Simon Cushing conducted the following interview with Elizabeth Anderson on 18 June 2014.

Belief-in reports of the form 'S believes in O' have been taken to have at least two senses: factual and evaluative. I begin by briefly suggesting that there is no evidence for two distinct senses, then spend most of the paper developing a general semantics for belief-in reports. I explore, and use my semantics to explain, several features of belief-in reports: the context-dependence of what belief-that reports they entail, their widespread lack of equivalence with belief-that reports, and their neg-raising property. (...) Put roughly, my semantics says that 'S believes in O' a) asserts that, for some contextually salient property F, S believes that O has F and b) presupposes that S either has a belief that O has F for which they meet an affective and evidential requirement or has a belief that O doesn’t have F for which they meet an affective and evidential requirement. (shrink)

The relationship of part to whole is one of the most fundamental there is; this is the first and only full-length study of this concept. This book shows that mereology, the formal theory of part and whole, is essential to ontology. Peter Simons surveys and criticizes previous theories, especially the standard extensional view, and proposes a more adequate account which encompasses both temporal and modal considerations in detail. 'Parts could easily be the standard book on mereology for the next twenty (...) or thirty years.' Timothy Williamson, Grazer Philosophische Studien. (shrink)

Scientific realism driven by inference to the best explanation (IBE) takes empirically confirmed objects to exist, independent, pace empiricism, of whether those objects are observable or not. This kind of realism, it has been claimed, does not need probabilistic reasoning to justify the claim that these objects exist. But I show that there are scientific contexts in which a non-probabilistic IBE-driven realism leads to a puzzle. Since IBE can be applied in scientific contexts in which empirical confirmation has not yet (...) been reached, realists will in these contexts be committed to the existence of empirically unconfirmed objects. As a consequence of such commitments, because they lack probabilistic features, the possible empirical confirmation of those objects is epistemically redundant with respect to realism. (shrink)

Which political principles should govern global politics? In his new book, Simon Caney engages with the work of philosophers, political theorists, and international relations scholars in order to examine some of the most pressing global issues of our time. Are there universal civil, political, and economic human rights? Should there be a system of supra- state institutions? Can humanitarian intervention be justified?

I explore the process of changes in the observability of entities and objects in science and how such changes impact two key issues in the scientific realism debate: the claim that predictively successful elements of past science are retained in current scientific theories, and the inductive defense of a specific version of inference to the best explanation with respect to unobservables. I provide a case-study of the discovery of radium by Marie Curie in order to show that the observability of (...) some entities can change and that such changes are relevant for arguments seeking to establish the reliability of success-to-truth inferences with respect to unobservables. (shrink)

The main purpose of this paper is to refute the metaphysicians ‘methodological continuation’ argument supporting epistemic realism in metaphysics. This argument aims to show that scientific realists have to accept that metaphysics is as rationally justified as science given that they both employ inference to the best explanation, i.e. that metaphysics and science are methodologically continuous. I argue that the reasons given by scientific realists as to why inference to the best explanation is reliable in science do not constitute a (...) reason to believe that it is reliable in metaphysics. The justification of IBE in science and the justification of IBE in metaphysics are two distinct issues with only superficial similarities, and one cannot rely on one for the other. This becomes especially clear when one analyses the debate about the legitimacy of IBE that has taken place between realists and empiricists. The metaphysician seeking to piggyback on the realist defense of IBE in science by the methodological continuation argument presupposes that the defense is straightforwardly applicable to metaphysics. I will argue that it is, in fact, not. The favoured defenses of IBE in scientific realism make extensive use of empirical considerations, predictive power and inductive evidence, all of which are paradigmatically absent in the metaphysical context. Furthermore, I argue that the metaphysician, even if the realist would concede to the methodological continuation argument, fails to offer any agreed upon conclusions resulting from its application in metaphysics. As a result, the scientific realist is not committed to believing that there is metaphysical knowledge. (shrink)

Recent advances in natural language processing have given rise to a new kind of AI architecture: the language agent. By repeatedly calling an LLM to perform a variety of cognitive tasks, language agents are able to function autonomously to pursue goals specified in natural language and stored in a human-readable format. Because of their architecture, language agents exhibit behavior that is predictable according to the laws of folk psychology: they function as though they have desires and beliefs, and then make (...) and update plans to pursue their desires given their beliefs. We argue that the rise of language agents significantly reduces the probability of an existential catastrophe due to loss of control over an AGI. This is because the probability of such an existential catastrophe is proportional to the difficulty of aligning AGI systems, and language agents significantly reduce that difficulty. In particular, language agents help to resolve three important issues related to aligning AIs: reward misspecification, goal misgeneralization, and uninterpretability. (shrink)

It is generally assumed that existing artificial systems are not phenomenally conscious, and that the construction of phenomenally conscious artificial systems would require significant technological progress if it is possible at all. We challenge this assumption by arguing that if Global Workspace Theory (GWT) — a leading scientific theory of phenomenal consciousness — is correct, then instances of one widely implemented AI architecture, the artificial language agent, might easily be made phenomenally conscious if they are not already. Along the way, (...) we articulate an explicit methodology for thinking about how to apply scientific theories of consciousness to artificial systems and employ this methodology to arrive at a set of necessary and sufficient conditions for phenomenal consciousness according to GWT. (shrink)

The descriptions 'good' and 'bad' are examples of thin concepts, as opposed to 'kind' or 'cruel' which are thick concepts. Simon Kirchin provides one of the first full-length studies of the crucial distinction between 'thin' and 'thick' concepts, which is fundamental to many debates in ethics, aesthetics and epistemology.

One Stoic response to the skeptical indistinguishability argument is that it fails to account for expertise: the Stoics allow that while two similar objects create indistinguishable appearances in the amateur, this is not true of the expert, whose appearances succeed in discriminating the pair. This paper re-examines the motivations for this Stoic response, and argues that it reveals the Stoic claim that, in generating a kataleptic appearance, the perceiver’s mind is active, insofar as it applies concepts matching the perceptual stimulus. (...) I argue that this claim is reflected in the Stoic definition of the kataleptic appearance, and that it respects their more general account of mental representation. I further suggest that, in attributing some activity to the mind in creating each kataleptic appearance, and in claiming that the expert’s mind allows her to form more kataleptic appearances than the amateur, the Stoics draw inspiration from the wax tablet model in Plato’s Theaetetus (190e–196d), where Socrates distinguishes the wise from the ignorant on the basis of how well they match sensory input with its appropriate mental ‘seal’ (σφραγίς). (shrink)

Here I propose an interpretation of the ancient Stoic psychological theory on which (i) the concepts that an adult human possesses affect the content of the perceptual impressions (φαντασίαι αἰσθητικαί) she forms, and (ii) the content of such impressions is exhausted by an ‘assertible’ (ἀξίωμα) of suitable complexity. What leads the Stoics to accept (i) and (ii), I argue, is their theory of assent and belief formation, which requires that the perceptual impression communicate information suitable to serve as the content (...) of belief. In arguing for (i), I reject a rival interpretation on which conceptualization occurs subsequently to the formation of a perceptual impression. In arguing for (ii), I deny that perceptual impressions have two kinds of content: one formulated in an assertible, the other sensory, featuring independently of this assertible. I explore the implications of (i) and (ii) for the Stoic theory of emotions, expertise, and rationality, and argue that they shed new light on the workings of impression, assent, and belief. (shrink)

This paper develops an information-sensitive theory of the semantics and probability of conditionals and statements involving epistemic modals. The theory validates a number of principles linking probability and modality, including the principle that the probability of a conditional If A, then C equals the probability of C, updated with A. The theory avoids so-called triviality results, which are standardly taken to show that principles of this sort cannot be validated. To achieve this, we deny that rational agents update their credences (...) via conditionalization. We offer a new rule of update, Hyperconditionalization, which agrees with Conditionalization whenever nonmodal statements are at stake but differs for modal and conditional sentences. (shrink)

What is it for two people to think of an object, natural kind or other entity under the same mode of presentation (MOP)? This has seemed a particularly difficult question for advocates of the Mental Files approach, the Language of Thought, or other ‘atomistic’ theories. In this paper I propose a simple answer. I first argue that, by parallel with the synchronic intrapersonal case, the sharing of a MOP should involve a certain kind of epistemic transparency between the token thoughts (...) of the two thinkers. I then explain how shared words help bring about this transparency. Finally, I show how this account can be extended for thoughts expressed using demonstratives or indexicals. (shrink)

Despite well-established results in survey methodology, many experimental philosophers have not asked whether and in what way conclusions about folk intuitions follow from people’s responses to their surveys. Rather, they appear to have proceeded on the assumption that intuitions can be simply read off from survey responses. Survey research, however, is fraught with difficulties. I review some of the relevant literature—particularly focusing on the conversational pragmatic aspects of survey research—and consider its application to common experimental philosophy surveys. I argue for (...) two claims. First, that experimental philosophers’ survey methodology leaves the facts about folk intuitions massively underdetermined; and second, that what has been regarded as evidence for the instability of philosophical intuitions is, at least in some cases, better accounted for in terms of subjects’ reactions to subtle pragmatic cues contained in the surveys. (shrink)

This paper explores the principle that knowledge is fragile, in that whenever S knows that S doesn’t know that S knows that p, S thereby fails to know p. Fragility is motivated by the infelicity of dubious assertions, utterances which assert p while acknowledging higher-order ignorance whether p. Fragility is interestingly weaker than KK, the principle that if S knows p, then S knows that S knows p. Existing theories of knowledge which deny KK by accepting a Margin for Error (...) principle can be conservatively extended with Fragility. (shrink)

A distorted representation of one's own body is a diagnostic criterion and core psychopathology of both anorexia nervosa (AN) and bulimia nervosa (BN). Despite recent technical advances in research, it is still unknown whether this body image disturbance is characterized by body dissatisfaction and a low ideal weight and/or includes a distorted perception or processing of body size. In this article, we provide an update and meta-analysis of 42 articles summarizing measures and results for body size estimation (BSE) from 926 (...) individuals with AN, 536 individuals with BN and 1920 controls. We replicate findings that individuals with AN and BN overestimate their body size as compared to controls (ES= 0.63). Our meta-regression shows that metric methods (BSE by direct or indirect spatial measures) yield larger effect sizes than depictive methods (BSE by evaluating distorted pictures), and that effect sizes are larger for patients with BN than for patients with AN. To interpret these results, we suggest a revised theoretical framework for BSE that accounts for differences between depictive and metric BSE methods regarding the underlying body representations (conceptual vs. perceptual, implicit vs. explicit). We also discuss clinical implications and argue for the importance of multimethod approaches to investigate body image disturbance. (shrink)

Under what conditions would an artificially intelligent system have wellbeing? Despite its clear bearing on the ethics of human interactions with artificial systems, this question has received little direct attention. Because all major theories of wellbeing hold that an individual’s welfare level is partially determined by their mental life, we begin by considering whether artificial systems have mental states. We show that a wide range of theories of mental states, when combined with leading theories of wellbeing, predict that certain existing (...) artificial systems have wellbeing. Along the way, we argue that there are good reasons to believe that artificial systems can have wellbeing even if they are not phenomenally conscious. While we do not claim to demonstrate conclusively that AI systems have wellbeing, we argue that there is a significant probability that some AI systems have or will soon have wellbeing, and that this should lead us to reassess our relationship with the intelligent systems we create. (shrink)

There is much to be said for a diachronic or interpersonal individuation of singular modes of presentation (MOPs) in terms of a criterion of epistemic transparency between thought tokens. This way of individuating MOPs has been discussed recently within the mental files framework, though the issues discussed here arise for all theories that individuate MOPs in terms of relations among tokens. All such theories face objections concerning apparent failures of the transitivity of the ‘same MOP’ relation. For mental files, these (...) transitivity failures most obviously occur because mental files can merge or undergo fission. In this paper I argue that this problem is easily resolved once mental files are properly construed as continuants, whose metaphysics is analogous to that of persons or physical objects. All continuants can undergo fission or fusion, leading to similar transitivity problems, but there are well-established theories of persistence that accommodate this. I suggest that, in particular, the stage theory best suits the purposes of a continuant theory of MOPs. (shrink)

This paper offers a new defense of the externalist interpretation of the kataleptic impression. My strategy is to situate the kataleptic impression within the larger context of the Stoic account of expertise. I argue that, given mastery in recognizing the limitations of her own state of mind, the subject can restrict her assent to kataleptic impressions, even if they are phenomenologically indistinguishable from those which are not kataleptic.

Following recent arguments that cultural practices in wild animal populations have important conservation implications, we argue that recognizing captive animals as cultural has important welfare implications. Having a culture is of deep importance for cultural animals, wherever they live. Without understanding the cultural capacities of captive animals, we will be left with a deeply impoverished view of what they need to flourish. Best practices for welfare should therefore require concern for animals’ cultural needs, but the relationship between culture and welfare (...) is also extremely complex, requiring us to rethink standard assumptions about what constitutes and contributes to welfare. (shrink)

A letter replying to Bayne et al. "Tests for consciousness in humans and beyond", 2024, arguing that the search for consciousness "beyond" healthy adult humans should begin with other animals.

We notice a number of interesting overlaps between the views on personhood of Ifeanyi Menkiti and Marya Schechtman. Both philosophers distance their views from the individualistic ones standard in western thought and foreground the importance of extrinsic or relational features to personhood. For Menkiti, it is ‘the community which defines the person as person’; for Schechtman, being a person is to have a place in person-space, which involves being seen as a person by others. But there are also striking differences. (...) Schechtman sees this aspect as expanding the scope of personhood to infants and those who are severely mentally disabled. Menkiti thinks that there is a line to be drawn at some point between those humans that are persons and those who are not. We consider the cases offered in questioning how the dispute between the two views should be resolved. (shrink)

This essay is an investigation into the sense of ‘able’ relevant to free will, where free will is understood as requiring the ability to do otherwise. I argue that van Inwagen's recent functional specification of the relevant sense of ‘able’ is flawed, and that explicating the powers involved in free will shall likely require paying detailed attention to the semantics and pragmatics of ‘can’ and ‘able’. Further, I argue that van Inwagen's promise-level ability requirement on free will is too strong. (...) I also argue that Mele's conjecture that the strength of the ability to perform the ‘alternative’ action be no higher than the strength of the ability exercised in performing an action is mistaken. I suggest there is an asymmetry in the strengths of the abilities which make up the n-way power that comprises free will, and that this looks to have some interesting consequences for the connection between the abilities required for free will and, e.g. the ‘up to us’ locution. (shrink)

Existing indexing systems used to arrange philosophical works have been shown to misrepresent the discipline in ways that reflect and perpetuate exclusionary attitudes within it. In recent years, there has been a great deal of effort to challenge those attitudes and to revise them. But as the discipline moves toward greater equality and inclusivity, the way it has indexed its work has unfortunately not. To course correct, we identify in this article some of the specific changes that are needed within (...) current indexing systems and propose a new model that could embody them. We use the Diversity Reading List in Philosophy as a case study and PhilPapers as a basis for comparison. The model we propose not only represents the discipline in a more inclusive and fair way, but it is also efficient, easy to use or implement, and adaptable for a variety of contexts. (shrink)

Although digital breast tomosynthesis has higher sensitivity than digital mammography and at least as high specificity, digital mammography remains the most common method for conducting mammographic screening. At the same time, mammography systems are now delivered “DBT-ready” and can be used for either digital mammography or digital breast tomosynthesis. In this paper, we ask whether it is ethically permissible to use such equipment for digital mammography, given its lower sensitivity. We argue it is not, and that clinics are ethically required (...) to use their DBTready equipment to screen with digital breast tomosynthesis whenever this is practically possible. Our argument relies on a comparison between digital breast tomosynthesis and a hypothesized improvement in the image quality of digital mammography. (shrink)

Recent research has identified a tension between the Safety principle that knowledge is belief without risk of error, and the Closure principle that knowledge is preserved by competent deduction. Timothy Williamson reconciles Safety and Closure by proposing that when an agent deduces a conclusion from some premises, the agent’s method for believing the conclusion includes their method for believing each premise. We argue that this theory is untenable because it implies problematically easy epistemic access to one’s methods. Several possible solutions (...) are explored and rejected. (shrink)

According to the B-theory, the passage of time is an illusion. The B-theory therefore requires an explanation of this illusion before it can be regarded as fullysatisfactory; yet very few B-theorists have taken up the challenge of trying to provide one. In this paper I take some first steps toward such an explanation by first making a methodological proposal, then a hypothesis about a key element in the phenomenology of temporal passage. The methodological proposal focuses onthe representational content of the (...) element of experience by virtue of which time seems to pass. The hypothesis involves the claim that the experience of changeinvolves the representation of something enduring, rather than perduring, through any change. (shrink)

Perry Hendricks’ ‘impairment argument’, which he has defended in this journal, is intended to demonstrate that the generally conceded wrongness of giving a fetus fetal alcohol syndrome (FAS) shows that abortion must also be immoral, even if we allow that the fetus is not a rights-bearing moral person. The argument fails because the harm of causing FAS is extrinsic but Hendricks needs it to be intrinsic for it to show anything about abortion. Either the subject of the wrong of causing (...) FAS is a person who does not exist in the case of abortion or the wrong is negligible. (shrink)

Stump and Timpe have recently proposed Thomistic based solutions to the traditional problem in Christian theology of how to relate grace and free will. By taking a closer look at the notion of control, I subject Timpe’s account – itself an extension of Stump’s account – to extended critique. I argue that the centrepiece of Timpe’s solution, his reliance on Dowe’s notion of quasi-causation, is misguided and irrelevant to the problem. As a result, Timpe’s account fails to avoid Semi-Pelagianism. I (...) canvass two alternatives, each of which adheres to the broad theological assumptions made by Stump and Timpe, including the positing of only one “unique” grace. I conclude that each of these proposals fails, although I argue that one comes as close as it is possible to get to a solution given the assumptions made. -/- . (shrink)

This paper considers the question of whether chimpanzees possess at least a primitive sense of normativity: i.e., some ability to internalize and enforce social norms—rules governing appropriate and inappropriate behaviour—within their social groups, and to make evaluations of others’ behaviour in light of such norms. A number of scientists and philosophers have argued that such a sense of normativity does exist in chimpanzees and in several other non-human primate and mammalian species. However, the dominant view in the scientific and philosophical (...) literature is that psychological capacities for social norms evolved uniquely in the human lineage, after our last common ancestor with chimpanzees and bonobos. After reviewing some of the existing evidence for normative capacity in chimpanzees, I defend the thesis of chimpanzee normativity against three key theoretical objections that have been presented in the literature, each of which have played a part in motivating the dominant sceptical position. I argue that, while we still have much to learn about the nature and extent of the normative capacities of other animals, there is strong prima facie evidence for social norms and normative evaluation in chimpanzees and the main theoretical objections to chimpanzee normativity are not at all compelling. (shrink)

Empirical studies of the social lives of non-human primates, cetaceans, and other social animals have prompted scientists and philosophers to debate the question of whether morality and moral cognition exists in non-human animals. Some researchers have argued that morality does exist in several animal species, others that these species may possess various evolutionary building blocks or precursors to morality, but not quite the genuine article, while some have argued that nothing remotely resembling morality can be found in any non-human species. (...) However, these different positions on animal morality generally appear to be motivated more by different conceptions of how the term “morality” is to be defined than by empirical disagreements about animal social behaviour and psychology. After delving deeper into the goals and methodologies of various of the protagonists, I argue that, despite appearances, there are actually two importantly distinct debates over animal morality going on, corresponding to two quite different ways of thinking about what it is to define “morality”, “moral cognition”, and associated notions. Several apparent skirmishes in the literature are thus cases of researchers simply talking past each other. I then focus on what I take to be the core debate over animal morality, which is concerned with understanding the nature and phylogenetic distribution of morality conceived as a psychological natural kind. I argue that this debate is in fact largely terminological and non-substantive. Finally, I reflect on how this core debate might best be re-framed. (shrink)

There is a libertarian argument for live donor organ markets, according to which live donor organ markets would be permitted if we simply refrained from imposing any substantive and controversial moral assumptions on people who reasonably disagree about morality and justice. I argue that, to the contrary, this endorsement of live donor organ markets depends upon the libertarians’ adoption of a substantive and deeply controversial conception of strong, extensive property rights. This is shown by the fact that these rights would (...) prevent states intervening in cases of preventable and intuitively impermissible wronging of others that can arise when free individuals engage in voluntary offers and exchanges. I outline two forms of such wronging: discrimination and disrespectful demands. I argue that although these types of acts are morally impermissible, the policy question of whether and how they should be regulated by states is non-trivial. I then argue that there is good reason to think that organ markets would rely on disrespectful demands. This may help explain the widespread moral repugnance people feel toward organ trading. It also provides a prima facie, though not decisive, case for states to prohibit such markets. (shrink)

This paper models knowledge in cases where an agent has multiple experiences over time. Using this model, we introduce a series of observations that undermine the pretheoretic idea that the evidential significance of experience depends on the extent to which that experience matches the world. On the basis of these observations, we model knowledge in terms of what is likely given the agent’s experience. An agent knows p when p is implied by her epistemic possibilities. A world is epistemically possible (...) when its probability given the agent’s experiences is not significantly lower than the probability of the actual world given that experience. (shrink)

Frankfurt-style cases purport to show that an agent can be morally responsible for an action despite not having any alternatives. Some critics have responded by highlighting various alternatives that remain in the cases presented, while Frankfurtians have objected that such alternatives are typically not capable of grounding responsibility. In this essay I address the recent suggestion by Seth Shabo that only alternatives associated with the ‘up to us’ locution ground moral responsibility. I distinguish a number of kinds of ability, suggest (...) which kinds of abilities ground the truth of the ‘up to us’ locution, and outline how these distinctions apply to the indeterministic buffer cases. (shrink)

The empirical phenomenon at the center of this paper is projection, which we define (uncontroversially) as follows: (1) Definition of projection An implication projects if and only if it survives as an utterance implication when the expression that triggers the implication occurs under the syntactic scope of an entailment-cancelling operator. Projection is observed, for example, with utterances containing aspectual verbs like stop, as shown in (2) and (3) with examples from English and Paraguayan Guaraní (Paraguay, Tupí-Guaraní).1 The Guaraní example in (...) (2) and its English translation have at least the following implications: (i) Carla has previously smoked, and (ii) Carla stopped smoking. The first but not the second of these implications is also conveyed by the question version of sentence (2), as in (3a), or when (2) is embedded under entailment-cancelling sentential operators, such as negation, as in (3b), the antecedent of a conditional, as in (3c), or an epistemic modal, as in (3d). Hence, by the definition in (14), the first but not the second implication of (2) projects. (shrink)

The work of Thomas Kuhn has been very influential in Anglo-American philosophy of science and it is claimed that it has initiated the historical turn. Although this might be the case for English speaking countries, in France an historical approach has always been the rule. This article aims to investigate the similarities and differences between Kuhn and French philosophy of science or ‘French epistemology’. The first part will argue that he is influenced by French epistemologists, but by lesser known authors (...) than often thought. The second part focuses on the reactions of French epistemologists on Kuhn’s work, which were often very critical. It is argued that behind some superficial similarities there are deep disagreements between Kuhn and French epistemology. This is finally shown by a brief comparison with the reaction of more recent French philosophers of science, who distance themselves from French epistemology and are more positive about Kuhn. Based on these diverse appreciations of Kuhn, a typology of the different positions within the philosophy of science is suggested. (shrink)

This paper develops a semantic solution to the puzzle of Free Choice permission. The paper begins with a battery of impossibility results showing that Free Choice is in tension with a variety of classical principles, including Disjunction Introduction and the Law of Excluded Middle. Most interestingly, Free Choice appears incompatible with a principle concerning the behavior of Free Choice under negation, Double Prohibition, which says that Mary can’t have soup or salad implies Mary can’t have soup and Mary can’t have (...) salad. Alonso-Ovalle 2006 and others have appealed to Double Prohibition to motivate pragmatic accounts of Free Choice. Aher 2012, Aloni 2018, and others have developed semantic accounts of Free Choice that also explain Double Prohibition. -/- This paper offers a new semantic analysis of Free Choice designed to handle the full range of impossibility results involved in Free Choice. The paper develops the hypothesis that Free Choice is a homogeneity effect. The claim possibly A or B is defined only when A and B are homogenous with respect to their modal status, either both possible or both impossible. Paired with a notion of entailment that is sensitive to definedness conditions, this theory validates Free Choice while retaining a wide variety of classical principles except for the transitivity of entailment. The homogeneity hypothesis is implemented in two different ways, homogeneous alternative semantics and homogeneous dynamic semantics, with interestingly different consequences. (shrink)

The 26 essays collected in this volume explore some crucial aspects of the philosophical and scientific traditions of the Middle Ages and the Renaissance. Using a historical-philological approach, the volume brings to light unpublished documents and offers new reconstructions of the intellectual paths of authors such as Meister Eckhart, Nicole Oresme, John Buridan, Siger of Brabant, William of Ockham, Peter Pomponazzi, and Giovanni Pico della Mirandola.

LLMs have dramatically improved in capabilities in recent years. This raises the question of whether LLMs could become genuine agents with beliefs and desires. This paper demonstrates an in principle limit to LLM agency, based on their architecture. LLMs are next word predictors: given a string of text, they calculate the probability that various words can come next. LLMs produce outputs that reflect these probabilities. I show that next word predictors are exploitable. If LLMs are prompted to make probabilistic predictions (...) about the world, these predictions are guaranteed to be incoherent, and so Dutch bookable. If LLMs are prompted to make choices over actions, their preferences are guaranteed to be intransitive, and so money pumpable. In short, the problem is that selecting an action based on its potential value is structurally different then selecting the description of an action that is most likely given a prompt: probability cannot be forced into the shape of expected value. The in principle exploitability of LLMs raises doubts about how agential they can become. This exploitability also offers an opportunity for humanity to safely control such AI systems. (shrink)

: I offer a new approach to the increasingly convoluted debate between the A- and B-theories of time, the ‘tensed’ and ‘tenseless’ theories. It is often assumed that the B-theory faces more difficulties than the A-theory in explaining the apparently tensed features of temporal experience. I argue that the A-theory cannot explain these features at all, because on any physicalist or supervenience theory of the mind, in which the nature of experience is fixed by the physical state of the world, (...) the tensed properties of time posited by the A-theory could play no role in shaping temporal experience. It follows that the A-theory is false; even a priori arguments for it fail, because they still require the tensed vocabulary which is used to describe temporal experience. (shrink)

According to some, free will requires alternative possibilities. But not any old alternative possibility will do. Sometimes, being able to bring about an alternative does not bestow any control on an agent. In order to bestow control, and so be directly relevant qua alternative to grounding the agent's moral responsibility, alternatives need to be robust. Here, I investigate the nature of robust alternatives. I argue that Derk Pereboom's latest robustness criterion is too strong, and I suggest a different criterion based (...) on the idea that what agents need to be able to do is keep open the possibility of securing their blamelessness, rather than needing to directly ensure their own blamelessness at the time of decision. (shrink)

This paper examines the question of whether Large Language Models (LLMs) like ChatGPT possess minds, focusing specifically on whether they have a genuine folk psychology encompassing beliefs, desires, and intentions. We approach this question by investigating two key aspects: internal representations and dispositions to act. First, we survey various philosophical theories of representation, including informational, causal, structural, and teleosemantic accounts, arguing that LLMs satisfy key conditions proposed by each. We draw on recent interpretability research in machine learning to support these (...) claims. Second, we explore whether LLMs exhibit robust dispositions to perform actions, a necessary component of folk psychology. We consider two prominent philosophical traditions, interpretationism and representationalism, to assess LLM action dispositions. While we find evidence suggesting LLMs may satisfy some criteria for having a mind, particularly in game-theoretic environments, we conclude that the data remains inconclusive. Additionally, we reply to several skeptical challenges to LLM folk psychology, including issues of sensory grounding, the "stochastic parrots" argument, and concerns about memorization. Our paper has three main upshots. First, LLMs do have robust internal representations. Second, there is an open question to answer about whether LLMs have robust action dispositions. Third, existing skeptical challenges to LLM representation do not survive philosophical scrutiny. (shrink)

Mereological nihilists typically employ a paraphrase strategy in order to mitigate the apparent absurdity of their denial of the existence of composite objects. I argue here that the nihilist's paraphrase strategy is incomplete, because no schema for generating nihilistically acceptable paraphrases of sentences concerning material constitution has ever been given. Nor can an adequate schema be arrived at by generalising things that nihilists have already said. I fill this lacuna in the nihilist's account by developing and defending a novel paraphrase (...) schema that makes essential reference to the determinable‐determinate relation that holds between arrangement‐properties instantiated by the simples. (shrink)

This paper offers the first careful analysis of the possibility that AI and humanity will go to war. The paper focuses on the case of artificial general intelligence, AI with broadly human capabilities. The paper uses a bargaining model of war to apply standard causes of war to the special case of AI/human conflict. The paper argues that information failures and commitment problems are especially likely in AI/human conflict. Information failures would be driven by the difficulty of measuring AI capabilities, (...) by the uninterpretability of AI systems, and by differences in how AIs and humans analyze information. Commitment problems would make it difficult for AIs and humans to strike credible bargains. Commitment problems could arise from power shifts, rapid and discontinuous increases in AI capabilities. Commitment problems could also arise from missing focal points, where AIs and humans fail to effectively coordinate on policies to limit war. In the face of this heightened chance of war, the paper proposes several interventions. War can be made less likely by improving the measurement of AI capabilities, capping improvements in AI capabilities, designing AI systems to be similar to humans, and by allowing AI systems to participate in democratic political institutions. (shrink)

Many defend the thesis that when someone knows p, they couldn’t easily have been wrong about p. But the notion of easy possibility in play is relatively undertheorized. One structural idea in the literature, the principle of Counterfactual Closure (CC), connects easy possibility with counterfactuals: if it easily could have happened that p, and if p were the case, then q would be the case, it follows that it easily could have happened that q. We first argue that while CC (...) is false, there is a true restriction of it to cases involving counterfactual dependence on a coin flip. The failure of CC falsifies a model where the easy possibilities are counterfactually similar to actuality. Next, we show that extant normality models, where the easy possibilities are the sufficiently normal ones, are incompatible with the restricted CC thesis involving coin flips. Next, we develop a new kind of normality theory that can accommodate the restricted version of CC. This new theory introduces a principle of Counterfactual Contamination, which says roughly that any world is fairly abnormal if at that world very abnormal events counterfactually depend on a coin flip. Finally, we explain why coin flips and other related events have a special status. A central take home lesson is that the correct principle in the vicinity of Safety is importantly normality-theoretic rather than (as it is usually conceived) similarity-theoretic. (shrink)

This volume has its roots in two recent developments within mainstream analytic epistemology: a growing recognition over the past two or three decades of the active and social nature of our epistemic lives; and, more recently still, the increasing appreciation of the various ways in which the epistemic practices of individuals and societies can, and often do, go wrong. The theoretical analysis of these breakdowns in epistemic practice, along with the various harms and wrongs that follow as a consequence, constitutes (...) an approach to epistemology that we refer to as non-ideal epistemology. In this introductory chapter we introduce and contextualise the ten essays that comprise this volume, situating them within four broad sub-fields: vice epistemology, epistemic injustice, inter-personal epistemic practices, and applied epistemology. We also provide a brief overview of several other important growth areas in non-ideal epistemology. (shrink)

Does Seneca entirely reject the utility of dialectical study for moral improvement? No, I argue here. Focusing on Letter 87, I propose that Seneca raises and disarms objections to formal Stoic arguments in order to help moral progressors avoid backsliding and advance towards ethical knowledge. I trace this method back to Chrysippus and show that reading Letter 87 in this Chrysippean framework yields a satisfying explanation of its otherwise puzzling features.

Frege held that indexical thoughts could be retained through changes of context that required a change of indexical term. I argue that Frege was partially right in that a singular mode of presentation can be retained through changes of indexical. There must, however, be a further mode of presentation that changes when the indexical term changes. This suggests that indexicals should be regarded as complex demonstratives; a change of indexical term is like a change between 'that φ' and 'that ψ', (...) where 'φ' and 'ψ' pick out relational properties that may nonetheless be conceived of by the thinker as intrinsic. (shrink)