Ioannis G. Ntagkas

In this paper we design a new logical system to explicitly model the different deductive reasoning steps of a boundedly rational agent. We present an adequate system in line with experimental findings about an agent’s reasoning limitations and the cognitive effort that is involved. Inspired by Dynamic Epistemic Logic, we work with dynamic operators denoting explicit applications of inference rules in our logical language. Our models are supplemented by (a) impossible worlds (not closed under logical consequence), suitably structured according to the effect of inference rules, and (b) quantitative components capturing the agent’s cognitive capacity and the cognitive costs of rules with respect to certain resources (e.g. memory, time). These ingredients allow us to avoid problematic logical closure principles, while at the same time deductive reasoning is reflected in our dynamic truth clauses. We further show that our models can be reduced to awareness-like plausibility structures that validate the same formulas and we give a sound and complete axiomatization with respect to them. This approach to the agent’s internal deductive reasoning is finally combined with actions of external information.

Natural modalities are often analysed from an abstract point of view where they are associated with putative laws of nature. However, the way possibilities are represented in physics is more complex. Lagrangian mechanics, for instance, involves two different layers of modalities: kinematical and dynamical possibilities. This paper examines the status of these two layers, both in the classical and quantum case. The quantum case is particularly problematic: we identify four possible interpretive options. The upshot is that a close inspection of the way possibilities are represented in physics could lead to new ways of thinking about natural modalities.

In this paper we first develop a logic independent account of relevant implication. We propose a stipulative denition of what it means for a multiset of premises to relevantly L-imply a multiset of conclusions, where L is a Tarskian consequence relation: the premises relevantly imply the conclusions iff there is an abstraction of the pair such that the abstracted premises L-imply the abstracted conclusions and none of the abstracted premises or the abstracted conclusions can be omitted while still maintaining valid L-consequence. Subsequently we apply this denition to the classical logic consequence relation to obtain NTR-consequence, i.e. the relevant CL-consequence relation in our sense, and develop a sequent calculus that is sound and complete with regard to relevant CL-consequence. We present a sound and complete sequent calculus for NTR. In a next step we add rules for an object language relevant implication to the sequent calculus. The object language implication reflects exactly the NTR-consequence relation. One can see the resulting logic NTR-> as a relevant logic in the traditional sense of the word. By means of a translation to the relevant logic R, we show that the presented logic NTR is very close to relevance logics in the Anderson-Belnap-Dunn-Routley-Meyer tradition. However, unlike usual relevant logics, NTR is decidable for the full language, Disjunctive Syllogism and Adjunction are valid, and neither Modus Ponens nor the Cut rule are admissible.

In this paper I discuss the political value of the right to privacy. The classical accounts of privacy do not differentiate between privacy as the right of a citizen against other citizens vs. the right to privacy as the right against the state or the government. I shall argue that this distinction should be made, since the new context of the privacy debate has surpassed the historical frames in which the intelligence methods used by governments were comparable to those available to individuals. I also present cases in which political privacy serves as an instrument of protecting important collective agendas exceeding the context of personal privacy. I argue that due to its function, political privacy should be considered a necessary element of democratic governance with the rule of law, imposing legal bounds on governments’ discretionary actions.

It is quite unequivocal that Kuhn was committed to (some version of) naturalism; that he defended, especially in his later work, the autonomy of scientific rationality; and that he rejected the correspondence theory of truth, i.e., the traditional realistic conception of the world’s mind-independence. In this paper, I argue that these three philosophical perspectives form an uneasy triangle, for while it is possible to coherently defend each of them separately or two of them combined, holding all three leads to incoherence. Hence, the only option is to reject one of them in a way that is both compatible with Kuhnian historiography and philosophically fruitful. My purpose in this paper is not exegetical. My ambition is to sketch a coherent philosophical framework which does justice to Kuhn’s historiographical model for scientific change and avoids the philosophical costs of relativism and anti-naturalism. In particular, I argue in favor of the following interconnected positions: (1) Given the role of anomalies in Kuhn’s historiographical model, the latter satisfies the two fundamental conditions for acknowledging the mind-independence of the world: the ‘irreducibility condition’ and the ‘objectivity condition’, (2) Kuhn’s rejection of the notion of an ‘Archimedean platform’ does not entail that the truth-value of beliefs doesn’t play a role in the explanation of their change, (3) Adopting scientific realism is the only way to defend the autonomy of scientific rationality within a naturalistic philosophical framework. However, adopting realism leads to a liberal rather than an eliminative version of naturalism.

This Element is an exposition of second- and higher-order logic and type theory. It begins with a presentation of the syntax and semantics of classical second-order logic, pointing up the contrasts with first-order logic. This leads to a discussion of higher-order logic based on the concept of a type. The second Section contains an account of the origins and nature of type theory, and its relationship to set theory. Section 3 introduces Local Set Theory, an important form of type theory based on intuitionistic logic. In Section 4 number of contemporary forms of type theory are described, all of which are based on the so-called 'doctrine of propositions as types'. We conclude with an Appendix in which the semantics for Local Set Theory - based on category theory - is outlined.

After defending the ontologically genuine existence of at least some of the actual nomic relations, I discuss some issues concerning their metaphysical features. I firstly argue in favour of the metaphysical contingency of nomic relations and then I suggest that their relata-specificity is the most plausible metaphysical view that guarantees the unity of facts that the laws of nature are. Finally, I present a novel account according to which some of the actual nomic relations are neither external nor external but contingently possess a kind of hybrid character

The Identity Theory of properties is an increasingly popular metaphysical view that aims to be a middle way between pure powerism and pure categoricalism. This paper’s goal is to highlight three major difficulties that IDT should address in order to be a plausible account of the nature of properties. First, although IDT needs a clear definition of the notion of qualitativity which is both adequate and compatible with the tenets of the theory, all the extant proposals fail to provide such a definition. Second, the suggested definitions of qualitativity are characterised by a metaphysical ‘neutrality’ that threatens to trivialise the important metaphysical debate regarding the question whether properties confer by themselves dispositions on their bearers or need the ‘assistance’ of laws of nature to do that. Finally, none of the extant alternative understandings of the core identity-claim of IDT can adequately meet the objection that the claim in question is unintelligible.

How do non-Humean laws govern regularities in nature? According to the Inference Problem, non-Humean accounts of governing face a central problem: it is not clear how such laws do perform their governing function. Recently, Jonathan Schaffer has argued that the introduction of a law-to-regularity axiom is sufficient to solve the Inference Problem. The authors argue that Schaffer’s solution faces a devastating dilemma: either the required axiom cannot, on its own, differentiate the non-Humean account from a Humean account of laws or, if more content is added to the primitive governing posit, it should be shown how and why the ‘outfitted’ posit obtains in the world. Furthermore, the authors show that those cases that Schaffer presents to motivate his approach are not analogous to the case of lawhood and so they cannot provide justification for his axiomatic solution.

Most metaphysicians agree that powers can exist without being manifested. The main goal of this paper is to show that adherents of an unrestricted version of Dispositional Monism cannot provide a plausible metaphysical account of the difference between a situation in which a power-instance is not manifested and a situation in which a manifestation of that power-instance actually occurs unless they undermine their own view. To this end, two kinds of manifestation-relation are introduced and it is argued that dispositional monists should appeal to the former in order to offer the required account. After defending the introduction of token-level-manifestation-relations against objections to their metaphysical robustness and explanatory non-redundancy, it is finally argued that their existence is incompatible with the core tenet of an unrestricted form of Dispositional Monism because they cannot be powers.

The bulk of the literature concerning the governing role of non-Humean laws has been concentrated on the alleged incapability of higher order nomic facts to determine the regularities in the behaviour of actual objects, the so-called Inference Problem. Most recently Ioannidis, Livanios and Psillos (2021) argue that an adequate solution to the Inference Problem requires an answer to the question of how nomic relations manage to ‘tell’ properties what to do. Ioannidis et al. dub the difficulty that all extant accounts of governing laws face to give such an answer, the Governing Problem and introduce a Dualist Model, according to which the specific behaviour of things in the world is the outcome of both the thin powers things have to be subjected to laws, and certain nomic features of the world. The aim of this paper is to show that the most plausible form of the Dualist Model can provide the basis for a novel version of the Powerful Qualities View about properties. To this end it presents ways of developing two aspects of the Dualist Model. It defends the view that nomic relations must be relata-specific in order to play their theoretical role within the Model and provides arguments for the view that natural properties should have the minimum (compatible with the core tenets of the Model) modal strength.

A well-known difficulty that affects all accounts of laws of nature according to which the latter are higher-order facts involving relations between universals (the so-called DTA accounts, from Dretske in Philosophy of Science 44:248–268, 1977; Tooley in Canadian Journal of Philosophy 7:667–698, 1977 and Armstrong (What is a Law of Nature?, Cambridge University Press, Cambridge, 1983)) is the Inference Problem: how can laws construed in that way determine the first-order regularities that we find in the actual world? Bird (Analysis 65:147–55, 2005) has argued that there is no solution to the Inference Problem which is consistent with both categorical monism (that is, the view that all natural properties are categorical) and basic tenets of Armstrong’s account of the laws of nature. This paper shows that, given Armstrong’s view about laws as first-order structural universals whose instantiation ‘produce’ nomic regularities and under specific plausible metaphysical assumptions concerning nomic relations which are consistent with a broadly construed DTA approach to laws, there is no extra difficulty regarding the Inference Problem in a categorical monistic context besides the ones that beset structural universals in general.

Recent authors have drawn attention to a new kind of defeating evidence commonly referred to as higher-order evidence. Such evidence works by inducing doubts that one’s doxastic state is the result of a flawed process – for instance, a process brought about by a reason-distorting drug. I argue that accommodating defeat by higher-order evidence requires a two-tiered theory of justification, and that the phenomenon gives rise to a puzzle. The puzzle is that at least in some situations involving higher-order defeaters the correct epistemic rules issue conflicting recommendations. For instance, a subject ought to believe p, but she ought also to suspend judgment in p. I discuss three responses. The first resists the puzzle by arguing that there is only one correct epistemic rule, an Über-rule. The second accepts that there are genuine epistemic dilemmas. The third appeals to a hierarchy or ordering of correct epistemic rules. I spell out problems for all of these responses. I conclude that the right lesson to draw from the puzzle is that a state can be epistemically rational or justified even if one has what looks to be strong evidence to think that it is not. As such, the considerations put forth constitute a non question-begging argument for a kind of externalism.

Machine learning researchers distinguish between reinforcement learning and supervised learning and refer to reinforcement learning systems as “agents”. This paper vindicates the claim that systems trained by reinforcement learning are agents while those trained by supervised learning are not. Systems of both kinds satisfy Dretske’s criteria for agency, because they both learn to produce outputs selectively in response to inputs. However, reinforcement learning is sensitive to the instrumental value of outputs, giving rise to systems which exploit the effects of outputs on subsequent inputs to achieve good performance over episodes of interaction with their environments. Supervised learning systems, in contrast, merely learn to produce better outputs in response to individual inputs.

Colburn (computer science, U. of Minnesota-Duluth) has a doctorate in philosophy and an advanced degree in computer science; he's worked as a philosophy professor, a computer programmer, and a research scientist in artificial intelligence. Here he discusses the philosophical foundations of artificial intelligence; the new encounter of science and philosophy (logic, models of the mind and of reasoning, epistemology); and the philosophy of computer science (touching on math, abstraction, software, and ontology).

The present volume stems from a book-proposal made about two years ago to College Publications, London. The main idea was that of illustrating the interplay between the contemporary work in logic and the mainstream mathematics. The division of the volume in two sections - topics in 'logic' vs topics in 'computing' - is more or less conventional. Some contributions are focussed on historical and technical details meant to put in perspective the impact of the work of some outstanding mathematicians and philosophers on the contemporary research in logic and computing science. Some other papers, also with a historical flavour, were supposed to evidentiate punctual methods of research and specific concepts or topics, as, e.g., decidability, computability, randomness, and computational or descriptive complexity. In general, the papers were intended as specific surveys of results. Other volumes - to be issued subsequently in the same series - will hopefully delineate aspects of the contemporary logic landscape that have not been illustrated here. The intended audience of the book includes graduate students in mathematical logic, foundations of matematics, and computing science, as well as philosophers, mathematicians, and, possibly, other scientists interested in the recent research on logic and computing.

A surge of empirical research demonstrating flexible cognition in animals and young infants has raised interest in the possibility of rational decision‐making in the absence of language. A venerable position, which I here call “Classical Inferentialism”, holds that nonlinguistic agents are incapable of rational inferences. Against this position, I defend a model of nonlinguistic inferences that shows how they could be practically rational. This model vindicates the Lockean idea that we can intuitively grasp rational connections between thoughts by developing the Davidsonian idea that practical inferences are at bottom categorization judgments. From this perspective, we can see how similarity‐based categorization processes widely studied in human and animal psychology might count as practically rational. The solution involves a novel hybrid of internalism and externalism: intuitive inferences are psychologically rational (in the explanatory sense) given the intensional sensitivity of the similarity assessment to the internal structure of the agent's reasons for acting, but epistemically rational (in the justificatory sense) given an ecological fit between the features matched by that assessment and the structure of the agent's environment. The essay concludes by exploring empirical results that show how nonlinguistic agents can be sensitive to these similarity assessments in a way that grants them control over their opaque judgments.

This book provides a framework for thinking about foundational philosophical questions surrounding machine learning as an approach to artificial intelligence. Specifically, it links recent breakthroughs in deep learning to classical empiricist philosophy of mind. In recent assessments of deep learning's current capabilities and future potential, prominent scientists have cited historical figures from the perennial philosophical debate between nativism and empiricism, which primarily concerns the origins of abstract knowledge. These empiricists were generally faculty psychologists; that is, they argued that the active engagement of general psychological faculties-such as perception, memory, imagination, attention, and empathy-enables rational agents to extract abstract knowledge from sensory experience. This book explains a number of recent attempts to model roles attributed to these faculties in deep neural network based artificial agents by appeal to the faculty psychology of philosophers such as Aristotle, Ibn Sina (Avicenna), John Locke David Hume, William James, and Sophie de Grouchy. It illustrates the utility of this interdisciplinary connection by showing how it can provide benefits to both philosophy and computer science: computer scientists can continue to mine the history of philosophy for ideas and aspirational targets to hit on the way to more robustly rational artificial agents, and philosophers can see how some of the historical empiricists' most ambitious speculations can be realized in specific computational systems.

Libertarian views of freedom claim that, although determinism would rule out our freedom, we are nevertheless free on some occasions. An odd implication of such views (to put it mildly) seems to be that indeterminism somehow enhances or contributes to our agency. But how could that be? What does indeterminism have to offer agency? This paper develops a novel answer, one that is centred around the notion of explanation. In short, it is argued that, if indeterminism holds in the right places, then the best explanation of the history of the world necessarily cites facts about our agency. Along the way, alternative proposals regarding the significance of indeterminism are considered and, ultimately, rejected in favour of the one developed in this paper.

In “Study of Concepts”, Peacocke puts forward an argument for non-conceptualism derived from the possession conditions of observational concepts. In this paper, I raise two objections to this argument. First, I argue that if non-conceptual perceptual contents are scenario contents, then perceptual experiences cannot present perceivers with the circumstances specified by the application conditions of observational concepts and, therefore, they cannot play the semantic and epistemic roles Peacocke wants them to play in the possession conditions of these concepts. Second, I argue that if non-conceptual perceptual contents are protopropositions, then Peacocke’s account of the possession conditions of observational concepts falls into the Myth of the Given.

To make the case for non-conceptualism, Heck draws on an apparent dichotomy between linguistic and iconic representations. According to Heck, whereas linguistic representations have conceptual content, the content of iconic representations is non-conceptual. Based on the case of cartographic systems, the authors criticize Heck’s dichotomous distinction. They argue that maps are composed of semantically arbitrary elements that play different syntactic roles. Based on this, they claim that maps have a predicative structure and convey conceptual content. Finally, the authors argue that, despite their differences, maps and sentences can logically interact with each other through heterogeneous inferences. These considerations not only challenge the view that conceptual content and inferential processes necessarily involve linguistic representations; furthermore, they provide a new perspective for thinking about maps, their semantics and syntax, and their interaction with linguistic systems.

Philosophers have always tried to explain what concepts are. Currently, most neo- Fregean philosophers identify concepts with abilities peculiar to cognitive agents. Philosophers who defend a psychological view, in contrast, identify concepts with representations located in the mind. In this paper, I argue that concepts should be understood neither in terms of mental representations nor in terms of abilities. Concepts, I argue, are rules for sorting an inferring. To support this, I follow Ginsborg’s Kantian conception of concepts. Nevertheless, unlike Ginsborg, I provide an explanation of the cognitive relationship between concepts and thinkers that presupposes no linguistic awareness of any normative concept. In doing so, a dispositional approach to the normativity of concepts is proposed.