consciousness and what exactly it means that it is a subjective phenomenon one needs a theory about what a first-person perspective really is. This theory has to be conceptually convincing, empirically plausible and, most of all, open to new developments. The chosen conceptual framework must be able to accommodate scientific progress. Its ba- sic assumptions have to be plastic as it were, so that new details and empirical data can continuously be fed into (...) the theoretical model as it grows and becomes more refined. This paper makes an attempt at sketching the outlines of such a theory, offering a representationalist analysis of the phenomenal first-person perspective. Three phenomenal target properties are centrally relevant. (shrink)

Recent work in the methodology of connectionist explanation has I'ocrrsccl on the notion of levels of explanation. Specific issucs in conncctionisrn hcrc intersect with rvider areas of debate in the philosophy of psychology and thc philosophy of science generally. The issues I raise in this chapter, then, are not unique to cognitive science; but they arise in new and important contexts when connectionism is taken seriously as a model of cognition. The general questions are the relation between levels and the (...) status of levels which have no obvious relation to others. In speaking of levels, what is the connection, if there is one, between explanation and ontology? Which, if any, conccpt of reduction is applicable to connectionist systems? What kind of legitinrtcy can the constructs of common sense psychology, or of that vclsion ol intentional realism represented by classical symbol-systems n I, hirvc irr ir full-scale connectionist theory of mind? (shrink)

Does the existence of body representations undermine the explanatory role of the body? Or do certain types of representation depend so closely upon the body that their involvement in a cognitive task implicates the body itself? In the introduction of this special issue we explore lines of tension and complement that might hold between the notions of embodiment and body representations, which remain too often neglected or obscure. To do so, we distinguish two conceptions of embodiment that either put weight (...) on the explanatory role of the body itself or body representations. We further analyse how and to what extent body representations can be said to be embodied. Finally, we give an overview of the full volume articulated around foundational issues (How should we define the notion of embodiment? To what extent and in what sense is embodiment compatible with representationalism? To what extent and in what sense are sensorimotor approaches similar to behaviourism?) and their applications in several cognitive domains (perception, concepts, selfhood, social cognition). (shrink)

This book contains a representationalist theory of self-consciousness and of the phenomenal first-person perspective. It draws on empirical data from the cognitive and neurosciences.

Mathematicians’ use of external representations, such as symbols and diagrams, constitutes an important focal point in current philosophical attempts to understand mathematical practice. In this paper, we add to this understanding by presenting and analyzing how research mathematicians use and interact with external representations. The empirical basis of the article consists of a qualitative interview study we conducted with active research mathematicians. In our analysis of the empirical material, we primarily used the empirically based frameworks provided by distributed cognition and (...) cognitive semantics as well as the broader theory of cognitive integration as an analytical lens. We conclude that research mathematicians engage in generative feedback loops with material representations, that they use representations to facilitate the use of experiences of handling the physical world as a resource in mathematical work, and that their use of representations is socially sanctioned and enabled. These results verify the validity of the cognitive frameworks used as the basis for our analysis, but also show the need for augmentation and revision. Especially, we conclude that the social and cultural context cannot be excluded from cognitive analysis of mathematicians’ use of external representations. Rather, representations are socially sanctioned and enabled in an enculturation process. (shrink)

In this article, I develop an account of the use of intentional predicates in cognitive neuroscience explanations. As pointed out by Maxwell Bennett and Peter Hacker, intentional language abounds in neuroscience theories. According to Bennett and Hacker, the subpersonal use of intentional predicates results in conceptual confusion. I argue against this overly strong conclusion by evaluating the contested language use in light of its explanatory function. By employing conceptual resources from the contemporary philosophy of science, I show that although the (...) use of intentional predicates in mechanistic explanations sometimes leads to explanatorily inert claims, intentional predicates can also successfully feature in mechanistic explanations as tools for the functional analysis of the explanandum phenomenon. Despite the similarities between my account and Daniel Dennett's intentional-stance approach, I argue that intentional stance should not be understood as a theory of subpersonal causal explanation, and therefore cannot be used to assess the explanatory role of intentional predicates in neuroscience. Finally, I outline a general strategy for answering the question of what kind of language can be employed in mechanistic explanations. (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)