Recently, there has been considerable interest in large language models: machine learning systems which produce human-like text and dialogue. Applications of these systems have been plagued by persistent inaccuracies in their output; these are often called “AI hallucinations”. We argue that these falsehoods, and the overall activity of large language models, is better understood as bullshit in the sense explored by Frankfurt (On Bullshit, Princeton, 2005): the models are in an important way indifferent to the truth of their (...) outputs. We distinguish two ways in which the models can be said to be bullshitters, and argue that they clearly meet at least one of these definitions. We further argue that describing AI misrepresentations as bullshit is both a more useful and more accurate way of predicting and discussing the behaviour of these systems. (shrink)

Until recently, cognitive science focused on such mental functions as problem solving, grammar, and pattern-the functions in which the human mind most closely resembles a computer. But humans are more than computers: we invent new meanings, imagine wildly, and even have ideas that have never existed before. Today the cutting edge of cognitive science addresses precisely these mysterious, creative aspects of the mind.The Way We Think is a landmark analysis of the imaginative nature of the mind. Conceptual (...) blending is already widely known in research laboratories throughout the world; this book, written to be accessible to both lay readers and interested scientists, is its definitive statement. Gilles Fauconnier and Mark Turner show that conceptual blending is the root of the cognitively modern human mind, and that conceptual blends themselves are continually combined and reblended to create the rich mental fabric in which we live.The Way We Think shows how this blending operates; how it is affected by (and gives rise to) language, identity, culture, and invention; and how we imagine what could be and what might have been. The result is a bold and exciting new view of how the mind works. (shrink)

In the era of “big data,” science is increasingly information driven, and the potential for computers to store, manage, and integrate massive amounts of data has given rise to such new disciplinary fields as biomedical informatics. Applied ontology offers a strategy for the organization of scientific information in computer-tractable form, drawing on concepts not only from computer and information science but also from linguistics, logic, and philosophy. This book provides an introduction to the field of applied (...) ontology that is of particular relevance to biomedicine, covering theoretical components of ontologies, best practices for ontology design, and examples of biomedical ontologies in use. After defining an ontology as a representation of the types of entities in a given domain, the book distinguishes between different kinds of ontologies and taxonomies, and shows how applied ontology draws on more traditional ideas from metaphysics. It presents the core features of the Basic Formal Ontology (BFO), now used by over one hundred ontology projects around the world, and offers examples of domain ontologies that utilize BFO. The book also describes Web Ontology Language (OWL), a common framework for Semantic Web technologies. Throughout, the book provides concrete recommendations for the design and construction of domain ontologies. (shrink)

Extract from Hofstadter's revew in Bulletin of American Mathematical Society : http://www.ams.org/journals/bull/1980-02-02/S0273-0979-1980-14752-7/S0273-0979-1980-14752-7.pdf -/- "Aaron Sloman is a man who is convinced that most philosophers and many other students of mind are in dire need of being convinced that there has been a revolution in that field happening right under their noses, and that they had better quickly inform themselves. The revolution is called "Artificial Intelligence" (Al)-and Sloman attempts to impart to others the "enlighten- ment" which he clearly regrets not having (...) experienced earlier himself. Being somewhat of a convert, Sloman is a zealous campaigner for his point of view. Now a Reader in Cognitive Science at Sussex, he began his academic career in more orthodox philosophy and, by exposure to linguistics and AI, came to feel that all approaches to mind which ignore AI are missing the boat. I agree with him, and I am glad that he has written this provocative book. The tone of Sloman's book can be gotten across by this quotation (p. 5): "I am prepared to go so far as to say that within a few years, if there remain any philosophers who are not familiar with some of the main developments in artificial intelligence, it will be fair to accuse them of professional incom- petence, and that to teach courses in philosophy of mind, epistemology, aesthetics, philosophy of science, philosophy of language, ethics, metaphysics, and other main areas of philosophy, without discussing the relevant aspects of artificial intelligence will be as irresponsible as giving a degree course in physics which includes no quantum theory." -/- (The author now regrets the extreme polemical tone of the book.). (shrink)

Often machine learning programs inherit social patterns reflected in their training data without any directed effort by programmers to include such biases. Computer scientists call this algorithmic bias. This paper explores the relationship between machine bias and human cognitive bias. In it, I argue similarities between algorithmic and cognitive biases indicate a disconcerting sense in which sources of bias emerge out of seemingly innocuous patterns of information processing. The emergent nature of this bias obscures the existence of the bias (...) itself, making it difficult to identify, mitigate, or evaluate using standard resources in epistemology and ethics. I demonstrate these points in the case of mitigation techniques by presenting what I call ‘the Proxy Problem’. One reason biases resist revision is that they rely on proxy attributes, seemingly innocuous attributes that correlate with socially-sensitive attributes, serving as proxies for the socially-sensitive attributes themselves. I argue that in both human and algorithmic domains, this problem presents a common dilemma for mitigation: attempts to discourage reliance on proxy attributes risk a tradeoff with judgement accuracy. This problem, I contend, admits of no purely algorithmic solution. (shrink)

This paper analyses the phenomenology and epistemology of chatbots such as ChatGPT and Bard. The computational architecture underpinning these chatbots are large language models (LLMs), which are generative artificial intelligence (AI) systems trained on a massive dataset of text extracted from the Web. We conceptualise these LLMs as multifunctional computational cognitive artifacts, used for various cognitive tasks such as translating, summarizing, answering questions, information-seeking, and much more. Phenomenologically, LLMs can be experienced as a “quasi-other”; when that happens, users anthropomorphise (...) them. For most users, current LLMs are black boxes, i.e., for the most part, they lack data transparency and algorithmic transparency. They can, however, be phenomenologically and informationally transparent, in which case there is an interactional flow. Anthropomorphising and interactional flow can, in some users, create an attitude of (unwarranted) trust towards the output LLMs generate. We conclude this paper by drawing on the epistemology of trust and testimony to examine the epistemic implications of these dimensions. Whilst LLMs generally generate accurate responses, we observe two epistemic pitfalls. Ideally, users should be able to match the level of trust that they place in LLMs to the degree that LLMs are trustworthy. However, both their data and algorithmic opacity and their phenomenological and informational transparency can make it difficult for users to calibrate their trust correctly. The effects of these limitations are twofold: users may adopt unwarranted attitudes of trust towards the outputs of LLMs (which is particularly problematic when LLMs hallucinate), and the trustworthiness of LLMs may be undermined. (shrink)

Large language models such as ChatGPT enable users to automatically produce text but also raise ethical concerns, for example about authorship and deception. This paper analyses and discusses some key philosophical assumptions in these debates, in particular assumptions about authorship and language and—our focus—the use of the appearance/reality distinction. We show that there are alternative views of what goes on with ChatGPT that do not rely on this distinction. For this purpose, we deploy the two phased approach of (...) deconstruction and relate our finds to questions regarding authorship and language in the humanities. We also identify and respond to two common counter-objections in order to show the ethical appeal and practical use of our proposal. (shrink)

A number of recent discussions comparing computer simulation and traditional experimentation have focused on the significance of “materiality.” I challenge several claims emerging from this work and suggest that computer simulation studies are material experiments in a straightforward sense. After discussing some of the implications of this material status for the epistemology of computer simulation, I consider the extent to which materiality (in a particular sense) is important when it comes to making justified inferences about target systems (...) on the basis of experimental results. (shrink)

This powerfully iconoclastic book reconsiders the influential nativist position toward the mind. Nativists assert that some concepts, beliefs, or capacities are innate or inborn: "native" to the mind rather than acquired. Fiona Cowie argues that this view is mistaken, demonstrating that nativism is an unstable amalgam of two quite different--and probably inconsistent--theses about the mind. Unlike empiricists, who postulate domain-neutral learning strategies, nativists insist that some learning tasks require special kinds of skills, and that these skills are hard-wired into our (...) brains at birth. This "faculties hypothesis" finds its modern expression in the views of Noam Chomsky. Cowie, marshaling recent empirical evidence from developmental psychology, psycholinguistics, computer science, and linguistics, provides a crisp and timely critique of Chomsky's nativism and defends in its place a moderately nativist approach to language acquisition. Also in contrast to empiricists, who view the mind as simply another natural phenomenon susceptible of scientific explanation, nativists suspect that the mental is inelectably mysterious. Cowie addresses this second strand in nativist thought, taking on the view articulated by Jerry Fodor and other nativists that learning, particularly concept acquisition, is a fundamentally inexplicable process. Cowie challenges this explanatory pessimism, and argues convincingly that concept acquisition is psychologically explicable. What's Within? is a clear and provocative achievement in the study of the human mind. (shrink)

Modern modal logic originated as a branch of philosophical logic in which the concepts of necessity and possibility were investigated by means of a pair of dual operators that are added to a propositional or first-order language. The field owes much of its flavor and success to the introduction in the 1950s of the “possible-worlds” semantics in which the modal operators are interpreted via some “accessibility relation” connecting possible worlds. In subsequent years, modal logic has received attention as an (...) attractive approach towards formalizing such diverse notions as time, knowledge, or action. Nowadays, modal logics are applied in various disciplines, ranging from economics to linguistics and computer science. Consequently, there is by now a large variety of modal languages, with an even greater wealth of interpretations. For instance, many applications require a poly-modal framework consisting of a language with a family of modal operators and a semantics in which the corresponding accessibility relations are connected somehow. (shrink)

It is the aim of INDICES to document recent explorations in the various fields of philosophical logic and formal linguistics and their applications in other disciplines. The main emphasis of this series is on self-contained monographs covering particular areas of recent research and surveys of methods, problems, and results in all fields of inquiry where recourse to logical analysis and logical methods has been fruitful. INDICES will contain monographs dealing with the central areas of philosophical logic (extensional and intensional systems, (...) indexical logics, non-classical logics, philosophy of logic, etc.) as well as studies in which these systems are applied to specific issues in philosophy, in the formal semantics of natural languages, the foundations of linguistic theory, in computational linguistics, and in theoretical computer science. Constructive type theory was first presented in 1970, by the Swedish logician Per Martin-Lof. It has become one of the main approaches used in the foundations of mathematics and computer science. But it has remained relatively unknown among linguists and philosophers, although it provides a considerable extension of the concepts and techniques of logic. The book first gives an introduction to type theory from the point of view of linguistics and the philosophy of language. Type theory is then applied in the areas of quantification, anaphora, temporal reference, and the structure of text and discourse. By virtue of the type-theoretical concepts of proof object and context, various phenomena of dependence and progression in language can be discussed in precise terms, and several well-known problems can be solved. A categorial grammar is presented togenerate formally a fragment of English, together with an example of a computer implementation. (shrink)

This paper examines the role of mathematical idealization in describing and explaining various features of the world. It examines two cases: first, briefly, the modeling of shock formation using the idealization of the continuum. Second, and in more detail, the breaking of droplets from the points of view of both analytic fluid mechanics and molecular dynamical simulations at the nano-level. It argues that the continuum idealizations are explanatorily ineliminable and that a full understanding of certain physical phenomena cannot be obtained (...) through completely detailed, non-idealized representations. (shrink)

Deductive inference is usually regarded as being “tautological” or “analytical”: the information conveyed by the conclusion is contained in the information conveyed by the premises. This idea, however, clashes with the undecidability of first-order logic and with the (likely) intractability of Boolean logic. In this article, we address the problem both from the semantic and the proof-theoretical point of view. We propose a hierarchy of propositional logics that are all tractable (i.e. decidable in polynomial time), although by means of growing (...) computational resources, and converge towards classical propositional logic. The underlying claim is that this hierarchy can be used to represent increasing levels of “depth” or “informativeness” of Boolean reasoning. Special attention is paid to the most basic logic in this hierarchy, the pure “intelim logic”, which satisfies all the requirements of a natural deduction system (allowing both introduction and elimination rules for each logical operator) while admitting of a feasible (quadratic) decision procedure. We argue that this logic is “analytic” in a particularly strict sense, in that it rules out any use of “virtual information”, which is chiefly responsible for the combinatorial explosion of standard classical systems. As a result, analyticity and tractability are reconciled and growing degrees of computational complexity are associated with the depth at which the use of virtual information is allowed. (shrink)

Computer simulations are an exciting tool that plays important roles in many scientific disciplines. This has attracted the attention of a number of philosophers of science. The main tenor in this literature is that computer simulations not only constitute interesting and powerful new science , but that they also raise a host of new philosophical issues. The protagonists in this debate claim no less than that simulations call into question our philosophical understanding of scientific ontology, the (...) epistemology and semantics of models and theories, and the relation between experimentation and theorising, and submit that simulations demand a fundamentally new philosophy of science in many respects. The aim of this paper is to critically evaluate these claims. Our conclusion will be sober. We argue that these claims are overblown and that simulations, far from demanding a new metaphysics, epistemology, semantics and methodology, raise few if any new philosophical problems. The philosophical problems that do come up in connection with simulations are not specific to simulations and most of them are variants of problems that have been discussed in other contexts before. (shrink)

Temporal Logic: From Ancient Ideas to Artificial Intelligence deals with the history of temporal logic as well as the crucial systematic questions within the field. The book studies the rich contributions from ancient and medieval philosophy up to the downfall of temporal logic in the Renaissance. The modern rediscovery of the subject, which is especially due to the work of A. N. Prior, is described, leading into a thorough discussion of the use of temporal logic in computer science (...) and the understanding of natural language. Temporal Logic: From Ancient Ideas to Artificial Intelligence thus interweaves linguistic, philosophical and computational aspects into an informative and inspiring whole. (shrink)

This paper shares results from a pedagogical experiment that assigns undergraduates to “cheat” on a final class essay by requiring their use of text-generating AI software. For this assignment, students harvested content from an installation of GPT-2, then wove that content into their final essay. At the end, students offered a “revealed” version of the essay as well as their own reflections on the experiment. In this assignment, students were specifically asked to confront the oncoming availability of AI as a (...) writing tool. What are the ethics of using AI this way? What counts as plagiarism? What are the conditions, if any, we should place on AI assistance for student writing? And how might working with AI change the way we think about writing, authenticity, and creativity? While students (and sometimes GPT-2) offered thoughtful reflections on these initial questions, actually composing with GPT-2 opened their perspectives more broadly on the ethics and practice of writing with AI. In this paper, I share how students experienced those issues, connect their insights to broader conversations in the humanities about writing and communication, and explain their relevance for the ethical use and evaluation of language models. (shrink)

This is the first comprehensive textbook on higher order logic that is written specifically to introduce the subject matter to graduate students in philosophy. The book covers both the formal aspects of higher-order languages -- their model theory and proof theory, the theory of λ-abstraction and its generalizations -- and their philosophical applications, especially to the topics of modality and propositional granularity. The book has a strong focus on non-extensional higher-order logics, making it more appropriate for foundational metaphysics than other (...) introductions to the subject from computer science, mathematics, and linguistics. A Philosophical Introduction to Higher Order Logics assumes only that readers have a basic knowledge of first-order logic. With an emphasis on exercises, it can be used as a textbook though is also ideal for self-study. Author Andrew Bacon organizes the book's 18 chapters around four main parts: I. Typed Language II. Higher Order Languages III. General Higher-Order Languages IV. Higher-Order Model Theory In addition, two appendices cover the Curry-Howard isomorphism and its applications for modeling propositional structure. Each chapter includes exercises that move from easier to more difficult, strategically placed throughout the chapter, and concludes with an annotated suggested reading list providing graduate students with most valuable additional resources. Key Features Is the first comprehensive introduction to higher-order logic as a grounding for addressing problems in metaphysics Introduces the basic formal tools that are needed to theorize in, and model, higher order languages Offers an abundance of: - Simple exercises throughout the book, serving as comprehension checks on basic concepts and definitions - More difficult exercises designed to facilitate long-term learning Contains annotated sections on further reading, pointing the reader to related literature, learning resources, and historical context. (shrink)

While the notion of the mind as information-processor--a kind of computational system--is widely accepted, many scientists and philosophers have assumed that this account of cognition shows that the mind's operations are characterizable independent of their relationship to the external world. Existential Cognition challenges the internalist view of mind, arguing that intelligence, thought, and action cannot be understood in isolation, but only in interaction with the outside world. Arguing that the mind is essentially embedded in the external world, Ron McClamrock provides (...) a schema that allows cognitive scientists to address such long-standing problems in artificial intelligence as the "frame" problem and the issue of "bounded" rationality. Extending this schema to cover progress in other studies of behavior, including language, vision, and action, McClamrock reinterprets the importance of the organism/environment distinction. McClamrock also considers the broader philosophical question of the place of mind in the world, particularly with regard to questions of intentionality, subjectivity, and phenomenology. With implications for philosophy, cognitive and computer science, AI, and psychology, this book synthesizes state-of-the-art work in philosophy and cognitive science on how the mind interacts with the world to produce thoughts, ideas, and actions. (shrink)

As machine learning has gradually entered into ever more sectors of public and private life, there has been a growing demand for algorithmic explainability. How can we make the predictions of complex statistical models more intelligible to end users? A subdiscipline of computer science known as interpretable machine learning (IML) has emerged to address this urgent question. Numerous influential methods have been proposed, from local linear approximations to rule lists and counterfactuals. In this article, I highlight three conceptual (...) challenges that are largely overlooked by authors in this area. I argue that the vast majority of IML algorithms are plagued by (1) ambiguity with respect to their true target; (2) a disregard for error rates and severe testing; and (3) an emphasis on product over process. Each point is developed at length, drawing on relevant debates in epistemology and philosophy of science. Examples and counterexamples from IML are considered, demonstrating how failure to acknowledge these problems can result in counterintuitive and potentially misleading explanations. Without greater care for the conceptual foundations of IML, future work in this area is doomed to repeat the same mistakes. (shrink)

Modeling and computer simulations, we claim, should be considered core philosophical methods. More precisely, we will defend two theses. First, philosophers should use simulations for many of the same reasons we currently use thought experiments. In fact, simulations are superior to thought experiments in achieving some philosophical goals. Second, devising and coding computational models instill good philosophical habits of mind. Throughout the paper, we respond to the often implicit objection that computer modeling is “not philosophical.”.

Call a bit of scientific discourse a description of a missing system when (i) it has the surface appearance of an accurate description of an actual, concrete system (or kind of system) from the domain of inquiry, but (ii) there are no actual, concrete systems in the world around us fitting the description it contains, and (iii) that fact is recognised from the outset by competent practitioners of the scientific discipline in question. Scientific textbooks, classroom lectures, and journal articles abound (...) with such passages; and there is a widespread practice of talking and thinking as though there are systems which fit the descriptions they contain perfectly, despite the recognition that no actual, concrete systems do so—call this the face value practice. There are, furthermore, many instances in which philosophers engage in the face value practice whilst offering answers to epistemological and methodological questions about the sciences. Three questions, then: (1) How should we interpret descriptions of missing systems? (2) How should we make sense of the face value practice? (3) Is there a set of plausible answers to (1) and (2) which legitimates reliance on the face value practice in our philosophical work, and can support the weight of the accounts which are entangled with that practice? In this paper I address these questions by considering three answers to the first: that descriptions of missing systems are straightforward descriptions of abstract objects, that they are indirect descriptions of “property-containing” abstracta, and that they are (in a different way) indirect descriptions of mathematical structures. All three proposals are present in the literature, but I find them wanting. The result is to highlight the importance of developing a satisfactory understanding of descriptions of missing systems and the face value practice, to put pressure on philosophical accounts which rely on the practice, and to help us assess the viability of certain approaches to thinking about models, theory structure, and scientific representation. (shrink)

In this paper I defend the theory that knowledge is credit-worthy true belief against a family of objections, one of which was leveled against it in a recent paper by Jennifer Lackey. In that paper, Lackey argues that testimonial knowledge is problematic for the credit-worthiness theory because when person A comes to know that p by way of the testimony of person B, it would appear that any credit due to A for coming to believe truly that p belongs to (...) the testifier, B, rather than the hearer, A. If so, then knowledge would appear not to be a matter of credit for true belief. I think that the problem this raises actually has little to do with the fact that the knowledge comes by way of testimony, and that similar objections can be formulated in terms of perceptual and memorial knowledge. I will attempt to neutralize these objections by drawing a distinction between credit as praiseworthiness and credit as attributability. (shrink)

In books such as The World Within the World and The Anthropic Cosmological Principle, astronomer John Barrow has emerged as a leading writer on our efforts to understand the universe. Timothy Ferris, writing in The Times Literary Supplement of London, described him as "a temperate and accomplished humanist, scientist, and philosopher of science--a man out to make a contribution, not a show." Now Barrow offers the general reader another fascinating look at modern physics, as he explores the quest for (...) a single, unifying theory that will unlock nature's secrets. Theories of Everything is more than a history of science, more than a popular report on recent research and discoveries. Barrow provides a reflective, intelligent commentary on what a true Theory of Everything would be--its ingredients, its limitations, and what it could tell us about the universe. Never before, he writes, have physicists been so confident and so eager in the hunt for this "cosmic Rosetta Stone," as he calls it: "a single all-embracing picture of all the laws of nature from which the inevitability of all things seen must follow with unimpeachable logic." He lays out eight essential ingredients for a Theory of Everything and then explores each in turn, tracing how our knowledge has developed and how scientific discovery relates to our changing philosophy and religious thought in each area. Some of these ingredients are obvious--the laws of nature must be explained, for example, as well as its organizing principles--but others may be surprising, such as broken symmetries and selection biases. A Theory of Everything must account for the fact that the universe is "messy and complicated," he tells us, and for the limitations imposed by the questions we ask and the information we can obtain. The key lies in the remarkable capacity of mathematics to express the fundamental workings of the physical world--a language that the human mind is uniquely equipped to understand and manipulate. Barrow examines what mathematics actually is and describes how it makes the universe intelligible and provides a path to the underlying coherence in nature--which has led, in fact, to arguments that the universe itself is a vast computer. Yet even the most complete theory, even the most comprehensive mathematical explanation, cannot account for the uncomputable varieties of human experience and thought. "No non-poetic account of reality," he writes, "can be complete." In a field where the authorities converse in equations and mathematical notations, John Barrow speaks with the voice of thoughtful and knowledgeable humanist. Written with eloquence and expertise, Theories of Everything establishes a new perspective on humanity's efforts to explain the universe. (shrink)

Recent artificial intelligence advances, especially those of large language models (LLMs), have increasingly shown glimpses of human-like intelligence. This has led to bold claims that these systems are no longer a mere “it” but now a “who,” a kind of person deserving respect. In this paper, I argue that this view depends on a Cartesian account of personhood, on which identifying someone as a person is based on their cognitive sophistication and ability to address common-sense reasoning problems. I contrast (...) this with a different account of personhood, one where an agent is a person if they are autonomous, responsive to norms, and culpable for their actions. On this latter account, I show that LLMs are not person-like, as evidenced by their propensity for dishonesty, inconsistency, and offensiveness. Moreover, I argue current LLMs, given the way they are designed and trained, cannot be persons—either social or Cartesian. The upshot is that contemporary LLMs are not, and never will be, persons. (shrink)

Recent work in the philosophy of science has generated an apparent conflict between theories attempting to explicate the nature of scientific representation. On one side, there are what one might call 'informational' views, which emphasize objective relations (such as similarity, isomorphism, and homomorphism) between representations (theories, models, simulations, diagrams, etc.) and their target systems. On the other side, there are what one might call 'functional' views, which emphasize cognitive activities performed in connection with these targets, such as interpretation and (...) inference. The main sources of the impression of conflict here are arguments by some functionalists to the effect that informational theories are flawed: it is suggested that relations typically championed by informational theories are neither necessary nor sufficient for scientific representation, and that any theory excluding functions is inadequate. In this paper I critically examine these arguments, and contend that, as it turns out, informational and functional theories are importantly complementary. (shrink)

In this paper, I distinguish scientific models in three kinds on the basis of their ontological status—material models, mathematical models and fictional models, and develop and defend an account of fictional models as fictional objects—i.e. abstract objects that stand for possible concrete objects.

The descriptions and theoretical laws scientists write down when they model a system are often false of any real system. And yet we commonly talk as if there were objects that satisfy the scientists’ assumptions and as if we may learn about their properties. Many attempt to make sense of this by taking the scientists’ descriptions and theoretical laws to define abstract or fictional entities. In this paper, I propose an alternative account of theoretical modelling that draws upon Kendall Walton’s (...) ‘make-believe’ theory of representation in art. I argue that this account allows us to understand theoretical modelling without positing any object of which scientists’ modelling assumptions are true. (shrink)

This book takes concepts developed by researchers in theoretical computer science and adapts and applies them to the study of natural language meaning. Summarizing over a decade of research, Chris Barker and Chung-chieh Shan put forward the Continuation Hypothesis: that the meaning of a natural language expression can depend on its own continuation.

From the late 1950s until 1975, cognition was understood mainly as disembodied symbol manipulation in cognitive psychology, linguistics, artificial intelligence, and the nascent field of Cognitive Science. The idea of embodied cognition entered the field of Cognitive Linguistics at its beginning in 1975. Since then, cognitive linguists, working with neuroscientists, computer scientists, and experimental psychologists, have been developing a neural theory of thought and language (NTTL). Central to NTTL are the following ideas: (a) we think with our (...) brains, that is, thought is physical and is carried out by functional neural circuitry; (b) what makes thought meaningful are the ways those neural circuits are connected to the body and characterize embodied experience; (c) so-called abstract ideas are embodied in this way as well, as is language. Experimental results in embodied cognition are seen not only as confirming NTTL but also explained via NTTL, mostly via the neural theory of conceptual metaphor. Left behind more than three decades ago is the old idea that cognition uses the abstract manipulation of disembodied symbols that are meaningless in themselves but that somehow constitute internal “representations of external reality” without serious mediation by the body and brain. This article uniquely explains the connections between embodied cognition results since that time and results from cognitive linguistics, experimental psychology, computational modeling, and neuroscience. (shrink)

This volume on the semantic complexity of natural language explores the question why some sentences are more difficult than others. While doing so, it lays the groundwork for extending semantic theory with computational and cognitive aspects by combining linguistics and logic with computations and cognition. -/- Quantifier expressions occur whenever we describe the world and communicate about it. Generalized quantifier theory is therefore one of the basic tools of linguistics today, studying the possible meanings and the inferential power of (...) quantifier expressions by logical means. The classic version was developed in the 1980s, at the interface of linguistics, mathematics and philosophy. Before this volume, advances in "classic" generalized quantifier theory mainly focused on logical questions and their applications to linguistics, this volume adds a computational component, the third pillar of language use and logical activity. This book is essential reading for researchers in linguistics, philosophy, cognitive science, logic, AI, and computer science. (shrink)

We discuss several features of coherent choice functions—where the admissible options in a decision problem are exactly those that maximize expected utility for some probability/utility pair in fixed set S of probability/utility pairs. In this paper we consider, primarily, normal form decision problems under uncertainty—where only the probability component of S is indeterminate and utility for two privileged outcomes is determinate. Coherent choice distinguishes between each pair of sets of probabilities regardless the “shape” or “connectedness” of the sets of probabilities. (...) We axiomatize the theory of choice functions and show these axioms are necessary for coherence. The axioms are sufficient for coherence using a set of probability/almost-state-independent utility pairs. We give sufficient conditions when a choice function satisfying our axioms is represented by a set of probability/state-independent utility pairs with a common utility. (shrink)

Zwart and Franssen’s impossibility theorem reveals a conflict between the possible-world-based content-definition and the possible-world-based likeness-definition of verisimilitude. In Sect. 2 we show that the possible-world-based content-definition violates four basic intuitions of Popper’s consequence-based content-account to verisimilitude, and therefore cannot be said to be in the spirit of Popper’s account, although this is the opinion of some prominent authors. In Sect. 3 we argue that in consequence-accounts, content-aspects and likeness-aspects of verisimilitude are not in conflict with each other, but in (...) agreement. We explain this fact by pointing towards the deep difference between possible-world- and the consequence-accounts, which does not lie in the difference between syntactic versus semantic formulations, but in the difference between ‘disjunction-of-possible-worlds’ versus ‘conjunction-of-parts’ representations of theories. Drawing on earlier work, we explain in Sect. 4 how the shortcomings of Popper’s original definition can be repaired by what we call the relevant element approach. We propose a quantitative likeness-definition of verisimilitude based on relevant elements which provably agrees with the qualitative relevant content-definition of verisimilitude on all pairs of comparable theories. We conclude the paper with a plea for consequence-accounts and a brief analysis of the problem of language-dependence. (shrink)

AI researchers have developed sophisticated language models capable of generating paragraphs of 'synthetic text' on topics specified by the user. While AI text generation has legitimate benefits, it could also be misused, potentially to grave effect. For example, AI text generators could be used to automate the production of convincing fake news, or to inundate social media platforms with machine-generated disinformation. This paper argues that AI text generators should be conceptualised as a dual-use technology, outlines some relevant lessons from (...) earlier debates on dual-use life sciences research, and calls for closer collaboration between ethicists and the machine learning community to address AI language models’ dual-use implications. (shrink)

As China and the United States strive to be the primary global leader in AI, their visions are coming into conflict. This is frequently painted as a fundamental clash of civilisations, with evidence based primarily around each country’s current political system and present geopolitical tensions. However, such a narrow view claims to extrapolate into the future from an analysis of a momentary situation, ignoring a wealth of historical factors that influence each country’s prevailing philosophy of technology and thus their overarching (...) AI strategies. In this article, we build a philosophy-of-technology-grounded framework to analyse what differences in Chinese and American AI policies exist and, on a fundamental level, why they exist. We support this with Natural Language Processing methods to provide an evidentiary basis for our analysis of policy differences. By looking at documents from three different American presidential administrations––Barack Obama, Donald Trump, and Joe Biden––as well as both national and local policy documents (many available only in Chinese) from China, we provide a thorough comparative analysis of policy differences. This article fills a gap in US–China AI policy comparison and constructs a framework for understanding the origin and trajectory of policy differences. By investigating what factors are informing each country’s philosophy of technology and thus their overall approach to AI policy, we argue that while significant obstacles to cooperation remain, there is room for dialogue and mutual growth. (shrink)

Why is a red face not really red? How do we decide that this book is a textbook or not? Conceptual spaces provide the medium on which these computations are performed, but an additional operation is needed: Contrast. By contrasting a reddish face with a prototypical face, one gets a prototypical ‘red’. By contrasting this book with a prototypical textbook, the lack of exercises may pop out. Dynamic contrasting is an essential operation for converting perceptions into predicates. The existence of (...) dynamic contrasting may contribute to explaining why lexical meanings correspond to convex regions of conceptual spaces. But it also explains why predication is most of the time opportunistic, depending on context. While off-line conceptual similarity is a holistic operation, the contrast operation provides a context-dependent distance that creates ephemeral predicative judgments that are essential for interfacing conceptual spaces with natural language and with reasoning. (shrink)

Much of our behavior is guided by our understanding of events. We perceive events when we observe the world unfolding around us, participate in events when we act on the world, simulate events that we hear or read about, and use our knowledge of events to solve problems. In this book, Gabriel A. Radvansky and Jeffrey M. Zacks provide the first integrated framework for event cognition and attempt to synthesize the available psychological and neuroscience data surrounding it. This synthesis leads (...) to new proposals about several traditional areas in psychology and neuroscience including perception, attention, language understanding, memory, and problem solving.Radvansky and Zacks have written this book with a diverse readership in mind. It is intended for a range of researchers working within cognitive science including psychology, neuroscience, computer science, philosophy, anthropology, and education. Readers curious about events more generally such as those working in literature, film theory, and history will also find it of interest. (shrink)

This paper discusses the neo-logicist approach to the foundations of mathematics by highlighting an issue that arises from looking at the Bad Company objection from an epistemological perspective. For the most part, our issue is independent of the details of any resolution of the Bad Company objection and, as we will show, it concerns other foundational approaches in the philosophy of mathematics. In the first two sections, we give a brief overview of the "Scottish" neo-logicist school, present a generic form (...) of the Bad Company objection and introduce an epistemic issue connected to this general problem that will be the focus of the rest of the paper. In the third section, we present an alternative approach within philosophy of mathematics, a view that emerges from Hilbert's Grundlagen der Geometrie (1899, Leipzig: Teubner; Foundations of geometry (trans.: Townsend, E.). La Salle, Illinois: Open Court, 1959.). We will argue that Bad Company-style worries, and our concomitant epistemic issue, also affects this conception and other foundationalist approaches. In the following sections, we then offer various ways to address our epistemic concern, arguing, in the end, that none resolves the issue. The final section offers our own resolution which, however, runs against the foundationalist spirit of the Scottish neo-logicist program. (shrink)

Transfer of information between senders and receivers, of one kind or another, is essential to all life. David Lewis introduced a game theoretic model of the simplest case, where one sender and one receiver have pure common interest. How hard or easy is it for evolution to achieve information transfer in Lewis signaling?. The answers involve surprising subtleties. We discuss some if these in terms of evolutionary dynamics in both finite and infinite populations, with and without mutation.

This paper sets out a philosophical interpretation of the model theory of Mares and Goldblatt (The Journal of Symbolic Logic 71, 2006). This interpretation distinguishes between truth conditions and information conditions. Whereas the usual Tarskian truth condition holds for universally quantified statements, their information condition is quite different. The information condition utilizes general propositions . The present paper gives a philosophical explanation of general propositions and argues that these are needed to give an adequate theory of general information.

Filling the need for an accessible, carefully structured introductory text in symbolic logic, Modern Logic has many features designed to improve students' comprehension of the subject, including a proof system that is the same as the award-winning computer program MacLogic, and a special appendix that shows how to use MacLogic as a teaching aid. There are graded exercises at the end of each chapter--more than 900 in all--with selected answers at the end of the book. Unlike competing texts, Modern (...) Logic gives equal weight to semantics and proof theory and explains their relationship, and develops in detail techniques for symbolizing natural language in first-order logic. After a general introduction featuring the notion of logical form, the book offers sections on classical sentential logic, monadic predicate logic, and full first-order logic with identity. A concluding section deals with extensions of and alternatives to classical logic, including modal logic, intuitionistic logic, and fuzzy logic. For students of philosophy, mathematics, computer science, or linguistics, Modern Logic provides a thorough understanding of basic concepts and a sound basis for more advanced work. (shrink)

In contemporary artificial intelligence, the challenge is making intricate connectionist systems—comprising millions of parameters—more comprehensible, defensible, and rationally grounded. Two prevailing methodologies address this complexity. The inaugural approach amalgamates symbolic methodologies with connectionist paradigms, culminating in a hybrid system. This strategy systematizes extensive parameters within a limited framework of formal, symbolic rules. Conversely, the latter strategy remains staunchly connectionist, eschewing hybridity. Instead of internal transparency, it fabricates an external, transparent proxy system. This ancillary system’s mandate is elucidating the principal system’s (...) decisions, essentially approximating its outcomes. Leveraging natural language processing as our analytical lens, this paper elucidates both methodologies: the hybrid method is underscored by the compositional vector semantics, whereas the purely connectionist method evolves as a derivative of neural semantic parsers. This discourse extols the merits of the purely connectionist approach for its inherent flexibility and for a pivotal delineation: segregating the explanatory apparatus from the operational core, thereby rendering artificial intelligence systems reminiscent of human cognition. (shrink)

Newly powerful large language models have burst onto the scene, with applications across a wide range of functions. We can now expect to encounter their outputs at rapidly increasing volumes and frequencies. Some commentators claim that large language models are bullshitting, generating convincing output without regard for the truth. If correct, that would make large language models distinctively dangerous discourse participants. Bullshitters not only undermine the norm of truthfulness (by saying false things) but the normative status of (...) truth itself (by treating it as entirely irrelevant). So, do large language models really bullshit? I argue that they can, in the sense of issuing propositional content in response to fact-seeking prompts, without having first assessed that content for truth or falsity. However, I further argue that they need not bullshit, given appropriate guardrails. So, just as with human speakers, the propensity for a large language model to bullshit depends on its own particular make-up. (shrink)

The aim of this paper is to present a topological method for constructing discretizations of topological conceptual spaces. The method works for a class of topological spaces that the Russian mathematician Pavel Alexandroff defined more than 80 years ago. The aim of this paper is to show that Alexandroff spaces, as they are called today, have many interesting properties that can be used to explicate and clarify a variety of problems in philosophy, cognitive science, and related disciplines. For instance, (...) recently, Ian Rumfitt used a special type of Alexandroff spaces to elucidate the logic of vague concepts in a new way. Moreover, Rumfitt’s class of Alexandroff spaces can be shown to provide a natural topological semantics for Susanne Bobzien’s “logic of clearness”. Mainly due to the work of Peter Gärdenfors and his collaborators, conceptual spaces have become an increasingly popular tool of dealing with a variety of problems in the fields of cognitive psychology, artificial intelligence, linguistics and philosophy. For Gärdenfors’s conceptual spaces, geometrically defined discretizations play an essential role. These tessellations can be shown to be extensionally equivalent to topological tessellations that can be constructed for Alexandroff spaces in general. Thereby, Rumfitt’s and Gärdenfors’s constructions turn out to be special cases of an approach that works for a more general class of spaces, namely, for weakly scattered Alexandroff spaces. The main aim of this paper is to show that this class of spaces provides a convenient framework for conceptual spaces as used in epistemology and related disciplines in general. Weakly scattered Alexandroff spaces are useful for elucidating problems related to the logic of vague concepts, in particular they offer a solution of the Sorites paradox. Further, they provide a semantics for the logic of clearness that overcomes certain problems of the concept of higher-order vagueness. Moreover, these spaces help find a natural place for classical syllogistics in the framework of conceptual spaces. The specialization order of Alexandroff spaces can be used to refine the all-or-nothing distinction between prototypical and non-prototypical stimuli in favor of a gradual distinction between more or less prototypical elements of conceptual spaces. The greater conceptual flexibility of the topological approach helps avoid some inherent inadequacies of the geometrical approach, for instance, the so-called “thickness problem”. Finally, it is shown that the Alexandroff spaces offer an appropriate framework to deal with digital conceptual spaces that are gaining more and more importance in the areas of artificial intelligence, computer science and related disciplines. (shrink)

Recent developments in artificial intelligence based on neural nets—deep learning and large language models which together I refer to as NEWAI—have resulted in startling improvements in language handling and the potential to keep up with changing human knowledge by learning from the internet. Nevertheless, examples such as ChatGPT, which is a ‘large language model’, have proved to have no moral compass: they answer queries with fabrications with the same fluency as they provide facts. I try to explain (...) why this is, basing the argument on the sociology of knowledge, particularly social studies of science, notably ‘studies of expertise and experience’ and the ‘fractal model’ of society. Learning from the internet is not the same as socialisation: NEWAI has no primary socialisation such as provides the foundations of human moral understanding. Instead, large language models are retrospectively socialised by human intervention in an attempt to align them with societally accepted ethics. Perhaps, as technology advances, large language models could come to understand speech and recognise objects sufficiently well to acquire the equivalent of primary socialisation. In the meantime, we must be vigilant about who is socialising them and be aware of the danger of their socialising us to align with them rather than vice-versa, an eventuality that would lead to the further erosion of the distinction between the true and the false giving further support to populism and fascism. (shrink)

In this article I have two primary goals. First, I present two recent views on the distinction between knowledge-that and knowledge-how (Stanley and Williamson, The Journal of Philosophy 98(8):411–444, 2001; Hetherington, Epistemology futures, 2006). I contend that neither of these provides conclusive arguments against the distinction. Second, I discuss studies from neuroscience and experimental psychology that relate to this distinction. Having examined these studies, I then defend a third view that explains certain relevant data from these studies by positing the (...) double dissociation of knowledge-that and knowledge-how and that is also able to do explanatory work elsewhere. (shrink)

This Handbook documents the main trends in current research between logic and language, including its broader influence in computer science, linguistic theory and cognitive science. The history of the combined study of Logic and Linguistics goes back a long way, at least to the work of the scholastic philosophers in the Middle Ages. At the beginning of this century, the subject was revitalized through the pioneering efforts of Gottlob Frege, Bertrand Russell, and Polish philosophical logicians such (...) as Kazimierz Ajdukiewicz. Around 1970, the landmark achievements of Richard Montague established a junction between state-of-the-art mathematical logic and generative linguistic theory. Over the subsequent decades, this enterprise of Montague Grammar has flourished and diversified into a number of research programs with empirical and theoretical substance. This appears to be the first Handbook to bring logic-language interface to the fore. Both aspects of the interaction between logic and language are demonstrated in the book i.e. firstly, how logical systems are designed and modified in response to linguistic needs and secondly, how mathematical theory arises in this process and how it affects subsequent linguistic theory. The Handbook presents concise, impartial accounts of the topics covered. Where possible, an author and a commentator have cooperated to ensure the proper breadth and technical content of the papers. The Handbook is self-contained, and individual articles are of the highest quality. (shrink)

Stein once urged us not to confuse the means of representation with that which is being represented. Yet that is precisely what philosophers of science appear to have done at the meta-level when it comes to representing the practice of science. Proponents of the so-called ‘syntactic’ view identify theories as logically closed sets of sentences or propositions and models as idealised interpretations, or ‘theoruncula, as Braithwaite called them. Adherents of the ‘semantic’ approach, on the other hand, are typically (...) characterised as taking them to be families of models that are set-theoretic, according to Suppes and others, or abstract, as Giere has argued. da Costa and French (Science and Partial Truth. OUP, Oxford, 2003) suggested that we should refrain from ontological speculation as to the nature of scientific theories and models and focus on their appropriate representation for various purposes within the philosophy of science. Such an approach allows both linguistic and non-linguistic resources to play their appropriate role (see also French and Saatsi, Philosophy of Science, Proceedings of the 2004 PSA Meeting, 78:548–559, 2006) and can be supported by recent case studies illustrating the heterogeneity of scientific practice. My aim in this paper is to further develop this ‘quietist’ view, and to indicate how it offers a fruitful way forward for the philosophy of science. (shrink)

The appropriate methodology for psychological research depends on whether one is studying mental algorithms or their implementation. Mental algorithms are abstract specifications of the steps taken by procedures that run in the mind. Implementational issues concern the speed and reliability of these procedures. The algorithmic level can be explored only by studying across-task variation. This contrasts with psychology's dominant methodology of looking for within-task generalities, which is appropriate only for studying implementational issues.The implementation-algorithm distinction is related to a number of (...) other “levels” considered in cognitive science. Its realization in Anderson's ACT theory of cognition is discussed. Research at the algorithmic level is more promising because it is hard to make further fundamental scientific progress at the implementational level with the methodologies available. Protocol data, which are appropriate only for algorithm-level theories, provide a richer source than data at the implementational level. Research at the algorithmic level will also yield more insight into fundamental properties of human knowledge because it is the level at which significant learning transitions are defined.The best way to study the algorithmic level is to look for differential learning outcomes in pedagogical experiments that manipulate instructional experience. This provides control and prediction in realistically complex learning situations. The intelligent tutoring paradigm provides a particularly fruitful way to implement such experiments.The implications of this analysis for the issue of modularity of mind, the status of language, research on human/computer interaction, and connectionist models are also examined. (shrink)

Imagine advanced computers that could, by virtue merely of being programmed in the right ways, act, react, communicate, and otherwise behave like humans. Might such computers be capable of understanding, thinking, believing, and the like? The framework developed in this paper for tackling challenging questions of concept application (in any realm of discourse) answers in the affirmative, contrary to Searle’s famous ‘Chinese Room’ thought experiment, which purports to prove that ascribing such mental processes to computers like these would be necessarily (...) incorrect. The paper begins by arguing that the core issue concerns language, specifically the discourse-community-guided mapping of phenomena onto linguistic categories. It then offers a model of how people adapt language to deal with novel states of affairs and thereby lend generality to their words, employing processes of assimilation, lexemic creation, and accommodation (in intersense and intrasense varieties). Attributions of understanding to some computers lie in the middle range on a spectrum of acceptability and are thus reasonable. Possible objections deriving from Searle’s writings require supplementing the model with distinctions between present and future acceptability, and between contemplated and uncontemplated word uses, as well as a literal-figurative distinction that is more sensitive than Searle’s to actual linguistic practice and the multiplicity of subsenses possible within a single literal sense. The paper then critiques two misleading rhetorical features of Searle’s Chinese Room presentation, and addresses a contemporary defense of Searle that seems to confront the sociolinguistic issue, but fails to allow for intrasense accommodation. It concludes with a brief consideration of the proper course for productive future discussion. (shrink)

In this paper I discuss two fundamental challenges concerning Crispin Wright's notion of entitlement of cognitive project: firstly, whether entitlement is an epistemic kind of warrant since, seemingly, it is not underwritten by epistemic reasons, and, secondly, whether, in the absence of such reasons, the kind of rationality associated with entitlement is epistemic in nature. The paper investigates three possible lines of response to these challenges. According to the first line of response, entitlement of cognitive project is underwritten by epistemic (...) reasons – and thus supports epistemic rationality – because, when P is an entitlement, trust in P is a dominant strategy with respect to promotion of epistemic value. The second line of response replaces dominance with maximization of expected utility. I argue that both of these proposals are flawed and develop an alternative line of response. (shrink)