Good reasoning can lead to success; bad reasoning can lead to catastrophe. Yet, it's not obvious how we reason, and why we make mistakes. This new book by one of the pioneers of the field, Philip Johnson-Laird, looks at the mental processes that underlie our reasoning. It provides the most accessible account yet of the science of reasoning.

This paper concerns the role of the structural properties of representations in determining the nature of their content. I take as a starting point Fodor's (2007) and Heck's (2007) recent arguments making the iconic structure of perceptual representations essential in establishing their content as content of a different (nonconceptual) kind. I argue that the prima facie state–content error this strategy seems to display is nothing but a case of “state–content error error,” i.e., the mistake of considering that the properties that (...) characterize the type of content certain representations have are indeed independent of their structural properties. I also consider another objection to the general strategy, namely that it falls short of showing that the nonconceptual content of perceptual experiences thus established is the kind of content that figures in true explanations of intentional behavior. I concede this point in respect of Fodor's version of the strategy, while denying that it has any bearing on Heck's approach. The success of this objection in Heck's case ultimately depends, I argue, upon an unjustified commitment to a certain empirically suspect model of perception. (shrink)

Research on human infants has begun to shed light on early-developing processes for segmenting perceptual arrays into objects. Infants appear to perceive objects by analyzing three-dimensional surface arrangements and motions. Their perception does not accord with a general tendency to maximize figural goodness or to attend to nonaccidental geometric relations in visual arrays. Object perception does accord with principles governing the motions of material bodies: Infants divide perceptual arrays into units that move as connected wholes, that move separately from one (...) another, that tend to maintain their size and shape over motion, and that tend to act upon each other only on contact. These findings suggest that o general representation of object unity and boundaries is interposed between representations of surfaces and representations of objects of familiar kinds. The processes that construct this representation may be related to processes of physical reasoning. (shrink)

Chomsky proposes a reformulation of the theory of transformational generative grammar that takes recent developments in the descriptive analysis of particular ...

Fodor advocates a view of cognitive processes as computations defined over the language of thought (or Mentalese). Even among those who endorse Mentalese, considerable controversy surrounds its representational format. What semantically relevant structure should scientific psychology attribute to Mentalese symbols? Researchers commonly emphasize logical structure, akin to that displayed by predicate calculus sentences. To counteract this tendency, I discuss computational models of navigation drawn from probabilistic robotics. These models involve computations defined over cognitive maps, which have geometric rather than logical (...) structure. They thereby demonstrate the possibility of rational cognitive processes in an exclusively non-logical representational medium. Furthermore, they offer much promise for the empirical study of animal navigation. (shrink)

Only human beings have a rich conceptual repertoire with concepts like tort, entropy, Abelian group, mannerism, icon and deconstruction. How have humans constructed these concepts? And once they have been constructed by adults, how do children acquire them? While primarily focusing on the second question, in The Origin of Concepts , Susan Carey shows that the answers to both overlap substantially. Carey begins by characterizing the innate starting point for conceptual development, namely systems of core cognition. Representations of core cognition (...) are the output of dedicated input analyzers, as with perceptual representations, but these core representations differ from perceptual representations in having more abstract contents and richer functional roles. Carey argues that the key to understanding cognitive development lies in recognizing conceptual discontinuities in which new representational systems emerge that have more expressive power than core cognition and are also incommensurate with core cognition and other earlier representational systems. Finally, Carey fleshes out Quinian bootstrapping, a learning mechanism that has been repeatedly sketched in the literature on the history and philosophy of science. She demonstrates that Quinian bootstrapping is a major mechanism in the construction of new representational resources over the course of childrens cognitive development. Carey shows how developmental cognitive science resolves aspects of long-standing philosophical debates about the existence, nature, content, and format of innate knowledge. She also shows that understanding the processes of conceptual development in children illuminates the historical process by which concepts are constructed, and transforms the way we think about philosophical problems about the nature of concepts and the relations between language and thought. (shrink)

Over the last quarter century, the dominant tendency in comparative cognitive psychology has been to emphasize the similarities between human and nonhuman minds and to downplay the differences as (Darwin 1871). In the present target article, we argue that Darwin was mistaken: the profound biological continuity between human and nonhuman animals masks an equally profound discontinuity between human and nonhuman minds. To wit, there is a significant discontinuity in the degree to which human and nonhuman animals are able to approximate (...) the higher-order, systematic, relational capabilities of a physical symbol system (PSS) (Newell 1980). We show that this symbolic-relational discontinuity pervades nearly every domain of cognition and runs much deeper than even the spectacular scaffolding provided by language or culture alone can explain. We propose a representational-level specification as to where human and nonhuman animals' abilities to approximate a PSS are similar and where they differ. We conclude by suggesting that recent symbolic-connectionist models of cognition shed new light on the mechanisms that underlie the gap between human and nonhuman minds. (shrink)

Most of us create and use a panoply of non-sentential representations throughout our ordinary lives: we regularly use maps to navigate, charts to keep track of complex patterns of data, and diagrams to visualize logical and causal relations among states of affairs. But philosophers typically pay little attention to such representations, focusing almost exclusively on language instead. In particular, when theorizing about the mind, many philosophers assume that there is a very tight mapping between language and thought. Some analyze utterances (...) as the outer vocalizations of inner thoughts (e.g. Grice 1957, Devitt 2005), while others treat thought as a form of inner speech (e.g. Sellars 1956/1997, Carruthers 2002). But even philosophers who take no stand on the relative priority of language and thought still tend to individuate mental states in terms of the sentences we use to ascribe them. Indeed, Dummett (1993) claims that it is constitutive of analytic philosophy that it approaches the mind by way of language. In many ways, this linguistic model is salutary. Our thoughts are often intimately intertwined with their linguistic expression, and public language does provide a comparatively tractable proxy for, and a window into, the messier realm of thought. However, an exclusive focus on thought as it is expressed in language threatens to leave other sorts of thought unexplained, or even to blind us to their possibility. In particular, many cognitive ethologists and psychologists find it useful to talk about humans, chimpanzees, birds, rats, and even bees as employing cognitive maps. We need to make sense of this way of talking about minds as well as more familiar sentential descriptions. In what follows, I investigate the theoretical and practical possibility of non-sentential thought. Ultimately, I am most interested in the contours of distinctively human thought: what forms does human thought take, and how do those different forms interact? How does human thought compare with that of other animals? In this essay, however, I focus on a narrower and more basic theoretical question: could thought occur in maps? Many philosophers are convinced that in some important sense, thought per se must be language-like.. (shrink)

What makes an object the same persisting individual over time? Philosophers and psychologists have both grappled with this question, but from different perspectives—philosophers conceptually analyzing the criteria for object persistence, and psychologists exploring the mental mechanisms that lead us to experience the world in terms of persisting objects. It is striking that the same themes populate explorations of persistence in these two very different fields—e.g. the roles of spatiotemporal continuity, persistence through property change, and cohesion violations. Such similarities may reflect (...) an underlying connection, in that psychological mechanisms of object persistence (especially relevant parts of mid-level visual object processing) may serve to underlie the intuitions about persistence that fuel metaphysical theories. This would be a way for cognitive science to join these two disparate fields, helping to explain the possible origins and reliability of some metaphysical intuitions, and perhaps leading to philosophical progress. (shrink)

I argue that we can reconcile two seemingly incompatible traditions for thinking about concepts. On the one hand, many cognitive scientists assume that the systematic redeployment of representational abilities suffices for having concepts. On the other hand, a long philosophical tradition maintains that language is necessary for genuinely conceptual thought. I argue that on a theoretically useful and empirically plausible concept of 'concept', it is necessary and sufficient for conceptual thought that a thinker be able to entertain many of the (...) potential thoughts produced by recombining her representational abilities apart from a direct confrontation with the states of affairs being represented. Such representational abilities support a cognitive engagement with the world that is flexible, abstract, and active. (shrink)

The idea that human thinking and machine computing are "radically the same" provides the central theme for this marvelously lucid and witty book on...

This paper explores the difference between Connectionist proposals for cognitive a r c h i t e c t u r e a n d t h e s o r t s o f m o d e l s t hat have traditionally been assum e d i n c o g n i t i v e s c i e n c e . W e c l a i m t h a t t h (...) e m a j o r d i s t i n c t i o n i s t h a t , w h i l e b o t h Connectionist and Classical architectures postulate representational mental states, the latter but not the former are committed to a symbol-level of representation, or to a ‘language of thought’: i.e., to representational states that have combinatorial syntactic and semantic structure. Several arguments for combinatorial structure in mental representations are then reviewed. These include arguments based on the ‘systematicity’ of mental representation: i.e., on the fact that cognitive capacities always exhibit certain symmetries, so that the ability to entertain a given thought implies the ability to entertain thoughts with semantically related contents. We claim that such arguments make a powerful case that mind/brain architecture is not Connectionist at the cognitive level. We then consider the possibility that Connectionism may provide an account of the neural (or ‘abstract neurological’) structures in which Classical cognitive architecture is implemented. We survey a n u m b e r o f t h e s t a n d a r d a r g u m e n t s t h a t h a v e b e e n o f f e r e d i n f a v o r o f Connectionism, and conclude that they are coherent only on this interpretation. (shrink)

The book also introduces a host of new experimental techniques and major hypotheses to guide future research.

It is generally accepted that there is something special about reasoning by using mental images. The question of how it is special, however, has never been satisfactorily spelled out, despite more than thirty years of research in the post-behaviorist tradition. This article considers some of the general motivation for the assumption that entertaining mental images involves inspecting a picture-like object. It sets out a distinction between phenomena attributable to the nature of mind to what is called the cognitive architecture, and (...) ones that are attributable to tacit knowledge used to simulate what would happen in a visual situation. With this distinction in mind, the paper then considers in detail the widely held assumption that in some important sense images are spatially displayed or are depictive, and that examining images uses the same mechanisms that are deployed in visual perception. I argue that the assumption of the spatial or depictive nature of images is only explanatory if taken literally, as a claim about how images are physically instantiated in the brain, and that the literal view fails for a number of empirical reasons – for example, because of the cognitive penetrability of the phenomena cited in its favor. Similarly, while it is arguably the case that imagery and vision involve some of the same mechanisms, this tells us very little about the nature of mental imagery and does not support claims about the pictorial nature of mental images. Finally, I consider whether recent neuroscience evidence clarifies the debate over the nature of mental images. I claim that when such questions as whether images are depictive or spatial are formulated more clearly, the evidence does not provide support for the picture-theory over a symbol-structure theory of mental imagery. Even if all the empirical claims were true, they do not warrant the conclusion that many people have drawn from them: that mental images are depictive or are displayed in some (possibly cortical) space. Such a conclusion is incompatible with what is known about how images function in thought. We are then left with the provisional counterintuitive conclusion that the available evidence does not support rejection of what I call the “null hypothesis”; namely, that reasoning with mental images involves the same form of representation and the same processes as that of reasoning in general, except that the content or subject matter of thoughts experienced as images includes information about how things would look. (shrink)

Preface 1 Introduction: The Persistence of the Attitudes 2 Individualism and Supervenience 3 Meaning Holism 4 Meaning and the World Order Epilogue Creation Myth Appendix Why There Still Has to be a Language of Thought Notes References Author Index.

The renowned philosopher Jerry Fodor, a leading figure in the study of the mind for more than twenty years, presents a strikingly original theory on the basic constituents of thought. He suggests that the heart of cognitive science is its theory of concepts, and that cognitive scientists have gone badly wrong in many areas because their assumptions about concepts have been mistaken. Fodor argues compellingly for an atomistic theory of concepts, deals out witty and pugnacious demolitions of rival theories, and (...) suggests that future work on human cognition should build upon new foundations. This lively, conversational, and superbly accessible book is the first volume in the Oxford Cognitive Science Series, where the best original work in this field will be presented to a broad readership. Concepts will fascinate anyone interested in contemporary work on mind and language. Cognitive science will never be the same again. (shrink)

Prior to the twentieth century, theories of knowledge were inherently perceptual. Since then, developments in logic, statis- tics, and programming languages have inspired amodal theories that rest on principles fundamentally different from those underlying perception. In addition, perceptual approaches have become widely viewed as untenable because they are assumed to implement record- ing systems, not conceptual systems. A perceptual theory of knowledge is developed here in the context of current cognitive science and neuroscience. During perceptual experience, association areas in the (...) brain capture bottom-up patterns of activation in sensory-motor areas. Later, in a top-down manner, association areas partially reactivate sensory-motor areas to implement perceptual symbols. The stor- age and reactivation of perceptual symbols operates at the level of perceptual components – not at the level of holistic perceptual expe- riences. Through the use of selective attention, schematic representations of perceptual components are extracted from experience and stored in memory (e.g., individual memories of green, purr, hot). As memories of the same component become organized around a com- mon frame, they implement a simulator that produces limitless simulations of the component (e.g., simulations of purr). Not only do such simulators develop for aspects of sensory experience, they also develop for aspects of proprioception (e.g., lift, run) and introspec- tion (e.g., compare, memory, happy, hungry). Once established, these simulators implement a basic conceptual system that represents types, supports categorization, and produces categorical inferences. These simulators further support productivity, propositions, and ab- stract concepts, thereby implementing a fully functional conceptual system. Productivity results from integrating simulators combinato- rially and recursively to produce complex simulations. Propositions result from binding simulators to perceived individuals to represent type-token relations. Abstract concepts are grounded in complex simulations of combined physical and introspective events. Thus, a per- ceptual theory of knowledge can implement a fully functional conceptual system while avoiding problems associated with amodal sym- bol systems. Implications for cognition, neuroscience, evolution, development, and artificial intelligence are explored. (shrink)

Thoughts and other mental states are defined by their role in a functional system. Since it is easier to determine when we have knowledge than when reasoning has occurred, Gilbert Harman attempts to answer the latter question by seeing what assumptions about reasoning would best account for when we have knowledge and when not. He describes induction as inference to the best explanation, or more precisely as a modification of beliefs that seeks to minimize change and maximize explanatory coherence. Originally (...) published in 1973. The Princeton Legacy Library uses the latest print-on-demand technology to again make available previously out-of-print books from the distinguished backlist of Princeton University Press. These editions preserve the original texts of these important books while presenting them in durable paperback and hardcover editions. The goal of the Princeton Legacy Library is to vastly increase access to the rich scholarly heritage found in the thousands of books published by Princeton University Press since its founding in 1905. (shrink)

In crossmodal identification, a subject token identifies an item perceived in one sensory modality with an item perceived in another sensory modality. Does crossmodal identification always occur in cognition, or does crossmodal identification sometimes take place in perception? This paper argues that crossmodal identification occurs in cognition, and not in perception. Nevertheless, multisensory perception is not unalive to crossmodal identity. Experimental evidence demonstrates that perception is differentially sensitive to the identity of individuals presented to distinct senses. Such sensitivity enhances recognition (...) and improves action. This approach relies on distinguishing crossmodal identification from perceiving crossmodal identity. Perception registers crossmodal identity, but crossmodal identification as such belongs to thought. (shrink)

The ability to perceive the shapes of things as enduring through changes in how they stimulate our sense organs is vital to our sense of stability in the world. But what sort of capacity is shape constancy, and how is it reflected in perceptual experience? This paper defends a pluralist account of shape constancy: There are multiple kinds of shape constancy centered on geometrical properties at various levels of abstraction, and properties at these various levels feature in the content of (...) perceptual experience, governing patterns of apparent shape similarity. I propose that the varieties of shape constancy are subserved by the syntactic complexity of perceptual shape representations. By assigning discrete constituents to various abstract shape parameters, these representations attune us to the preservation of certain abstract shape properties through changes in more determinate shape properties. Finally, I draw broader lessons concerning the nature and function of perceptual constancy. (shrink)

A basic deep neural network (DNN) is trained to exhibit a large set of input–output dispositions. While being a good model of the way humans perform some tasks automatically, without deliberative reasoning, more is needed to approach human‐like artificial intelligence. Analysing recent additions brings to light a distinction between two fundamentally different styles of computation: content‐specific and non‐content‐specific computation (as first defined here). For example, deep episodic RL networks draw on both. So does human conceptual reasoning. Combining the two takes (...) advantage of the complementary costs and benefits of each. It also offers a better model of human cognitive competence. (shrink)

Deep neural networks (DNNs) have had extraordinary successes in classifying photographic images of objects and are often described as the best models of biological vision. This conclusion is largely based on three sets of findings: (1) DNNs are more accurate than any other model in classifying images taken from various datasets, (2) DNNs do the best job in predicting the pattern of human errors in classifying objects taken from various behavioral datasets, and (3) DNNs do the best job in predicting (...) brain signals in response to images taken from various brain datasets (e.g., single cell responses or fMRI data). However, these behavioral and brain datasets do not test hypotheses regarding what features are contributing to good predictions and we show that the predictions may be mediated by DNNs that share little overlap with biological vision. More problematically, we show that DNNs account for almost no results from psychological research. This contradicts the common claim that DNNs are good, let alone the best, models of human object recognition. We argue that theorists interested in developing biologically plausible models of human vision need to direct their attention to explaining psychological findings. More generally, theorists need to build models that explain the results of experiments that manipulate independent variables designed to test hypotheses rather than compete on making the best predictions. We conclude by briefly summarizing various promising modeling approaches that focus on psychological data. (shrink)

Perceptual representations pick out individuals and attribute properties to them. This paper considers the role of perceptual attribution in determining or guiding perceptual reference to objects. We consider three extant models of the relation between perceptual attribution and perceptual reference–all attribution guides reference, no attribution guides reference, or a privileged subset of attributions guides reference–and argue that empirical evidence undermines all three. We then defend a flexible-attributives model, on which the range of perceptual attributives used to guide reference shifts adaptively (...) with context. This model underscores the remarkable and dynamic intelligence of our perceptual capacities. We elucidate implications of the model for the boundary between perception and propositional thought. (shrink)

Paul M. Pietroski presents an ambitious new account of human languages as generative procedures that respect substantive constraints. He argues that meanings are neither concepts nor extensions, and sentences do not have truth conditions; meanings are composable instructions for how to access and assemble concepts of a special sort.

In this paper, I make a case for the modularity of the motor system. I start where many do in discussions of modularity, by considering the extent to which the motor system is cognitively penetrable, i.e., the extent to which its processing and outputs are causally influenced, in a semantically coherent way, by states of central cognition. I present some empirical findings from a range of sensorimotor adaptation studies that strongly suggest that there are limits to such influence under certain (...) conditions. These results cry out for an explanation. In the remainder of the paper, I provide one: The motor system is cognitively penetrable, but nonetheless modular along broadly Fodorian lines, insofar as it is informationally encapsulated. This means that its access is limited to its own proprietary database in computing its function from input to output, which does not include the information stored in central cognition. I then offer a model of action control, from distal intention to action outcomes, that further helps to illustrate this picture and can accommodate the target empirical findings. (shrink)

On a now orthodox view, humans and many other animals possess a “number sense,” or approximate number system, that represents number. Recently, this orthodox view has been subject to numerous critiques that question whether the ANS genuinely represents number. We distinguish three lines of critique – the arguments from congruency, confounds, and imprecision – and show that none succeed. We then provide positive reasons to think that the ANS genuinely represents numbers, and not just non-numerical confounds or exotic substitutes for (...) number, such as “numerosities” or “quanticals,” as critics propose. In so doing, we raise a neglected question: numbers of what kind? Proponents of the orthodox view have been remarkably coy on this issue. But this is unsatisfactory since the predictions of the orthodox view, including the situations in which the ANS is expected to succeed or fail, turn on the kind of number being represented. In response, we propose that the ANS represents not only natural numbers, but also non-natural rational numbers. It does not represent irrational numbers, however, and thereby fails to represent the real numbers more generally. This distances our proposal from existing conjectures, refines our understanding of the ANS, and paves the way for future research. (shrink)

Most theories of concepts take concepts to be structured bodies of information used in categorization and inference. This paper argues for a version of atomism, on which concepts are unstructured symbols. However, traditional Fodorian atomism is falsified by polysemy and fails to provide an account of how concepts figure in cognition. This paper argues that concepts are generative pointers, that is, unstructured symbols that point to memory locations where cognitively useful bodies of information are stored and can be deployed to (...) resolve polysemy. The notion of generative pointers allows for unresolved ambiguity in thought and provides a basis for conceptual engineering. (shrink)

Artificial Intelligence: A Modern Approach, 3e offers the most comprehensive, up-to-date introduction to the theory and practice of artificial intelligence. Number one in its field, this textbook is ideal for one or two-semester, undergraduate or graduate-level courses in Artificial Intelligence. Dr. Peter Norvig, contributing Artificial Intelligence author and Professor Sebastian Thrun, a Pearson author are offering a free online course at Stanford University on artificial intelligence. According to an article in The New York Times, the course on artificial intelligence is (...) “one of three being offered experimentally by the Stanford computer science department to extend technology knowledge and skills beyond this elite campus to the entire world.” One of the other two courses, an introduction to database software, is being taught by Pearson author Dr. Jennifer Widom. Artificial Intelligence: A Modern Approach, 3e is available to purchase as an eText for your Kindle™, NOOK™, and the iPhone®/iPad®. (shrink)

Does the human mind resemble the machines that can behave like it? Biologically inspired machine-learning systems approach “human-level” accuracy in an astounding variety of domains, and even predict human brain activity—raising the exciting possibility that such systems represent the world like we do. However, even seemingly intelligent machines fail in strange and “unhumanlike” ways, threatening their status as models of our minds. How can we know when human–machine behavioral differences reflect deep disparities in their underlying capacities, vs. when such failures (...) are only superficial or peripheral? This article draws on a foundational insight from cognitive science—the distinction between performance and competence—to encourage “species-fair” comparisons between humans and machines. The performance/competence distinction urges us to consider whether the failure of a system to behave as ideally hypothesized, or the failure of one creature to behave like another, arises not because the system lacks the relevant knowledge or internal capacities (“competence”), but instead because of superficial constraints on demonstrating that knowledge (“performance”). I argue that this distinction has been neglected by research comparing human and machine behavior, and that it should be essential to any such comparison. Focusing on the domain of image classification, I identify three factors contributing to the species-fairness of human–machine comparisons, extracted from recent work that equates such constraints. Species-fair comparisons level the playing field between natural and artificial intelligence, so that we can separate more superficial differences from those that may be deep and enduring. (shrink)

Philosophical Issues, Volume 30, Issue 1, Page 273-292, October 2020.

I argue that semantics is the study of the proprietary database of a centrally inaccessible and informationally encapsulated input–output system. This system’s role is to encode and decode partial and defeasible evidence of what speakers are saying. Since information about nonlinguistic context is therefore outside the purview of semantic processing, a sentence’s semantic value is not its content but a partial and defeasible constraint on what it can be used to say. I show how to translate this thesis into a (...) detailed compositional-semantic theory based on the influential framework of Heim and Kratzer. This approach situates semantics within an independently motivated account of human cognitive architecture and reveals the semantics–pragmatics interface to be grounded in the underlying interface between modular and central systems. (shrink)

Many debates in philosophy focus on whether folk or scientific psychological notions pick out cognitive natural kinds. Examples include memory, emotions and concepts. A potentially interesting type of kind is: kinds of mental representations (as opposed, for example, to kinds of psychological faculties). In this chapter we outline a proposal for a theory of representational kinds in cognitive science. We argue that the explanatory role of representational kinds in scientific theories, in conjunction with a mainstream approach to explanation in cognitive (...) science, suggest that representational kinds are multi-level. This is to say that representational kinds’ properties cluster at different levels of explanation and allow for intra- and inter-level projections. (shrink)

The core of Zimmerman’s picture posits an inverse correlation between an action’s automaticity and belief’s role in the action’s execution. This proposal faces serious problems. First, high-attention, high-control actions don’t seem to heighten awareness of one’s beliefs. Second, low-attention, low-control actions are caused by the same states at play when executing high-attention, high-control actions, in which case there is no ontological difference in the states involved in these behaviors. Third, on Zimmerman’s view it is unclear what it is for a (...) state to be involved in behaviors at all, as the basic realist response—that beliefs cause behavior—is unavailable to a Zimmerman-style pragmatist. Lastly, if Zimmerman's view were right and low-level behaviors weren't caused by beliefs, then we should turn our attention to those states instead, as most of our behavior isn’t executed under conditions of high control and attention. (shrink)

Unconscious logical inference seems to rely on the syntactic structures of mental representations (Quilty-Dunn & Mandelbaum 2018). Other transitions, such as transitions using iconic representations and associative transitions, are harder to assimilate to syntax-based theories. Here we tackle these difficulties head on in the interest of a fuller taxonomy of mental transitions. Along the way we discuss how icons can be compositional without having constituent structure, and expand and defend the “symmetry condition” on Associationism (the idea that associative links and (...) transitions are perfectly symmetric). In the end, we show how a BIT (“bare inferential transition”) theory can cohabitate with these other non-inferential mental transitions. (shrink)

Deep learning is currently the most prominent and widely successful method in artificial intelligence. Despite having played an active role in earlier artificial intelligence and neural network research, philosophers have been largely silent on this technology so far. This is remarkable, given that deep learning neural networks have blown past predicted upper limits on artificial intelligence performance—recognizing complex objects in natural photographs and defeating world champions in strategy games as complex as Go and chess—yet there remains no universally accepted explanation (...) as to why they work so well. This article provides an introduction to these networks as well as an opinionated guidebook on the philosophical significance of their structure and achievements. It argues that deep learning neural networks differ importantly in their structure and mathematical properties from the shallower neural networks that were the subject of so much philosophical reflection in the 1980s and 1990s. The article then explores several different explanations for their success and ends by proposing three areas of inquiry that would benefit from future engagement by philosophers of mind and science. (shrink)

Short‐term memory in vision is typically thought to divide into at least two memory stores: a short, fragile, high‐capacity store known as iconic memory, and a longer, durable, capacity‐limited store known as visual working memory (VWM). This paper argues that iconic memory stores icons, i.e., image‐like perceptual representations. The iconicity of iconic memory has significant consequences for understanding consciousness, nonconceptual content, and the perception–cognition border. Steven Gross and Jonathan Flombaum have recently challenged the division between iconic memory and VWM by (...) arguing against the idea of capacity limits in favor of a flexible resource‐based model of short‐term memory. I argue that, while VWM capacity is probably governed by flexible resources rather than a sharp limit, the two memory stores should still be distinguished by their representational formats. Iconic memory stores icons, while VWM stores discursive (i.e., language‐like) representations. I conclude by arguing that this format‐based distinction between memory stores entails that prominent views about consciousness and the perception–cognition border will likely have to be revised. (shrink)

How can we think about things in the outside world? There is still no widely accepted theory of how mental representations get their meaning. In light of pioneering research, Nicholas Shea develops a naturalistic account of the nature of mental representation with a firm focus on the subpersonal representations that pervade the cognitive sciences.

Perceptual systems respond to proximal stimuli by forming mental representations of distal stimuli. A central goal for the philosophy of perception is to characterize the representations delivered by perceptual systems. It may be that all perceptual representations are in some way proprietarily perceptual and differ from the representational format of thought (Dretske 1981; Carey 2009; Burge 2010; Block ms.). Or it may instead be that perception and cognition always trade in the same code (Prinz 2002; Pylyshyn 2003). This paper rejects (...) both approaches in favor of perceptual pluralism, the thesis that perception delivers a multiplicity of representational formats, some proprietary and some shared with cognition. The argument for perceptual pluralism marshals a wide array of empirical evidence in favor of iconic (i.e., image-like, analog) representations in perception as well as discursive (i.e., language-like, digital) perceptual object representations. (shrink)

My topic is the intelligent guidance of action. In this paper I offer an empirically grounded case for four ideas: that [a] cognitive processes of practical reasoning play a key role in the intelligent guidance of action, [b] these processes could not do so without significant enabling work done by both perception and the motor system, [c] the work done by perceptual and motor systems can be characterized as the generation of information specialized for action guidance, which in turn suggests (...) that [d] the cognitive processes of practical reasoning that play a key role in the guidance of intelligent action are not the abstract, syllogistic ones philosophers often treat as the paradigm of practical reasoning. Rather, these cognitive processes are constrained by, and work well with, the specialized concepts outputted by perception and the feedback outputted by sensorimotor processes. (shrink)

Recent progress in artificial intelligence has renewed interest in building systems that learn and think like people. Many advances have come from using deep neural networks trained end-to-end in tasks such as object recognition, video games, and board games, achieving performance that equals or even beats that of humans in some respects. Despite their biological inspiration and performance achievements, these systems differ from human intelligence in crucial ways. We review progress in cognitive science suggesting that truly human-like learning and thinking (...) machines will have to reach beyond current engineering trends in both what they learn and how they learn it. Specifically, we argue that these machines should build causal models of the world that support explanation and understanding, rather than merely solving pattern recognition problems; ground learning in intuitive theories of physics and psychology to support and enrich the knowledge that is learned; and harness compositionality and learning-to-learn to rapidly acquire and generalize knowledge to new tasks and situations. We suggest concrete challenges and promising routes toward these goals that can combine the strengths of recent neural network advances with more structured cognitive models. (shrink)

The notion of an object file figures prominently in recent work in philosophy and cognitive science. Object files play a role in theories of singular reference, object individuation, perceptual memory, and the development of cognitive capacities. However, the philosophical literature lacks a detailed, empirically informed theory of object files. In this paper, we articulate and defend the multiple-slots view, which specifies both the format and architecture of object files. We argue that object files represent in a non-iconic, propositional format that (...) incorporates discrete symbols for separate features. Moreover, we argue that features of separate categories are stored in separate memory slots within an object file. We supplement this view with a computational framework that characterizes how information about objects is stored and retrieved. (shrink)

Dispositionalism about belief has had a recent resurgence. In this paper we critically evaluate a popular dispositionalist program pursued by Eric Schwitzgebel. Then we present an alternative: a psychofunctional, representational theory of belief. This theory of belief has two main pillars: that beliefs are relations to structured mental representations, and that the relations are determined by the generalizations under which beliefs are acquired, stored, and changed. We end by describing some of the generalizations regarding belief acquisition, storage, and change.

ABSTRACTThis paper provides a naturalistic account of inference. We posit that the core of inference is constituted by bare inferential transitions, transitions between discursive mental representations guided by rules built into the architecture of cognitive systems. In further developing the concept of BITs, we provide an account of what Boghossian [2014] calls ‘taking’—that is, the appreciation of the rule that guides an inferential transition. We argue that BITs are sufficient for implicit taking, and then, to analyse explicit taking, we posit (...) rich inferential transitions, which are transitions that the subject is disposed to endorse. (shrink)