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 present interpretation‐based processing—a theory of sentence processing that builds a syntactic and a semantic representation for a sentence and assigns an interpretation to the sentence as soon as possible. That interpretation can further participate in comprehension and in lexical processing and is vital for relating the sentence to the prior discourse. Our theory offers a unified account of the processing of literal sentences, metaphoric sentences, and sentences containing semantic illusions. It also explains how text can prime lexical access. We (...) show that word literality is a matter of degree and that the speed and quality of comprehension depend both on how similar words are to their antecedents in the preceding text and how salient the sentence is with respect to the preceding text. Interpretation‐based processing also reconciles superficially contradictory findings about the difference in processing times for metaphors and literals. The theory has been implemented in ACT‐R [Anderson and Lebiere, The Atomic Components of Thought, Lawrence Erlbaum Associates Publishers, Mahwah, NJ, 1998]. (shrink)

. References. Critical Review: Vol. 18, Democratic Competence, pp. 331-360.

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

The Darwinian theory of evolution is itself evolving and this book presents the details of the core of modern Darwinism and its latest developmental directions. The authors present current scientific work addressing theoretical problems and challenges in four sections, beginning with the concepts of evolution theory, its processes of variation, heredity, selection, adaptation and function, and its patterns of character, species, descent and life. The second part of this book scrutinizes Darwinism in the philosophy of science and its usefulness in (...) understanding ecosystems, whilst the third section deals with its application in disciplines beyond the biological sciences, including evolutionary psychology and evolutionary economics, Darwinian morality and phylolinguistics. The final section addresses anti-Darwinism, the creationist view and issues around teaching evolution in secondary schools. The reader learns how current experimental biology is opening important perspectives on the sources of variation, and thus of the very power of natural selection. This work examines numerous examples of the extension of the principle of natural selection and provides the opportunity to critically reflect on a rich theory, on the methodological rigour that presides in its extensions and exportations, and on the necessity to measure its advantages and also its limits. Scholars interested in modern Darwinism and scientific research, its concepts, research programs and controversies will find this book an excellent read, and those considering how Darwinism might evolve, how it can apply to the human sciences and other disciplines beyond its origins will find it particularly valuable. Originally produced in French (Les Mondes Darwiniens), the scope and usefulness of the book have led to the production of this English text, to reach a wider audience. This book is a milestone in the impressive penetration by Francophone scholars into the world of Darwinian science, its historiography and philosophy over the last two decades. Alex Rosenberg, R. Taylor Cole Professor of Philosophy, Duke University Until now this useful and comprehensive handbook has only been available to francophones. Thanks to this invaluable new translation, this collection of insightful and original essays can reach the global audience it deserves. Tim Lewens, University of Cambridge. (shrink)

As artificial intelligence (AI) thrives and propagates through modern life, a key question to ask is how to include humans in future AI? Despite human involvement at every stage of the production process from conception and design through to implementation, modern AI is still often criticized for its “black box” characteristics. Sometimes, we do not know what really goes on inside or how and why certain conclusions are met. Future AI will face many dilemmas and ethical issues unforeseen by their (...) creators beyond those commonly discussed (e.g., trolley problems and variants of it) and to which solutions cannot be hard-coded and are often still up for debate. Given the sensitivity of such social and ethical dilemmas and the implications of these for human society at large, when and if our AI make the “wrong” choice we need to understand how they got there in order to make corrections and prevent recurrences. This is particularly true in situations where human livelihoods are at stake (e.g., health, well-being, finance, law) or when major individual or household decisions are taken. Doing so requires opening up the “black box” of AI; especially as they act, interact, and adapt in a human world and how they interact with other AI in this world. In this article, we argue for the application of cognitive architectures for ethical AI. In particular, for their potential contributions to AI transparency, explainability, and accountability. We need to understand how our AI get to the solutions they do, and we should seek to do this on a deeper level in terms of the machine-equivalents of motivations, attitudes, values, and so on. The path to future AI is long and winding but it could arrive faster than we think. In order to harness the positive potential outcomes of AI for humans and society (and avoid the negatives), we need to understand AI more fully in the first place and we expect this will simultaneously contribute towards greater understanding of their human counterparts also. (shrink)

I provide an interpretation of Wittgenstein's much criticized remarks on Gödel's First Incompleteness Theorem in the light of paraconsistent arithmetics: in taking Gödel's proof as a paradoxical derivation, Wittgenstein was right, given his deliberate rejection of the standard distinction between theory and metatheory. The reasoning behind the proof of the truth of the Gödel sentence is then performed within the formal system itself, which turns out to be inconsistent. I show that the models of paraconsistent arithmetics (obtained via the Meyer-Mortensen (...) collapsing filter on the standard model) match with many intuitions underlying Wittgensteins philosophy of mathematics, such as its strict finitism and the insistence on the decidability of any meaningful mathematical question. (shrink)

Cognitive science is an interdisciplinary research endeavor focusing on human cognitive phenomena such as memory, language use, and reasoning. It emerged in the second half of the 20th century and is charting new directions at the beginning of the 21st century. This chapter begins by identifying the disciplines that contribute to cognitive science and reviewing the history of the interdisciplinary engagements that characterize it. The second section examines the role that mechanistic explanation plays in cognitive science, while the third focuses (...) on the importance of mental representations in specifically cognitive explanations. The fourth section considers the interdisciplinary nature of cognitive science and explores how multiple disciplines can contribute to explanations that exceed what any single discipline might accomplish. The conclusion sketches some recent developments in cognitive science and their implications for philosophers. (shrink)

The relation between logic and thought has long been controversial, but has recently influenced theorizing about the nature of mental processes in cognitive science. One prominent tradition argues that to explain the systematicity of thought we must posit syntactically structured representations inside the cognitive system which can be operated upon by structure sensitive rules similar to those employed in systems of natural deduction. I have argued elsewhere that the systematicity of human thought might better be explained as resulting from the (...) fact that we have learned natural languages which are themselves syntactically structured. According to this view, symbols of natural language are external to the cognitive processing system and what the cognitive system must learn to do is produce and comprehend such symbols. In this paper I pursue that idea by arguing that ability in natural deduction itself may rely on pattern recognition abilities that enable us to operate on external symbols rather than encodings of rules that might be applied to internal representations. To support this suggestion, I present a series of experiments with connectionist networks that have been trained to construct simple natural deductions in sentential logic. These networks not only succeed in reconstructing the derivations on which they have been trained, but in constructing new derivations that are only similar to the ones on which they have been trained. (shrink)

Cognitive psychologists, like biologists, frequently describe mechanisms when explaining phenomena. Unlike biologists, who can often trace material transformations to identify operations, psychologists face a more daunting task in identifying operations that transform information. Behavior provides little guidance as to the nature of the operations involved. While not itself revealing the operations, identification of brain areas involved in psychological mechanisms can help constrain attempts to characterize the operations. In current memory research, evidence that the same brain areas are involved in what (...) are often taken to be different memory phenomena or in other cognitive phenomena is playing such a heuristic function. †To contact the author, please write to: Department of Philosophy, 0119, University of California, San Diego, La Jolla, CA 92093‐0119; e‐mail: [email protected]. (shrink)

Laboratory-based studies of problem solving suggest that transfer of solution principles from an analogue to a target arises only minimally without the presence of directive hints. Recently, however, real-world studies indicate that experts frequently and spontaneously use analogies in domain-based problem solving. There is also some evidence that in certain circumstances domain novices can draw analogies designed to illustrate arguments. It is less clear, however, whether domain novices can invoke analogies in the sophisticated manner of experts to enable them to (...) progress problem solving. In the current study groups of novices and experts tackled large-scale management problems. Spontaneous analogising was observed in both conditions, with no marked differences between expertise levels in the frequency, structure, or function of analogising. On average four analogies were generated by groups per hour, with significantly more relational mappings between analogue and target being produced than superficial object-and-attribute mappings. Analogising served two different purposes: problem solving (dominated by relational mappings), and illustration (which for novices was dominated by object-and-attribute mappings). Overall, our novices showed a sophistication in domain-based analogical reasoning that is usually only observed with experts, in addition to a sensitivity to the pragmatics of analogy use. (shrink)

This paper is a commentary on the target article by Lokendra Shastri & Venkat Ajjanagadde [S&A]: “From simple associations to systematic reasoning: A connectionist representation of rules, variables and dynamic bindings using temporal synchrony” in same issue of the journal, pp.417–451. -/- It puts S&A's temporal-synchrony binding method in a broader context, comments on notions of pointing and other ways of associating information - in both computers and connectionist systems - and mentions types of reasoning that are a challenge to (...) S&A's system (amongst many other connectionist systems, recognizing that S&A's is much more capable as regards reasoning than most other contemporary systems). (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)

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)

Several recent criticisms of the somatic marker hypothesis (SMH) identify multiple ambiguities in the way it has been formulated by its chief proponents. Here we provide evidence that this hypothesis has also been interpreted in various different ways by the scientific community. Our diagnosis of this problem is that SMH lacks an adequate computational-level account of practical decision making. Such an account is necessary for drawing meaningful links between neurological- and psychological-level data. The paper concludes by providing a simple, five-step (...) model of practical decision making. Recasting SMH in terms of this model generates more precise and empirically tractable computational-level hypotheses about the various ways that somatic markers might influence practical decisions. (shrink)

It is often argued that human emotions, and the cognitions that accompany them, involve refinements of, and extensions to, more basic functionality shared with other species. Such refinements may rely on common or on distinct processes and representations. Multi-level theories of cognition and affect make distinctions between qualitatively different types of representations often dealing with bodily, affective and cognitive attributes of self-related meanings. This paper will adopt a particular multi-level perspective on mental architecture and show how a mechanism of subsystem (...) differentiation could have allowed an evolutionarily “old” role for emotion in the control of action to have altered into one more closely coupled to meaning systems. We conclude by outlining some illustrative consequences of our analysis that might usefully be addressed in research in comparative psychology, cognitive archaeology, and in laboratory research on memory for emotional material. (shrink)

Recently, there has been a growing interest, both within theoretical biology and the philosophy of biology, in the possibility and desirability of a theory of development. Among the many issues raised within this debate, the questions of the spatial and temporal boundaries of development have received particular attention. In this article, noting that so far the discussion has mostly centered on the processes of morphogenesis and organogenesis, we argue that an important missing element in the equation, namely the development of (...) language and cognition in general, may play an important role in settling the issue of temporal boundaries. After examining the idea that the development of language, cognition, and action are bona fide biological processes, we explore the consequences for a general theory of development of taking them into consideration. (shrink)

Sciences able to identify appropriate analytical units for their domain, their natural kinds, have tended to be more progressive. In the biological sciences, evolutionary natural kinds are adaptations that can be identified by their common history of selection for some function. Human brains are the product of an evolutionary history of selection for component systems which produced behaviours that gave adaptive advantage to their hosts. These structures, behaviour production systems, are the natural kinds that psychology seeks. We argue these can (...) be identified deductively by classing behaviour first according to its level of behavioural control. Early animals in our lineage used only reactive production, Vertebrates evolved motivation, and later Primates developed executive control. Behaviour can also be classified by the type of evolutionary benefit it bestows: it can deliver either immediate benefits (food, gametes), improvements in the individual’s position with respect to the world (resource access, social status), or improvements in the ability to secure future benefits (knowledge, skill). Combining history and function implies the existence of seven types of behaviour production systems in human brains responsible for reflexive, instinctual, exploratory, driven, emotional, playful and planned behaviour. Discovering scientifically valid categories of behaviour can provide a fundamental taxonomy and common language for understanding, predicting and changing behaviour, and a way of discovering the organs in the brain––its natural kinds––that are responsible for behaviour. (shrink)

Since Freud and his co-author Breuer spoke of dissociation in 1895, a scientific paradigm was painstakingly established in the field of unconscious cognition. This is the dissociation paradigm. However, recent critical analysis of the many and various reported dissociations reveals their blurred, or unveridical, character. Moreover, we remain ignorant with respect to the ways cognitive phenomena transition from consciousness to an unconscious mode. This hinders us from filling in the puzzle of the unified mind. We conclude that we have reached (...) a Kuhnian crisis in the field of unconscious cognition, and we predict that new models, incorporating partly the relevant findings of the dissociation paradigm—but also of dynamic psychology—, will soon be established. We further predict that some of these models will be largely based on the pairs representation–process and analog–digital. (shrink)

The representational nature of human cognition and thought in general has been a source of controversies. This is particularly so in the context of studies of unconscious cognition, in which representations tend to be ontologically and structurally segregated with regard to their conscious status. However, it appears evolutionarily and developmentally unwarranted to posit such segregations, as,otherwise, artifact structures and ontologies must be concocted to explain them from the viewpoint of the human cognitive architecture. Here, from a by-and-large Classical cognitivist viewpoint, (...) I show why this segregation is wrong, and elaborate on the need to postulate an ontological and structural continuity between unconscious and conscious representations. Specifically, I hypothesize that this continuity is to be found in the symbolic-based interplay between the syntax and the semantics of thought, and I propose a model of human information processing characterized by the integration of syntactic and semantic representations. (shrink)

Contemporary studies in unconscious cognition are essentially founded on dissociation, i.e., on how it dissociates with respect to conscious mental processes and representations. This is claimed to be in so many and diverse ways that one is often lost in dissociation. In order to reduce this state of confusion we here carry out two major tasks: based on the central distinction between cognitive processes and representations, we identify and isolate the main dissociation paradigms; we then critically analyze their key tenets (...) and reported findings. (shrink)

This article addresses the question of the mechanisms of the emergence of structure and meaning in the biological and physical sciences. It proceeds from an examination of the concept of intentionality and proposes a model of intentional behavior on the basis of results of computer simulations of structural and functional self-organization. Current attempts to endow intuitive aspects of meaningful complexity with operational content are analyzed and the metaphor of DNA as a computer program (the `genetic program') is critically examined in (...) relation to an alternative metaphor of DNA as data. It is argued that relatively simple networks of boolean automata can classify and recognize patterns of binary strings on the basis of non-programmed, self-generated criteria, but lack a capacity for self-observation and interpretation. To overcome this problem it is necessary to clarify the relationships between the goals and underlying mechanisms of a process and between a system and its environment. It will be shown that memory devices that record the histories of interactions are essential for models of conscious and unconscious intentional behavior and that the possibility of infinitely sophisticated - and therefore unprogrammable - machines cannot be avoided. It will be argued that the notion of infinite sophistication allows the ideas of self-organization and physical determinism to be reconciled. These models will be used to suggest how the voluntary aspect of decision-making in general can emerge out of functional self-organizing processes. The conclusion will introduce the notion of `underdetermination' of theories, which imposes an intrinsic limitation on models of complex natural systems - a limitation that, at the same time, may be precisely what makes possible mutual understanding and intersubjectivity. (shrink)

Sometimes, we face choices between actions most likely to lead to valuable outcomes, and actions which put us in a better position to learn. These choices exemplify what is called the exploration/exploitation trade-off. In computer science and psychology, this trade-off has fruitfully been applied to modulating the way agents or systems make choices over time. This article extends the trade-off to belief. We can be torn between two ways of believing, one of which is expected to be more accurate in (...) light of current evidence, whereas the other is expected to lead to more learning opportunity and accuracy in the long run. Further, it is sometimes rationally permissible to choose the latter. The article breaks down the features of action which give rise to the trade-off, and then argues that each feature applies equally well to belief. This conclusion is an instance of a systematic, foreseeable way in which what is rational to believe now depends on what one expects to be doing in the future. That is, epistemic rationality fundamentally concerns time. (shrink)

Much of cognitive psychology focuses on effects measured in tens of milliseconds while significant educational outcomes take tens of hours to achieve. The task of bridging this gap is analyzed in terms of Newell's (1990) bands of cognition—the Biological, Cognitive, Rational, and Social Bands. The 10 millisecond effects reside in his Biological Band while the significant learning outcomes reside in his Social Band. The paper assesses three theses: The Decomposition Thesis claims that learning occurring at the Social Band can be (...) reduced to learning occurring at lower bands. The Relevance Thesis claims that instructional outcomes at the Social Band can be improved by paying attention to cognition at the lower bands. The Modeling Thesis claims that cognitive modeling provides a basis for bridging between events on the small scale and desired outcomes on the large scale. The unit-task level, at the boundary of the Cognitive and Rational Bands, is useful for assessing these theses. There is good evidence for all three theses in efforts that bridge from the unit-task level to educational applications. While there is evidence for the Decomposition Thesis all the way down to the 10 millisecond level, more work needs to be done to establish the Relevance Thesis and particularly the Modeling Thesis at the lower levels. (shrink)