Artificial Neural Networks have reached “grandmaster” and even “super-human” performance across a variety of games, from those involving perfect information, such as Go, to those involving imperfect information, such as “Starcraft”. Such technological developments from artificial intelligence (AI) labs have ushered concomitant applications across the world of business, where an “AI” brand-tag is quickly becoming ubiquitous. A corollary of such widespread commercial deployment is that when AI gets things wrong—an autonomous vehicle crashes, a chatbot exhibits “racist” behavior, automated credit-scoring processes (...) “discriminate” on gender, etc.—there are often significant financial, legal, and brand consequences, and the incident becomes major news. As Judea Pearl sees it, the underlying reason for such mistakes is that “... all the impressive achievements of deep learning amount to just curve fitting.” The key, as Pearl suggests, is to replace “reasoning by association” with “causal reasoning” —the ability to infer causes from observed phenomena. It is a point that was echoed by Gary Marcus and Ernest Davis in a recent piece for theNew York Times: “we need to stop building computer systems that merely get better and better at detecting statistical patterns in data sets—often using an approach known as ‘Deep Learning’—and start building computer systems that from the moment of their assembly innately grasp three basic concepts: time, space, and causality.” In this paper, foregrounding what in 1949 Gilbert Ryle termed “a category mistake”, I will offer an alternative explanation for AI errors; it is not so much that AI machinery cannot “grasp” causality, but that AI machinery (qua computation) cannot understand anything at all. (shrink)

When experimental philosophers carry out studies on thought experiments, some participants are excluded based on certain exclusion criteria, mirroring standard social science vignette methodology. This involves excluding people that do not pay attention or who miscomprehend the scenario presented in thought experiments. However, experimental philosophy studies sometimes exclude an alarmingly high number of participants. We argue that this threatens the external and internal validity of the conclusions being drawn and we show how a simple visualization of thought experiments can reduce (...) exclusion rates significantly. Furthermore, we argue that focus should not merely be on how many are excluded, but also why they are excluded, and we highlight the role of comprehension questions in this regard. Philosophical thought experiments often rely on the acceptance of certain key premises that may be regarded contestable, and asking comprehension questions involving such key assumptions could be problematic as that may result in some participants being inadvertently excluded from the study, potentially creating a selection bias. (shrink)

The article discusses the case of Microsoft's Twitter chatbot Tay that “turned into a Nazi” after less than 24 hours from its release on the Internet. The first section presents a brief recapitulation of Alan Turing's proposal for a test for artificial intelligence and the way it influenced subsequent discussions in the philosophy of mind. In the second section, I offer a few arguments appealing for caution regarding the identification of an accomplished chatbot as a thinking being. These are motivated (...) principally by Ludwig Wittgenstein's discussions of mind and soul and by some Wittgensteinian philosophers' criticisms of AI endeavours. I will try to show that it is very difficult, if not impossible, to make sense of AIs such as chatbots as thinking beings, rather independently of their technical perfection and accomplishment. In the third section, the case of the “Nazi chatbot” Tay will offer me material for some light to be shed on the peculiar character of our interconnected concepts of thinking, soul and person and on the importance of their further ramified connections. (shrink)

Ned Block. Psychologism and behaviorism. Philosophical Review, 90, 5-43.) argued that a behaviorist conception of intelligence is mistaken, and that the nature of an agent's internal processes is relevant for determining whether the agent has intelligence. He did that by describing a machine which lacks intelligence, yet can answer questions put to it as an intelligent person would. The nature of his machine's internal processes, he concluded, is relevant for determining that it lacks intelligence. I argue against Block that it (...) is not the nature of its processes but of its linguistic behavior which is responsible for his machine's lack of intelligence. As I show, not only has Block failed to establish that the nature of internal processes is conceptually relevant for psychology, in fact his machine example actually supports some version of behaviorism. As Wittgenstein has maintained, as far as psychology is concerned, there may be chaos inside. (shrink)

Searle's Chinese Room was supposed to prove that computers can't understand: the man in the room, following, like a computer, syntactical rules alone, though indistinguishable from a genuine Chinese speaker, doesn't understand a word. But such a room is impossible: the man won't be able to respond correctly to questions like What is the time?, even though such an ability is indispensable for a genuine Chinese speaker. Several ways to provide the room with the required ability are considered, and it (...) is concluded that for each of these the room will have understanding. Hence, Searle's argument is invalid. (shrink)

The object of the study is a special branch of scientific knowledge, formed at the intersection of interests of various humanitarian disciplines – the philosophy of artificial intelligence; the subject of the study is the problems it considers. The author identifies the main traditional questions that the philosophy of artificial intelligence tried to answer earlier. Such questions in the earlier stages of the development of science were concentrated around the possibility of artificial intelligence to become aware of itself, learn to (...) think, feel and create like a person. Scientists since the 1960s interested in the problems of equality of natural and artificial intelligence, the ability of a computer to show benevolence or hostility towards its creators. The research was carried out using a historical and philosophical analysis of issues in the philosophy of artificial intelligence, the method of interpretive synthesis, etc. Systemic-structural, dialectical, cultural-historical, value-based, interdisciplinary approaches to the problem were used. The scientific novelty of the research lies in identifying the problems of the philosophy of artificial intelligence that are relevant at this stage of development of this scientific field. It is noted that the dynamics of the philosophy of artificial intelligence are significantly influenced by the fact that “strong” artificial intelligence has not been created for several decades. This has caused a transformation in the research field of the philosophy of artificial intelligence, and at the moment this branch of scientific knowledge is considering new questions, including why it is not possible to create “strong” AI that completely replicates human thinking; can a computer think, albeit differently, than a human; what is the difference between human and machine thinking; who should be responsible for the decisions and actions of artificial intelligence, and etc. The issues facing the philosophy of artificial intelligence are constantly updated. (shrink)

We can readily identify goal-directed systems and distinguish them from non-goal-directed systems. A woodpecker hunting for grubs is the first, a pendulum returning to rest is the second. But what is it to be a goal-directed system? Perhaps the dominant answer to this question, inspired by systems theories such as cybernetics, is that goal-directed systems are distinguished by their tendency to seek, aim at, or maintain some more-or-less easily identifiable goal. Cybernetics and the like would hold that physical systems subject (...) only to physical laws can exhibit such behavior. If sound, this systems approach to teleology would unify a diverse range of goal-directed phenomena and neatly side-step many traditional bogey-men of teleology, such as anthropomorphism and future causation. Goal-directed phenomena would be a normal feature of the natural causal world that could be described in purely descriptive and quantitative terms, and receive ordinary causal explanations. Thus, the systems approach promises to provide a naturalistic-cum-descriptive account of teleology suitable for use in naturalistic accounts of other phenomena, including the intentionality of mental states and even self-consciousness. (shrink)

In this paper the author tries to disentangle some of the problems tied up in John Searle's famous Chinese-room-argument. In a first step to answer the question what it would be for a system to have not only syntax, but also semantics the author gives a brief account of the functioning of the language understanding systems (LUS) so far developed in the framework of AI research thereby making clear that systems like Winograd's SHRDLU are indeed doing little more than mere (...) number crunching. But things would be entirely different, the author argues, if the database of a LUS were built up by the system itself via some perceptual component-at least, if this perceptual component had the capacity to distinguish objects having a certain property F from objects which do not. For in this case the system could store an internal representation of the fact that the object has the property F in its database if and only if the object in fact has that property. And this would be a good basis for calling such a system a genuine LUS. But Searle has objected to a very similar account of J. Fodor that nothing could be further from true language understanding. The reason for this complaint seems to be that Searle holds the view that a true LUS must e.g., know that the word hamburgers refers to hamburgers and that he moreover claims that this knowledge must be explicit or that the system must be aware of the reference of hamburgers to hamburgers. The author argues that this is asking too much. For it seems plausible to say that a system is able to understand e.g., the word hamburger even if it has only implicit knowledge of the fact that hamburger refers to hamburgers in the sense that it has the capacity to tell hamburgers from non hamburgers and the capacity to bring the word hamburger together just with objects of the former kind. (shrink)

Many of the mathematicians and scientists who guided the development of digital computers in the late 1940s, such as Alan Turing and John von Neumann, saw these new devices not just as tools for calculation but as devices that might employ the same principles as are exhibited in rational human thought. Thus, a subfield of what came to be called computer science assumed the label artificial intelligence. The idea of building artificial systems which could exhibit intelligent behavior comparable to that (...) of humans was a heady prospect, and the claims made on behalf of AI during the 1950s and 1960s were impossibly ambitious. Despite some theoretical and applied successes within the field, serious problems soon became evident. Instead of fulfilling the goal of quickly producing artificial intelligent agents which could compete with or outperform human beings, by the 1970s and 1980s AI had settled into a pattern of slower but real progress in modeling or simulating aspects of human intelligence. (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)

In philosophy the term intentionality refers to the feature possessed by mental states of beingabout things others than themselves. A serious question has been how to explain the intentionality of mental states. This paper starts with linguistic representations, and explores how an organism might use linguistic symbols to represent other things. Two research projects of Sue Savage-Rumbaugh, one explicity teaching twopan troglodytes to use lexigrams intentionally, and the other exploring the ability of several members ofpan paniscus to learn lexigram use (...) and comprehension of English speech spontaneously when raised in an appropriate environment, are examined to explore the acquisition process. Although it is controversial whether intentionality of mental states or linguistic symbols is primary, it is argued that the intentionality of linguistic symbols is primary and that studying how organisms learn to use linguistic symbols provides an avenue to understanding how intentionality is acquired by cognitive systems. (shrink)

This paper reviews four significant advances on the feedforward architecture that has dominated discussions of connectionism. The first involves introducing modularity into networks by employing procedures whereby different networks learn to perform different components of a task, and a Gating Network determines which network is best equiped to respond to a given input. The second consists in the use of recurrent inputs whereby information from a previous cycle of processing is made available on later cycles. The third development involves developing (...) compressed representations of strings in which there is no longer an explicit encoding of the components but where information about the structure of the original string can be recovered and so is present functionally. The final advance entails using connectionist learning procedures not just to change weights in networks but to change the patterns used as inputs to the network. These advances significantly increase the usefulness of connectionist networks for modeling human cognitive performance by, among other things, providing tools for explaining the productivity and systematicity of some mental activities, and developing representations that are sensitive to the content they are to represent. (shrink)

Philosophy of science is positioned to make distinctive contributions to cognitive science by providing perspective on its conceptual foundations and by advancing normative recommendations. The philosophy of science I embrace is naturalistic in that it is grounded in the study of actual science. Focusing on explanation, I describe the recent development of a mechanistic philosophy of science from which I draw three normative consequences for cognitive science. First, insofar as cognitive mechanisms are information-processing mechanisms, cognitive science needs an account of (...) how the representations invoked in cognitive mechanisms carry information about contents, and I suggest that control theory offers the needed perspective on the relation of representations to contents. Second, I argue that cognitive science requires, but is still in search of, a catalog of cognitive operations that researchers can draw upon in explaining cognitive mechanisms. Last, I provide a new perspective on the relation of cognitive science to brain sciences, one which embraces both reductive research on neural components that figure in cognitive mechanisms and a concern with recomposing higher-level mechanisms from their components and situating them in their environments. (shrink)

Philosophy of science is positioned to make distinctive contributions to cognitive science by providing perspective on its conceptual foundations and by advancing normative recommendations. The philosophy of science I embrace is naturalistic in that it is grounded in the study of actual science. Focusing on explanation, I describe the recent development of a mechanistic philosophy of science from which I draw three normative consequences for cognitive science. First, insofar as cognitive mechanisms are information‐processing mechanisms, cognitive science needs an account of (...) how the representations invoked in cognitive mechanisms carry information about contents, and I suggest that control theory offers the needed perspective on the relation of representations to contents. Second, I argue that cognitive science requires, but is still in search of, a catalog of cognitive operations that researchers can draw upon in explaining cognitive mechanisms. Last, I provide a new perspective on the relation of cognitive science to brain sciences, one which embraces both reductive research on neural components that figure in cognitive mechanisms and a concern with recomposing higher‐level mechanisms from their components and situating them in their environments. (shrink)

In his paper "Machines and the Mental" Fred Dretske claims that there is a difference on principle between men on the one and animals and machines on the other side which arises from the fact that men are able to respond to the meaning of symbols whereas animals and machines can only respond to the symbols that have the meaning. In this paper it is argued that this claim does not bear closer scrutiny. Mainly for two reasons: 1. The mere (...) having of a certain meaning does not rank among the possible causal properties of a symbol, not even in the case of men; for it is not the having of the meaning that may have certain effects on the behavior of a person, but the fact that the person associates this meaning with the symbol. 2. Even animals are capable to associate a meaning with a symbol, i. e. even animals can learn to come to believe that p if they become aware of a symbol which has the meaning that p. (shrink)

I start with some of the early challenges to the widely-employed philosophical method of cases—the very challenges that originally prompted the new movement of experimental philosophy—and with some fundamental questions about the method that are yet to have been given satisfying answers. I then propose that what has allowed both ‘armchair’ and ‘experimental’ participants in the ongoing debates concerning the method to ignore or repress those early challenges—and in particular Robert Cummins’s ‘calibration objection’—and to discount fundamental disagreements about those questions, (...) is in large part ‘the claim of continuity’, which is the claim, or assumption, that there is no philosophically significant difference between whatever it is that we do when we give our answer to the theorist’s question of whether some philosophically interesting word ‘applies’ (positively or negatively) to some theoretically significant case, and our use of that same word in the course of ordinary, non-philosophical discourse. I then summarize the ordinary language philosophy (OLP) argument against the claim of continuity, and explain why answers to the theorist’s questions are vulnerable to the calibration objection in a way that our non-philosophical, everyday employment of the same words is not. This then leads me to question another widespread assumption, which is the assumption that the theorist’s questions—as raised in the theorist’s context—have clear enough sense and correct answers. I end by responding to a series of objections to my argument and to OLP more generally. (shrink)

This paper explores the history of ELIZA, a computer programme approximating a Rogerian therapist, developed by Jospeh Weizenbaum at MIT in the 1970s, as an early AI experiment. ELIZA’s reception provoked Weizenbaum to re-appraise the relationship between ‘computer power and human reason’ and to attack the ‘powerful delusional thinking’ about computers and their intelligence that he understood to be widespread in the general public and also amongst experts. The root issue for Weizenbaum was whether human thought could be ‘entirely computable’. (...) This also provoked him to re-consider the nature of machine intelligence and to question the instantiation of its logics in the social world, which would come to operate, he said, as a ‘slow acting poison’. Exploring Weizenbaum’s 20th Century apostasy, in the light of ELIZA, illustrates ways in which contemporary anxieties and debates over machine smartness connect to earlier formations. In particular, this article argues that it is in its designation as a computational therapist that ELIZA is most significant today. ELIZA points towards a form of human–machine relationship now pervasive, a precursor of the ‘machinic therapeutic’ condition we find ourselves in, and thus speaks very directly to questions concerning modulation, autonomy, and the new behaviorism that are currently arising. (shrink)

Those who believe that explicitly Christian thinking is possible in the scientific disciplines tend to assume that it must be antithetical to the world’s thinking. Based on some of the author’s experience, this article examines a different approach, in which Christian thinking is used to account for and enrich the world’s thinking by transplanting it from its current ground-motive into the arguably more fertile soil of the creation-fall-redemption ground-motive. The article shows how Dooyeweerd’s version of Christian thinking has been employed (...) in two areas of thinking in information systems : thinking about the nature of computers and information, with the artificial intelligence question of whether computer is like human being soft systems methodology, by which perspectives on ‘human activity systems’ are orchestrated into new learning and plans. In both areas, the original ideas are accounted for, given philosophical underpinning, reinterpreted and enriched. These two show that Dooyeweerd’s philosophy can be equally useful in thinking grounded in both positivist and interpretivist cultures. (shrink)

Huxley’s Explanatory Gap: There can be no explanation of how states of consciousness arise from interaction among a collection of physical things.

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)

Modern biology has not yet come to terms with the presence of many organic codes in Nature, despite the fact that we can prove their existence. As a result, it has not yet accepted the idea that the great events of macroevolution were associated with the origin of new organic codes, despite the fact that this is the most parsimonious and logical explanation of those events. This is probably due to the fact that the existence of organic codes in all (...) fundamental processes of life, and in all major transitions in the history of life, has enormous theoretical implications. It requires nothing less than a new theoretical framework, and that kind of change is inevitably slow. There are too many facts to reconsider, too many bits of history to weave together in a new mosaic. But this is what science is about, and the purpose of the present paper is to show that it can be done. More precisely, it is shown that the whole natural history of the brain can be revisited in the light of the organic codes. What is described here is only a bird’s-eye view of brain macroevolution, but it is hoped that the extraordinary potential of the organic codes can nevertheless come through. The paper contains also another message. The organic codes prove that life is based on semiosis, and are in fact the components of organic semiosis, the first and the most diffused form of semiosis on Earth, but not the only one. It will be shown that the evolution of the brain was accompanied by the development of two new types of sign processes. More precisely, it gave origin first to interpretive semiosis, mostly in vertebrates, and then to cultural semiosis, in our species. (shrink)

Thirty years ago, grounded cognition had roots in philosophy, perception, cognitive linguistics, psycholinguistics, cognitive psychology, and cognitive neuropsychology. During the next 20 years, grounded cognition continued developing in these areas, and it also took new forms in robotics, cognitive ecology, cognitive neuroscience, and developmental psychology. In the past 10 years, research on grounded cognition has grown rapidly, especially in cognitive neuroscience, social neuroscience, cognitive psychology, social psychology, and developmental psychology. Currently, grounded cognition appears to be achieving increased acceptance throughout cognitive (...) science, shifting from relatively minor status to increasing importance. Nevertheless, researchers wonder whether grounded mechanisms lie at the heart of the cognitive system or are peripheral to classic symbolic mechanisms. Although grounded cognition is currently dominated by demonstration experiments in the absence of well-developed theories, the area is likely to become increasingly theory driven over the next 30 years. Another likely development is the increased incorporation of grounding mechanisms into cognitive architectures and into accounts of classic cognitive phenomena. As this incorporation occurs, much functionality of these architectures and phenomena is likely to remain, along with many original mechanisms. Future theories of grounded cognition are likely to be heavily influenced by both cognitive neuroscience and social neuroscience, and also by developmental science and robotics. Aspects from the three major perspectives in cognitive science—classic symbolic architectures, statistical/dynamical systems, and grounded cognition—will probably be integrated increasingly in future theories, each capturing indispensable aspects of intelligence. (shrink)

Very plausibly, nothing can be a genuine computing system unless it meets an input-sensitivity requirement. Otherwise all sorts of objects, such as rocks or pails of water, can count as performing computations, even such as might suffice for mentality—thus threatening computationalism about the mind with panpsychism. Maudlin in J Philos 86:407–432, ( 1989 ) and Bishop ( 2002a , b ) have argued, however, that such a requirement creates difficulties for computationalism about conscious experience, putting it in conflict with the (...) very intuitive thesis that conscious experience supervenes on physical activity. Klein in Synthese 165:141–153, ( 2008 ) proposes a way for computationalists about experience to avoid panpsychism while still respecting the supervenience of experience on activity. I argue that his attempt to save computational theories of experience from Maudlin’s and Bishop’s critique fails. (shrink)

The paper is a commentary on the target article by Christine A. Skarda & Walter J. Freeman, “How brains make chaos in order to make sense of the world”, in the same issue of the journal, pp.161–195. -/- I confine my comments largely to some philosophical claims that Skarda & Freeman make and to the relationship of their model to connectionism. Some of the comments hinge on what symbols are and how they might sit in neural systems.