We propose developing AIs whose only final goal is being shut down. We argue that this approach to AI safety has three benefits: (i) it could potentially be implemented in reinforcement learning, (ii) it avoids some dangerous instrumental convergence dynamics, and (iii) it creates trip wires for monitoring dangerous capabilities. We also argue that the proposal can overcome a key challenge raised by Soares et al. (2015), that shutdown-seeking AIs will manipulate humans into shutting them down. We conclude by comparing (...) our approach with Soares et al.'s corrigibility framework. (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)

The views on contradiction and consistency that Wittgenstein expressed in his later writings have met with misunderstanding and almost uniform hositility. In this paper, I trace the roots of these views by attempting to show that, in his early writings, Wittgenstein accorded a ?unique status? to tautologies and contradictions, marking them off logically from genuine propositions. This is integral both to his Tractatus project of furnishing a theory of inference, and to the enterprise of explaining the nature of the Satz (...) (statement, proposition). Wittgenstein mantained that contradictions are not false. In his early writings this surprising thesis is a consequence of his view that contradictions are not statements. In his late writings he continues to advocate the thesis, but for quite different reasons. In these late writings, I contend, Wittgenstein succeeds in making the surprising thesis plausible. (shrink)

Quantum philosophy, a peculiar twentieth-century malady, is responsible for most of the conceptual muddle plaguing the foundations of quantum physics. When this philosophy is eschewed, one naturally arrives at Bohmian mechanics, which is what emerges from Schrodinger's equation for a nonrelativistic system of particles when we merely insist that 'particles' means particles. While distinctly non-Newtonian, Bohmian mechanics is a fully deterministic theory of particles in motion, a motion choreographed by the wave function. The quantum formalism emerges when measurement situations are (...) analyzed according to this theory. When the quantum formalism is regarded as arising in this way, the paradoxes and perplexities so often associated with quantum theory simply evaporate.Bohr's ... approach to atomic problems ... is really remarkable. He is completely convinced that any understanding in the usual sense of the word is impossible. Therefore the conversation is almost immediately driven into philosophical questions, and soon you no longer know whether you really take the position he is attacking, or whether you really must attack the position he is defending. (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)

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)

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)

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)

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)

This paper presents a direction for narrative ethics based on ethical ideas found in the works of Michel Foucault. Narrative ethics is understood here at the meta-level of cultural discourse to see how the moral subject is constituted by the discursive practices that structure the contemporary debate on reproductive technologies. At this level it becomes meta-narrative-ethics. After a theoretical discussion, this paper uses two literary narratives representing the polarized views in the debate to show how the moral subject may be (...) compelled to relate to its self. Ethics is redefined as Foucauldian rapport à soi, and ethical analysis, at this meta-level, shows how the moral self is intimately connected to cultural discourse. (shrink)

I was commissioned by Barry Smith, Editor of The Monist , to act as Advisory Editor for issue 88.1, January 2005 on the topic Humor, and we drafted the appended description. The deadline for submissions is January 31, 2004, and you are welcome to submit an article to me for consideration (word limit 7,500 words, including footnotes). What the Editor and I are, hoping for, is some serious and seriously good philosophical writing on this topic.

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)

You omega know something when you know it, and know that you know it, and know that you know that you know it, and so on. This paper first argues that omega knowledge matters, in the sense that it is required for rational assertion, action, inquiry, and belief. The paper argues that existing accounts of omega knowledge face major challenges. One account is skeptical, claiming that we have no omega knowledge of any ordinary claims about the world. Another account embraces (...) the KK thesis, and identifies knowledge with omega knowledge. This position faces counterexamples, and struggles to make sense of inexact knowledge. The paper then develops a new account of knowledge, by proposing the principle of Reflective Luminosity: if you know that you know something, then you omega know it. I argue that Reflective Luminosity allows for omega knowledge while avoiding the problems for KK. (shrink)

From the acclaimed writer and thinker--whose award-winning books include both fiction and non-fiction--a dazzlingly original plunge into the drama of philosophy, revealing its hidden but essential role in today's debates on love, religion, politics, and science. Imagine that Plato came to life in the 21st century and set out on a multi-city speaking tour: How would he handle a host on Fox News who challenges him on religion and morality? How would he mediate a debate on the best way to (...) raise a child between a Freudian psychoanalyst and a Tiger Mom? How would he answer a neuroscientist who, about to scan Plato's brain, argues that all his philosophical problems can be solved by our new technologies? What would he make of Google, and the idea that knowledge can be crowdsourced rather than reasoned out by experts? With a philosopher's depth and a novelist's imagination, Rebecca Newberger Goldstein probes the deepest issues confronting us--from sexuality and child-rearing to morality and the meaning of life--by allowing us to eavesdrop on Plato as he encounters the modern world. By reviving the Platonic art of the dialogue for the 21st century, she demonstrates that the questions he first posed continue to confound and enlarge us. (shrink)

Contextology is the science of the dynamics of the conversational context. Contextology formulates laws governing how the shared information states of interlocutors evolve in response to assertion. More precisely, the contextologist attempts to construct a function which, when provided with just a conversation’s pre-update context and the content of an assertion, delivers that conversation’s post-update context. Most contextologists have assumed that the function governing the evolution of the context is simple: the post-update context is just the pre-update context intersected with (...) the content of the assertion. We argue that this assumption is wrong: not only is it false, it is also incoherent given standard contextological assumptions. Moreover, it is impossible in principle to revise it to correctly describe the dynamics of context. We conclude that there can be no science of Contextology. The laws governing the evolution of the context in response to assertion must make essential reference to the private information states of interlocutors. (shrink)

You omega know p when you possess every iteration of knowledge of p. This book argues that omega knowledge plays a central role in philosophy. In particular, the book argues that omega knowledge is necessary for permissible assertion, action, inquiry, and belief. Although omega knowledge plays this important role, existing theories of omega knowledge are unsatisfying. One theory, KK, identifies knowledge with omega knowledge. This theory struggles to accommodate cases of inexact knowledge. The other main theory is skeptical, claiming that (...) we do not omega know any ordinary claims about the world. This book develops and critically compares three new theories of omega knowledge. (shrink)

We discuss the content and significance of John von Neumann’s quantum ergodic theorem (QET) of 1929, a strong result arising from the mere mathematical structure of quantum mechanics. The QET is a precise formulation of what we call normal typicality, i.e., the statement that, for typical large systems, every initial wave function ψ0 from an energy shell is “normal”: it evolves in such a way that |ψt ψt| is, for most t, macroscopically equivalent to the micro-canonical density matrix. The QET (...) has been mostly forgotten after it was criticized as a dynamically vacuous statement in several papers in the 1950s. However, we point out that this criticism does not apply to the actual QET, a correct statement of which does not appear in these papers, but to a different (indeed weaker) statement. Furthermore, we formulate a stronger statement of normal typicality, based on the observation that the bound on the deviations from the average specified by von Neumann is unnecessarily coarse and a much tighter (and more relevant) bound actually follows from his proof. (shrink)

Demokratisierungsvorhaben sind auf Konzepte angewiesen, die Demokratieideale nicht leichtfertig preisgeben und alternative Möglichkeiten nicht grundsätzlich blockieren. Dabei sollten sie an bestehende Strukturen anknüpfen und die Entscheidungsfähigkeit politischer Gremien sowie die Zurechenbarkeit von Verantwortung erhalten können. Zugleich sollten sie die Qualität und den Einfluss öffentlicher Deliberation stärken und mobilisierbares politisches Engagement richtig einschätzen, ohne vorhandene Defizite als notwendige festzuschreiben. Auf der Grundlage einer Durchsicht von Gesellschafts-, Organisations-, Steuerungs- und Demokratietheorien skizziert Björn Thies ein solches Konzept, das an Arbeiten amerikanischer, britischer (...) und deutscher Autoren anknüpft. (shrink)

Recent philosophical literature on epistemic harms has paid little attention to the difference between deliberate and non-deliberate harms. In this paper, I analyze the “Curare Case,” a case from the 1940’s in which patient testimony was disregarded by physicians. This case has been described as an instance of epistemic injustice. I problematize this description, arguing instead that the case shows an instance of “epistemic disadvantage.” I propose epistemic disadvantage indicates when harms result from warranted asymmetric relations that justifiably exclude individuals (...) from hermeneutical participation. Epistemic disadvantage categorizes harms that result from justifiable exclusions, are non-deliberate, and result from poor epistemic environments. This analysis brings out a meaningful difference between accidental and deliberate harms in communicative exchanges. (shrink)

[Correction Notice: An erratum for this article was reported in Vol 109 of Psychological Review. Due to circumstances that were beyond the control of the authors, the studies reported in "Models of Ecological Rationality: The Recognition Heuristic," by Daniel G. Goldstein and Gerd Gigerenzer overlap with studies reported in "The Recognition Heuristic: How Ignorance Makes Us Smart," by the same authors and with studies reported in "Inference From Ignorance: The Recognition Heuristic". In addition, Figure 3 in the Psychological Review (...) article was originally published in the book chapter and should have carried a note saying that it was used by permission of Oxford University Press.] One view of heuristics is that they are imperfect versions of optimal statistical procedures considered too complicated for ordinary minds to carry out. In contrast, the authors consider heuristics to be adaptive strategies that evolved in tandem with fundamental psychological mechanisms. The recognition heuristic, arguably the most frugal of all heuristics, makes inferences from patterns of missing knowledge. This heuristic exploits a fundamental adaptation of many organisms: the vast, sensitive, and reliable capacity for recognition. The authors specify the conditions under which the recognition heuristic is successful and when it leads to the counter-intuitive less-is-more effect in which less knowledge is better than more for making accurate inferences. (shrink)

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)

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)

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)

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)

A major disagreement between different views about the foundations of quantum mechanics concerns whether for a theory to be intelligible as a fundamental physical theory it must involve a ‘primitive ontology’ (PO), i.e. variables describing the distribution of matter in four-dimensional space–time. In this article, we illustrate the value of having a PO. We do so by focusing on the role that the PO plays for extracting predictions from a given theory and discuss valid and invalid derivations of predictions. To (...) this end, we investigate a number of examples based on toy models built from the elements of familiar interpretations of quantum theory. (shrink)

This paper explores the relationship between dynamic and truth conditional semantics for epistemic modals. It provides a generalization of a standard dynamic update semantics for modals. This new semantics derives a Kripke semantics for modals and a standard dynamic semantics for modals as special cases. The semantics allows for new characterizations of a variety of principles in modal logic, including the inconsistency of ‘p and might not p’. Finally, the semantics provides a construction procedure for transforming any truth conditional semantics (...) for modals into a dynamic semantics for modals with similar properties. (shrink)