We normally take it for granted that other people will live on after we ourselves have died. Even if we do not believe in a personal afterlife in which we survive our own deaths, we assume that there will be a "collective afterlife" in which humanity survives long after we are gone. Samuel Scheffler maintains that this assumption plays a surprising - indeed astonishing - role in our lives.

Mintz examines seventeenth-century reactions to the political philosophy of Thomas Hobbes.

Can an AGI create a more intelligent AGI? Under idealized assumptions, for a certain theoretical type of intelligence, our answer is: “Not without outside help”. This is a paper on the mathematical structure of AGI populations when parent AGIs create child AGIs. We argue that such populations satisfy a certain biological law. Motivated by observations of sexual reproduction in seemingly-asexual species, the Knight-Darwin Law states that it is impossible for one organism to asexually produce another, which asexually produces another, and (...) so on forever: that any sequence of organisms (each one a child of the previous) must contain occasional multi-parent organisms, or must terminate. By proving that a certain measure (arguably an intelligence measure) decreases when an idealized parent AGI single-handedly creates a child AGI, we argue that a similar Law holds for AGIs. (shrink)

Legg and Hutter, as well as subsequent authors, considered intelligent agents through the lens of interaction with reward-giving environments, attempting to assign numeric intelligence measures to such agents, with the guiding principle that a more intelligent agent should gain higher rewards from environments in some aggregate sense. In this paper, we consider a related question: rather than measure numeric intelligence of one Legg- Hutter agent, how can we compare the relative intelligence of two Legg-Hutter agents? We propose an elegant answer (...) based on the following insight: we can view Legg-Hutter agents as candidates in an election, whose voters are environments, letting each environment vote (via its rewards) which agent (if either) is more intelligent. This leads to an abstract family of comparators simple enough that we can prove some structural theorems about them. It is an open question whether these structural theorems apply to more practical intelligence measures. (shrink)

We define a notion of the intelligence level of an idealized mechanical knowing agent. This is motivated by efforts within artificial intelligence research to define real-number intelligence levels of compli- cated intelligent systems. Our agents are more idealized, which allows us to define a much simpler measure of intelligence level for them. In short, we define the intelligence level of a mechanical knowing agent to be the supremum of the computable ordinals that have codes the agent knows to be codes (...) of computable ordinals. We prove that if one agent knows certain things about another agent, then the former necessarily has a higher intelligence level than the latter. This allows our intelligence no- tion to serve as a stepping stone to obtain results which, by themselves, are not stated in terms of our intelligence notion (results of potential in- terest even to readers totally skeptical that our notion correctly captures intelligence). As an application, we argue that these results comprise evidence against the possibility of intelligence explosion (that is, the no- tion that sufficiently intelligent machines will eventually be capable of designing even more intelligent machines, which can then design even more intelligent machines, and so on). (shrink)

This work has been selected by scholars as being culturally important, and is part of the knowledge base of civilization as we know it. This work was reproduced from the original artifact, and remains as true to the original work as possible. Therefore, you will see the original copyright references, library stamps (as most of these works have been housed in our most important libraries around the world), and other notations in the work. This work is in the public domain (...) in the United States of America, and possibly other nations. Within the United States, you may freely copy and distribute this work, as no entity (individual or corporate) has a copyright on the body of the work. As a reproduction of a historical artifact, this work may contain missing or blurred pages, poor pictures, errant marks, etc. Scholars believe, and we concur, that this work is important enough to be preserved, reproduced, and made generally available to the public. We appreciate your support of the preservation process, and thank you for being an important part of keeping this knowledge alive and relevant. (shrink)

Can an agent's intelligence level be negative? We extend the Legg-Hutter agent-environment framework to include punishments and argue for an affirmative answer to that question. We show that if the background encodings and Universal Turing Machine (UTM) admit certain Kolmogorov complexity symmetries, then the resulting Legg-Hutter intelligence measure is symmetric about the origin. In particular, this implies reward-ignoring agents have Legg-Hutter intelligence 0 according to such UTMs.

We present a game mechanic called pseudo-visibility for games inhabited by non-player characters (NPCs) driven by reinforcement learning (RL). NPCs are incentivized to pretend they cannot see pseudo-visible players: the training environment simulates an NPC to determine how the NPC would act if the pseudo-visible player were invisible, and penalizes the NPC for acting differently. NPCs are thereby trained to selectively ignore pseudo-visible players, except when they judge that the reaction penalty is an acceptable tradeoff (e.g., a guard might accept (...) the penalty in order to protect a treasure because losing the treasure would hurt even more). We describe an RL agent transformation which allows RL agents that would not otherwise do so to perform some limited self-reflection to learn the training environments in question. (shrink)

After generalizing the Archimedean property of real numbers in such a way as to make it adaptable to non-numeric structures, we demonstrate that the real numbers cannot be used to accurately measure non-Archimedean structures. We argue that, since an agent with Artificial General Intelligence (AGI) should have no problem engaging in tasks that inherently involve non-Archimedean rewards, and since traditional reinforcement learning rewards are real numbers, therefore traditional reinforcement learning probably will not lead to AGI. We indicate two possible ways (...) traditional reinforcement learning could be altered to remove this roadblock. (shrink)

“There is no world of thought that is not a world of language,” Walter Benjamin remarked, “and one only sees in the world what is preconditioned by ...

In light of Jewish tradition and the metaphysics of time, we argue that God can and will change the past. The argument makes for a new answer to the problem of evil and a new theory of atonement.

Bertrand Russell and the Nature of Propositions offers the first book-length defence of the Multiple Relation Theory of Judgement (MRTJ). Although the theory was much maligned by Wittgenstein and ultimately rejected by Russell himself, Lebens shows that it provides a rich and insightful way to understand the nature of propositional content. In Part I, Lebens charts the trajectory of Russell’s thought before he adopted the MRTJ. Part II reviews the historical story of the theory: What led Russell to deny the (...) existence of propositions altogether? Why did the theory keep evolving throughout its short life? What role did G. F. Stout play in the evolution of the theory? What was Wittgenstein’s concern with the theory, and, if we can’t know what his concern was exactly, then what are the best contending hypotheses? And why did Russell give the theory up? In Part III, Lebens makes the case that Russell’s concerns with the theory weren’t worth its rejection. Moreover, he argues that the MRTJ does most of what we could want from an account of propositions at little philosophical cost. This book bridges the history of early analytic philosophy with work in contemporary philosophy of language. It advances a bold reading of the theory of descriptions and offers a new understanding of the role of Stout and the representation concern in the evolution of the MRTJ. It also makes a decisive contribution to philosophy of language by demonstrating the viability of a no-proposition theory of propositions. (shrink)

In 2011, Hibbard suggested an intelligence measure for agents who compete in an adversarial sequence prediction game. We argue that Hibbard’s idea should actually be considered as two separate ideas: first, that the intelligence of such agents can be measured based on the growth rates of the runtimes of the competitors that they defeat; and second, one specific (somewhat arbitrary) method for measuring said growth rates. Whereas Hibbard’s intelligence measure is based on the latter growth-rate-measuring method, we survey other methods (...) for measuring function growth rates, and exhibit the resulting Hibbard-like intelligence measures and taxonomies. Of particular interest, we obtain intelligence taxonomies based on Big-O and Big-Theta notation systems, which taxonomies are novel in that they challenge conventional notions of what an intelligence measure should look like. We discuss how intelligence measurement of sequence predictors can indirectly serve as intelligence measurement for agents with Artificial General Intelligence (AGIs). (shrink)

In this paper, I provide what I believe to be Descartes's own solution to the problem of the Cartesian Circle. As I argue, Descartes thinks he can have certain knowledge of the premises of the Third Meditation proof of God's existence and veracity (i.e., the 3M-Proof) without presupposing God's existence. The key, as Broughton (1984) once argued, is that the premises of the 3M-Proof are knowable by the natural light. The major objection to this "natural light" gambit is that Descartes (...) identifies the natural light with the faculty of clear and distinct perception, a faculty that cannot be known to be reliable in advance of the 3M-Proof. I explain that Descartes distinguishes between three kinds of clear and distinction perceptions depending on their source; the senses, the imagination, or the intellect. I claim that although the First Meditation is designed to cast doubt on the clear and distinct perceptions of the senses and of the imagination, it is not designed to cast doubt on the clear and distinct perceptions of the intellect. The "natural light" gambit relies on the assumption that the natural light, by which propositions can be certainly known without presupposing knowledge of God's existence, is to be identified with the faculty of *intellectual* clear and distinct perception. (shrink)

A fundamental question asked in modal logic is whether a given theory is consistent. But consistent with what? A typical way to address this question identifies a choice of background knowledge axioms (say, S4, D, etc.) and then shows the assumptions codified by the theory in question to be consistent with those background axioms. But determining the specific choice and division of background axioms is, at least sometimes, little more than tradition. This paper introduces generic theories for propositional modal logic (...) to address consistency results in a more robust way. As building blocks for background knowledge, generic theories provide a standard for categorical determinations of consistency. We argue that the results and methods of this paper help to elucidate problems in epistemology and enjoy sufficient scope and power to have purchase on problems bearing on modalities in judgement, inference, and decision making. (shrink)

This morning I intended to get out of bed when my alarm went off. Hearing my alarm, I formed the intention to get up now. Yet, for a time, I remained in bed, irrationally lazy. It seems I irrationally failed to execute my intention. Such cases of execution failure pose a challenge for Mentalists about rationality, who believe that facts about rationality supervene on facts about the mind. For, this morning, my mind was in order; it was my action that (...) apparently made me irrational. What, then, should Mentalists say about the phenomenon of execution failure? The phenomenon of execution failure, and the puzzle it raises for Mentalists, have rarely been discussed. This paper addresses the puzzle in two parts. First, it argues that execution failure is a real phenomenon. It is possible for agents to irrationally fail to act on their present-directed intentions. It follows that Mentalists in the philosophy of action must solve the puzzle of explaining what is irrational about cases of execution failure. Second, this paper begins the search for such a solution. It considers six possible resolutions to the puzzle, arguing that none is obviously the most attractive. These resolutions include a requirement of overall conative consistency, an appeal to the norm of intention consistency, a form of Volitionalism, an appeal to factive mental states, and a proposal due to Garrett Cullity, and a novel requirement of proper functioning. (shrink)

What if "liberal democracy" were a contradiction in terms? This book distinguishes liberalism from democracy to defend a Rancirean vision of impure politics. Disclosing Rancire's refusal of ontology as political, The Lessons of Rancire enacts a critical theory beyond unmasking and a democratic politics beyond liberalism.

Inspired by recent progress in multi-agent Reinforcement Learning (RL), in this work we examine the collective intelligent behaviour of theoretical universal agents by introducing a weighted mixture operation. Given a weighted set of agents, their weighted mixture is a new agent whose expected total reward in any environment is the corresponding weighted average of the original agents' expected total rewards in that environment. Thus, if RL agent intelligence is quantified in terms of performance across environments, the weighted mixture's intelligence is (...) the weighted average of the original agents' intelligences. This operation enables various interesting new theorems that shed light on the geometry of RL agent intelligence, namely: results about symmetries, convex agent-sets, and local extrema. We also show that any RL agent intelligence measure based on average performance across environments, subject to certain weak technical conditions, is identical (up to a constant factor) to performance within a single environment dependent on said intelligence measure. (shrink)

We provide an intuitive motivation for the hyperreal numbers via electoral axioms. We do so in the form of a Socratic dialogue, in which Protagoras suggests replacing big-oh complexity classes by real numbers, and Socrates asks some troubling questions about what would happen if one tried to do that. The dialogue is followed by an appendix containing additional commentary and a more formal proof.

Philosophical Investigations, Volume 45, Issue 3, Page 320-333, July 2022.

This paper is an attempt to lay out a meta‐ethical position that is inspired by the framework of Wittgenstein's later philosophy. To achieve this goal, this paper is divided into two parts. First, I explore recent attempts to tie Wittgenstein's epistemology in On Certainty to moral epistemology. I argue that there can be a meaningful parallel drawn between the epistemic certainties discussed in On Certainty and what I consider to be moral certainties. These moral certainties are unjustified fundamental moral attitudes (...) that underlie our moral practices. Then, I show how the debate over moral certainty has branched into two directions. One direction presents the concept of moral certainty as a naturalistic concept. On this reading, moral certainties transcend time and place since they are rooted in our natural tendencies to act or not act in certain ways. The other direction presents moral certainty as a distinctly relativistic concept. On this reading, we have our moral certainties because we belong to communities that agree on these certainties. In the second section, I argue that we have both natural, universal certainties and localized, relative certainties. I also argue that our localized certainties are constrained by non‐moral facts about ourselves and about the world. To make this argument, I rely on Wittgenstein's concept of “general facts of nature.” The result of the paper is a meta‐ethical position that can be located in between moral relativism and moral realism. (shrink)

In the First Section of the Groundwork of the Metaphysics of Morals, Kant argues that a good-willed person “under subjective limitations and hindrances” (G 397) is required “never to act except in such a way that [she] could also will that [her] maxim should become a universal law” (G 402).2 This requirement has come to be known as the Formula of Universal Law (FUL) version of the Categorical Imperative, an “ought” statement expressing a command of reason that “represent[s] an action (...) as objectively necessary of itself, without reference to another end” (G 414). The question of how to understand and apply the FUL has received a great deal of attention in recent years. But Kant’s reasoning for the claim that a good-willed person (of a certain kind) acts in accordance with the FUL has not been investigated with the same degree of thoroughness. My purpose in this paper is to render the structure of Kant’s argument perspicuous, and thereby contribute to its proper evaluation by identifying its true strengths and weaknesses.3 Let us call Kant’s argument “K” and its conclusion “C”. The task at hand requires identification of K’s premises and of the reasoning by means of which C is derived. One of the advantages of clarifying K’s structure in this way is that doing so permits us to answer questions that continue to puzzle even the most sympathetic readers of the First Section. Some of these questions concern three “Propositions” to which Kant 1 draws our attention in the course of his presentation. The First Proposition (call it “P1”) is left unstated, but most commentators agree that it (or at least part of it) should be rendered as follows: (P1) A human action has moral worth only if it is done from duty.4 Kant states the other two propositions (call the Second “P2” and Third “P3”) explicitly: (P2) An action from duty has its moral worth not in the purpose to be attained by it but in the maxim in accordance with which it is decided upon. (G 399). (shrink)

Skepticism in epistemology refers to the supposedly irrational attitude of suspending judgment about all beliefs, particularly those taken for granted. The skeptical attitude presses philosophy to investigate the conditions under which knowledge and justification rather than accidental truths can be arrived at. In the first chapter, we will investigate how to construct a form of radical skepticism known as Pyrrhonian skepticism; we will see how Agrippa's Trilemma builds three ways of skeptical defense to object to three possible conditions in the (...) structuring of epistemic justification. In the second chapter, we will examine infinitism as a possible answer to the trilemma, as well as the conceptual bases that make possible the existence of this possible solution to the skeptical challenge. In the third, the main objection to infinitism known as the finite mind problem will be analyzed. In the fourth chapter, we will look at an alternative formulation of the trilemma, as well as the challenges this approach poses to the anti-skeptical response. (shrink)

In this rejoinder, we respond to two commentaries on the study by Greiff, S.; Stadler, M.; Sonnleitner, P.; Wolff, C.; Martin, R. Sometimes less is more: Comparing the validity of complex problem solving measures. Intelligence 2015, 50, 100–113. The study was the first to address the important comparison between a classical measure of complex problem solving (CPS) and the more recent multiple complex systems (MCS) approach regarding their validity. In the study, we investigated the relations between one classical microworld as (...) the initially developed method (here, the Tailorshop) with three more recently developed multiple complex systems (MCS; here, MicroDYN, Genetics Lab, and MicroFIN) tests. We found that the MCS tests showed higher levels of convergent validity with each other than with the Tailorshop even after reasoning was controlled for, thus empirically distinguishing between the two approaches. The commentary by Kretzschmar and the commentary by Funke, Fischer, and Holt expressed several concerns with how our study was conducted, our data was analyzed, and our results were interpreted. Whereas we acknowledge and agree with some of the more general statements made in these commentaries, we respectfully disagree with others, or we consider them to be at least partially in contrast with the existing literature and the currently available empirical evidence. (shrink)

Taking a thematic approach to Butler as a political thinker, this text develops Butler's theory of the political through an exploration of her politics of troubling.

Do we live in a computer simulation? I will present an argument that the results of a certain experiment constitute empirical evidence that we do not live in, at least, one type of simulation. The type of simulation ruled out is very specific. Perhaps that is the price one must pay to make any kind of Popperian progress.

Nelson Goodman has constructed two theories of simplicity: one of predicates; one of hypotheses. I offer a simpler theory by generalization and abstraction from his. Generalization comes by dropping special conditions Goodman imposes on which unexcluded extensions count as complicating and which excluded extensions count as simplifying. Abstraction is achieved by counting only nonisomorphic models and subinterpretations. The new theory takes into account all the hypotheses of a theory in assessing its complexity, whether they were projected prior to, or result (...) from, projection of a given hypothesis. It assigns simplicity post-projection priority over simplicity pre-projection. It better orders compound conditionals than does the theory of simplicity of hypotheses, and it does not inherit an anomaly of the theory of simplicity of predicates — its failure to order the ordering relations. Drop Goodman's special conditions, and the problems fall away with them. (shrink)

Abstract: Contemporary analytic philosophers of the “reactive attitudes” tend to share a simple conception of guilt as “self-directed blame”—roughly, an “unpleasant affect” felt in combination with, or in response to, the thought that one has violated a moral requirement, evinced substandard “quality of will,” or is blameworthy. I believe that this simple conception is inadequate. As an alternative, I offer my own theory of guilt’s logic and its connection to morality. In doing so, I attempt to articulate guilt’s defining thought (...) or proposition through an extended investigation and analysis of guilt’s many competing metaphors, which I trace from the Hebrew Bible to the New Testament to Shakespeare to contemporary vernacular. My interpretation of this literary genealogy offers a way to understand guilt’s seemingly disparate metaphors in terms of a single master-image that illuminates our self-conceptions and our relationship to morality. I conclude by making a very brief start toward a moral vindication of guilt against the backdrop of Nietzschean and Freudian analyses that explicitly call guilt’s place in a healthy social and personal life into question. I suggest that once guilt’s logic is made explicit, we can see that it makes sense of, honors, and addresses some of our deepest aspirations and needs. (shrink)

Cook and Reckhow defined a propositional formulation of the pigeonhole principle. This paper shows that there are Frege proofs of this propositional pigeonhole principle of polynomial size. This together with a result of Haken gives another proof of Urquhart's theorem that Frege systems have an exponential speedup over resolution. We also discuss connections to provability in theories of bounded arithmetic.

The threat of bioterrorism, the emergence of the SARS epidemic, and a recent focus on professionalism among physicians, present a timely opportunity for a review of, and renewed commitment to, physician obligations to care for patients during epidemics. The professional obligation to care for contagious patients is part of a larger "duty to treat," which historically became accepted when 1) a risk of nosocomial infection was perceived, 2) an organized professional body existed to promote the duty, and 3) the public (...) came to rely on the duty. Physicians' responses to epidemics from the Hippocratic era to the present suggests an evolving acceptance of the professional duty to treat contagious patients, reaching a long-held peak between 1847 and the1950's. There has been some professional retrenchment against this duty to treat in the last 40 years but, we argue, conditions favoring acceptance of the duty are met today. A renewed embrace of physicians' duty to treat patients during epidemics, despite conditions of personal risk, might strengthen medicine's relationship with society, improve society's capacity to prepare for threats such as bioterrorism and new epidemics, and contribute to the development of a more robust and meaningful medical professionalism. (shrink)

Recently there has been a growing literature on the concept of moral certainty. This concept, which is inspired by Wittgenstein’s reflections in On Certainty, is most prominently argued for by Nigel Pleasants. Pleasants contends that there is a meaningful parallel to be drawn between the epistemic certainties discussed by Wittgenstein and moral certainties. These moral certainties are unreflective, non-propositional, and show in the ways that we act. In addition, these certainties cannot be doubted by a reasonable moral agent. In a (...) recently published article, Fairhurst is critical of Pleasants’ account. Fairhurst argues that there are problems with Pleasants’ example of a moral certainty, 'it is wrong to kill innocent non-threatening people'. Fairhurst appeals to euthanasia as an example of when a reasonable person could meaningfully doubt that it is wrong to kill innocent, non-threatening people. If this is a certainty, then a doubt such as this should not be possible. In addition, he argues that Pleasants’ claim that moral certainties are natural is not well substantiated. In this article, I attempt to defend an account of moral certainty against Fairhurst’s criticisms. I divide his criticisms into two parts. His first set of criticisms deal with Pleasants’ example of a moral certainty. I argue that some of his criticisms are inconsistent while others can be addressed by making a small amendment to Pleasants’ account. His second criticism deals with Pleasants’ claim that moral certainties are natural. In order to address this criticism, I appeal to Machery’s Nomological Account of human nature and use it as a basis that allows us to consistently make the claim that there are natural moral certainties. (shrink)

The received view of Kripke's Wittgenstein on Rules and Private Language is that it fails as an interpretation because, inter alia, it ignores or overlooks what Wittgenstein has to say in the second paragraph of Philosophical Investigations 201. In this paper, I demonstrate that the paragraph in question is in fact fully accommodated within Kripke's reading, and cannot therefore be reasonably utilised to object to it. -/- In part one I characterise the objection; in part two I explain why it (...) fails; in part three I suggest why commentators might have been motivated to offer it; and in part four I claim that two commentators who have offered it also imply otherwise. (shrink)

Bearing Society in Mind disrupts the disciplinary boundaries of economic, political and cultural theory by exploring the theory and politics of the social formation.

I present an argument that for any computer-simulated civilization we design, the mathematical knowledge recorded by that civilization has one of two limitations. It is untrustworthy, or it is weaker than our own mathematical knowledge. This is paradoxical because it seems that nothing prevents us from building in all sorts of advantages for the inhabitants of said simulation.

Much contemporary metaphysics, moved by an apparent necessity to take reality to consist of given beings and properties, presents us with what appear to be deep problems requiring radical changes in the common sense conception of persons and the world. Contemporary meta-ethics ignores questions about logical form and formulates questions in ways that make the possibility of correct value judgments mysterious. In this book, Wheeler argues that given a Davidsonian understanding of truth, predication, and interpretation, and given a relativised version (...) of Aristotelian essentialism compatible with Davidson’s basic thinking, many metaphysical problems are not very deep. Likewise, many philosophers' claims that common sense needs to be modified are unfounded. He argues further that a proper consideration of questions of logical form clarifies and illuminates meta-ethical questions. Although the analyses and arguments he gives are often at odds with those at which Davidson arrived, they apply the central Davidsonian insights about semantics, understanding, and interpretation. (shrink)