A number of moral philosophers have endorsed instances of the following curious argument: it would be better if a certain moral theory were true; therefore, we have reason to believe that the theory is true. In other words, the mere truth of the theory—quite apart from the results of our believing it or acting in accord with it—would make for a better world than the truth of its rivals, and this fact provides evidence of the theory’s truth. This form of (...) argument may seem to be an obvious non-starter. After all, the fact that the truth of some empirical claim, say, the claim that there is an afterlife, would be desirable does not, by itself, give us any reason to believe it. But I argue that, when it is properly understood, this form of argument—which I call the better world argument—is valid in moral philosophy. I develop and defend a version of the argument that rests on the view that the correct moral theory cannot exhibit a certain form of self-defeat—a form that, as far as I know, has not been discussed in the literature. I also identify two promising applications of this form of argument. The first is a defense of permissions to promote one’s own private aims, rather than promote the greater good, and the second, an argument against the possibility of moral dilemmas. (shrink)

Associative duties are agent-centered duties to give defeasible moral priority to our special ties. Our strongest associative duties are to close friends and family. According to reductionists, our associative duties are just special duties—i.e., duties arising from what I have done to others, or what others have done to me. These include duties to abide by promises and contracts, compensate our benefactors in ways expressing gratitude, and aid those whom we have made especially vulnerable to our conduct. I argue, (...) though, that reductionism faces a problem: special duties are not strong enough to account for the strength of our associative duties. At the bar of associative duties, we are required to do what no special duty can warrant. I then present an alternative reductionist analysis of associative duties—the ‘Identity-Enactment Account’—which not only accounts for the peculiar strength of our associative duties, but also characterizes them in an intuitively compelling way. On this account, our strongest associative duties are special duties to protect or promote the welfare of the duty’s beneficiary by adopting and enacting a practical identity in which the duty’s beneficiary features prominently. There are persons who can legitimately demand a prominent place in our mental lives, for the protection and intimacy it affords. They can, in effect, legitimately demand to be among our nearest and dearest. The correlative of such a demand is, on our part, an associative duty we have toward them. (shrink)

I argue that the lives of domestic and enemy civilians should not receive equal weight in our proportionality calculations. Rather, the lives of enemy civilians ought to be “partially discounted” relative to the lives of domestic civilians. We ought to partially discount the lives of enemy civilians for the following reason (or so I argue). When our military wages a just war, we as civilians vest our right to self-defense in our military. This permits our military to weigh our (...) lives more heavily. Before arguing for this view I first explain why recent accounts attempting to show the opposite – that enemy civilians ought to be weighed more heavily – are mistaken. (shrink)

According to some philosophers, if moral agency is understood in behaviourist terms, robots could become moral agents that are as good as or even better than humans. Given the behaviourist conception, it is natural to think that there is no interesting moral difference between robots and humans in terms of moral agency (call it the _equivalence thesis_). However, such moral differences exist: based on Strawson’s account of participant reactive attitude and Scanlon’s relational account of blame, I argue that a distinct (...) kind of reason available to humans—call it _human-relative reason_—is not available to robots. The difference in moral reason entails that sometimes an action is morally permissible for humans, but not for robots. Therefore, when developing moral robots, we cannot consider only what humans can or cannot do. I use examples of paternalism to illustrate my argument. (shrink)

The agent-relative/agent-neutral distinction is one of the most important in contemporary moral theory. Yet, providing an adequate formal account of it has proven difficult. In this article I defend a new formal account of the distinction, one that avoids various problems faced by other accounts. My account is based on an influential account of the distinction developed by McNaughton and Rawling. I argue that their approach is on the right track but that it succumbs to two serious (...) objections. I then show how to formulate a new account that follows the key insights of McNaughton and Rawling’s approach yet avoids the two objections. (shrink)

It is now generally understood that constraints play an important role in commonsense moral thinking and generally accepted that they cannot be accommodated by ordinary, traditional consequentialism. Some have seen this as the most conclusive evidence that consequentialism is hopelessly wrong,1 while others have seen it as the most conclusive evidence that moral common sense is hopelessly paradoxical.2 Fortunately, or so it is widely thought, in the last twenty-five years a new research program, that of Agent-Relative Teleology, has (...) come to the rescue on all sides. While consequentialism says that every agent ought always to do that action that will bring about the most good, according to Agent- Relative Teleology. (shrink)

Andrew Forcehimes and Luke Semrau argue that agent-relative consequentialism is implausible because in some circumstances it classes an act as impermissible yet holds that the outcome of all agents performing that impermissible act is preferable. I argue that their problem is closely related to Derek Parfit's problem of ‘direct collective self-defeat’ and show how Parfit's plausible solution to his problem can be adapted to solve their problem.

Rights forfeiture or liability are not a path to the permissibility of self-defense (not even barring extraordinary circumstances), and the necessity condition is not intrinsic to justified self-defense. Rather, necessity in the context of justification must be distinguished from necessity in the context of rights forfeiture. While innocent aggressors only forfeit their right against necessary self-defense, culpable aggressors also forfeit, on grounds of a principle of reciprocity, certain rights against unnecessary self-defense. Yet, while culpable aggressors would therefore not be wronged (...) by certain unnecessary defensive means, the use of such means against them would still not be justified. The underlying rationale of this necessity requirement lies not in the rights of the aggressor, but in an agent-relative requirement to take fair precautions against violating the rights of the innocent. This concern is also expressed in the necessity criterion defended and formulated in this paper, which is very harsh on aggressors. To wit, the necessity condition for justified self-defense must not be interpreted as requiring the employment of literally the least harmful means or of means that the defender reasonably believes to be, literally, the least harmful ones. What he must believe about the properties and possible effects of the means he employs is something that is much less demanding. Finally, the necessity condition of justified self-defense is also harsh (on the aggressor) in not implying a “success condition” worth its name. (shrink)

In a recent paper, Jonathan Quong tries to offer further support for “the proposition that there are sometimes agent-relative prerogatives to harm nonliable persons.” In this brief paper, I will demonstrate that Quong’s argument implicitly relies on the premise that the violinist in Thomson’s famous example has a right not to be unplugged. Yet, first, Quong provides no argument in support of this premise; and second, the premise is clearly wrong. Moreover, throughout his paper Quong just question-beggingly and (...) without argument assumes that one cannot lose rights in other ways than by one’s own responsible action. I conclude that Quong has failed to provide further support for his thesis. (shrink)

Could it be true that even though we as a group ought to do something, you as an individual ought not to do your part? And under what conditions, in particular, could this happen? In this article, I discuss how a certain kind of case, introduced by David Copp, illustrates the possibility that you ought not to do your part even when you would be playing a crucial causal role in the group action. This is because you may have special (...) agent-relative reasons against participating that are not shared by the group as a whole. I defend the claim that these are indeed cases in which you ought not to do your part in what the group ought to do. I then argue that we can expect these cases to produce a troubling kind of rational conflict. (shrink)

Victor Tadros thinks the idea that in a conflict both sides may permissibly use force should (typically) be rejected. Thus, he thinks that two shipwrecked persons should not fight for the only available flotsam (which can only carry one person) but instead toss a coin, and that a bomber justifiably attacking an ammunitions factory must not be counterattacked by the innocent bystanders he endangers. I shall argue that Tadros’s claim rests on unwarranted assumptions and is also mistaken in the light (...) of the moral reasoning that he himself offers in support of his ‘means principle’. (shrink)

This paper is not so much concerned with the question under which circumstances self-defense is justified, but rather with other normative features of self-defense as well as with the source of the self-defense justification. I will argue that the aggressor’s rights-forfeiture alone – and hence the liberty-right of the defender to defend himself – cannot explain the intuitively obvious fact that a prohibition on self-defense would wrong victims of attack. This can only be explained by conceiving of self-defense also as (...) a claim-right. However, I will also argue that a claim-right cannot ground the self-defense justification either. Rather, what grounds the self-defense justification and its particular strength and scope is the fact that self-defense is an act-specific agent-relative prerogative: a defender is allowed to give particularly grave weight to his interest in engaging in self-defense, which distinguishes self-defense from most other acts. This is not the same as saying that he has a right or a liberty to engage in self-defense. Thus, self-defense, understood as a normative concept, is a claim-right, a liberty-right, and an act-specific agent-relative prerogative. (shrink)

It has been argued that an epistemically rational agent’s evidence is subjectively mediated through some rational epistemic standards, and that there are incompatible but equally rational epistemic standards available to agents. This supports Permissiveness, the view according to which one or multiple fully rational agents are permitted to take distinct incompatible doxastic attitudes towards P (relative to a body of evidence). In this paper, I argue that the above claims entail the existence of a unique and more reliable (...) epistemic standard. My strategy relies on Condorcet’s Jury Theorem. This gives rise to an important problem for those who argue that epistemic standards are permissive, since the reliability criterion is incompatible with such a type of Permissiveness. (shrink)

Absolutism—the idea that some kinds of acts are absolutely wrong and must never be done—plays an important role in medical ethics. Nicholas Denyer has defended it from some influential consequentialist critics who have alleged that absolutism is committed to “agent-relative constraints” and therefore intolerably complex and messy. Denyer ingeniously argues that, if there are problems with agent-relative constraints, then they are problems for consequentialism, since it contains agent-relative constraints, too. I show that, despite its (...) ingenuity, Denyer’s argument does not succeed. The defense of absolutism must move to other grounds. (shrink)

In the standard theory of deontic modals, ‘ought’ is understood as expressing a propositional operator. However, this view has been called into question by Almotahari and Rabern’s puzzle about deontic ‘ought’, according to which the ethical principle that one ought to be wronged by another person rather than wrong them is intuitively coherent but the standard theory makes it incoherent. In this paper, I take up Almotahari and Rabern’s challenge and offer a refinement of the standard theory to handle the (...) puzzle. I propose that ‘ought’ is evaluated relative to contextual parameters (e.g., Kratzer’s conversational backgrounds, Finlay and Snedegar’s alternative sets) and those contextual parameters are sensitive to agents as well as possible worlds. (shrink)

Many consequentialists take their theory to be anchored by a deeply intuitive idea, the “Compelling Idea” that it is always permissible to promote the best outcome. I demonstrate that this Idea is not, in fact, intuitive at all either in its agent-neutral or its evaluator-relative form. There are deeply intuitive ideas concerning the relationship of deontic to telic evaluation, but the Compelling Idea is at best a controversial interpretation of such ideas, not itself one of them. Because there (...) is no Compelling Idea at the heart of consequentialism, there is no initial burden of proof to be discharged nor any air of paradox to be cleared away by its opponents. (shrink)

This article examines how specific realist and projectivist versions of manipulability theories of causation deal with the problem of objectivity. Does an agent-dependent concept of manipulability imply that conflicting causal claims made by agents with different capacities can come out as true? In defence of the projectivist stance taken by the agency view, I argue that if the agent’s perspective is shown to be uniform across different agents, then the truth-values of causal claims do not vary arbitrarily and, (...) thus, reach a satisfactory level of objectivity. My argument connects Price’s considerations on the situation of deliberation, whose structure, common to all agents, is the same with respect to both decision making and causal claims on a concept inspired by Douglas’s classification of objectivity of thought processes: the perspective of the detached agent. I further argue that, despite his agent-independent concept of intervention, Woodward’s claim of a stronger objectivity standard cannot be achieved, as the relativity of causal concepts to a variable set brings about the issue of the agent’s choice of variables. Consequently, a more permissive objectivity standard applies to both views. (shrink)

Agents know some but not all logical consequences of what they know. Agents seem to be neither logically omniscient nor logically incompetent. Yet finding an intermediate standard of minimal rationality has proven difficult. In this paper, I take suggestions found in the literature (Lewis, 1988; Hawke, Özgün and Berto, 2020; Plebani and Spolaore, 2021) and join the forces of subject matter and impossible worlds approaches to devise a new solution to this quandary. I do so by combining a space of (...) FDE worlds (Berto and Jago, 2019) with a Lewisian (1988) understanding of subject matters as partitions. By doing so, I show how subject matters impose some order in the anarchic space of FDE worlds, while the worlds allow for distinctions between contents which would not otherwise be available. Combining the two approaches, then, brings us closer to the desired closure principles for knowledge of minimally rational agents. (shrink)

It is common to distinguish moral rules, reasons, or values that are agent-relative from those that are agent-neutral. One can also distinguish moral rules, reasons, or values that are moment-relative from those that are moment-neutral. In this article, I introduce a third distinction that stands alongside these two distinctions—the distinction between moral rules, reasons, or values that are patient-relative and those that are patient-neutral. I then show how patient-relativity plays an important role in several moral (...) theories, gives us a better understanding of agent-relativity and moment-relativity, and provides a novel objection to Derek Parfit’s “appeal to full relativity” argument. (shrink)

Nicomachean Ethics/NE, Book III, Chapters 1-5, provides Aristotle’s account of “Voluntary Movement.” It, thus, draws the Passion-Action distinction, only posited earlier in Categories, while also serving as the linchpin of NE’s discussion of Virtue, in explicitly connecting it to Right Reason. My explication of this text renders its terminology consistent with the Law of Excluded Middle and rebuts two criticisms of the Eudaimonistic Axiology on which it is based. These results are shown to be entailments of Aristotle’s doctrine that Voluntary (...) is (a) Relative. That categorization is then used to help treat a problematic case of Deliberation beyond his purview. (shrink)

It is commonly held that all deontological moral theories are agent-relative in the sense that they give each agent a special concern that she does not perform acts of a certain type rather than a general concern with the actions of all agents. Recently, Tom Dougherty has challenged this orthodoxy by arguing that agent-neutral deontology is possible. In this article I counter Dougherty's arguments and show that agent-neutral deontology is not possible.

According to the “Textbook View,” there is an extensional dispute between consequentialists and deontologists, in virtue of the fact that only the latter defend “agent-relative” principles—principles that require an agent to have a special concern with making sure that she does not perform certain types of action. I argue that, contra the Textbook View, there are agent-neutral versions of deontology. I also argue that there need be no extensional disagreement between the deontologist and consequentialist, as characterized (...) by the Textbook View. (shrink)

An idea that has attracted a lot of attention lately is the thought that consequentialism is a theory characterized basically by its agent neutrality.1 The idea, however, has also met with skepticism. In particular, it has been argued that agent neutrality cannot be what separates consequentialism from other types of theories of reasons for action, since there can be agent-neutral non-consequentialist theories as well as agent-relative consequentialist theories. I will argue in this paper that this (...) last claim is false. The paper is divided into four sections. Section one specifies two senses in which consequentialism is agent-neutral. Section two and three examine and reject, respectively, the claim that there are agent-relative consequentialist views as well as agent-neutral non-consequentialist views. I end the paper with some remarks on the plausibility, or better, the implausibility of characterizing consequentialism in terms other than agent neutrality. (shrink)

There is, on a given moral view, an agent-centered restriction against performing acts of a certain type if that view prohibits agents from performing an instance of that act-type even to prevent two or more others from each performing a morally comparable instance of that act-type. The fact that commonsense morality includes many such agent-centered restrictions has been seen by several philosophers as a decisive objection against consequentialism. Despite this, I argue that agent-centered restrictions are more plausibly (...) accommodated within a consequentialist framework than within the more standard side-constraint framework. For I argue that when we combine agent-relative consequentialism with a Kantian theory of value, we arrive at a version of consequentialism, which I call 'Kantsequentialism', that has several advantages over the standard side-constraint approach to accommodating constraints. What’s more, I argue that Kantsequentialism doesn’t have any of the disadvantages that critics of consequentializing have presumed that such a theory must have. (shrink)

I propose and defend a novel view called “de se consequentialism,” which is noteworthy for two reasons. First, it demonstrates—contra Doug Portmore, Mark Schroeder, Campbell Brown, and Michael Smith, among others—that agent-neutral consequentialism is consistent with agent-centered constraints. Second, it clarifies the nature of agent-centered constraints, thereby meriting attention from even dedicated nonconsequentialists. Scrutiny reveals that moral theories in general, whether consequentialist or not, incorporate constraints by assessing states in a first-personal guise. Consequently, de se consequentialism enacts (...) constraints through the very same feature that nonconsequentialist theories do. (shrink)

Morality is intrapersonally permissive: cases abound in which an agent has more than one morally permitted option. In contrast, there is a dearth of cases in which an agent has more than one epistemically permitted response to her evidence. Given the structural parallels between morality and epistemology, why do sources of moral permissiveness fail to have parallel permissive effects in the epistemic domain? This asymmetry between morality and epistemology cries out for explanation. The paper's task is to offer (...) an answer to that call. We explain the asymmetry by tracing moral permissiveness to two factors to which rationality is morally but not epistemically sensitive. (shrink)

We present a formal semantics for epistemic logic, capturing the notion of knowability relative to information (KRI). Like Dretske, we move from the platitude that what an agent can know depends on her (empirical) information. We treat operators of the form K_AB (‘B is knowable on the basis of information A’) as variably strict quantifiers over worlds with a topic- or aboutness- preservation constraint. Variable strictness models the non-monotonicity of knowledge acquisition while allowing knowledge to be intrinsically stable. (...) Aboutness-preservation models the topic-sensitivity of information, allowing us to invalidate controversial forms of epistemic closure while validating less controversial ones. Thus, unlike the standard modal framework for epistemic logic, KRI accommodates plausible approaches to the Kripke-Harman dogmatism paradox, which bear on non-monotonicity, or on topic-sensitivity. KRI also strikes a better balance between agent idealization and a non-trivial logic of knowledge ascriptions. (shrink)

There are two kinds of view in the literature concerning the relevance of intention to permissibility. While subjectivism assumes that an agent acts permissibly if he or she believes that the conduct is necessary for a moral purpose, for objectivism the de facto presence of an objective reason to justify one’s deeds is what matters. Recently, Scanlon and Hanser defend a moderate version of objectivism and subjectivism, respectively. Although I have a degree of sympathy toward both views, I will (...) argue that the truth lies somewhere in between. The view that I suggest in this paper hopefully occupies a space between subjectivism and objectivism and can accommodate the intuitions that neither of those views cannot account for. (shrink)

The complexities of how justice comes to be realized, and by which agents, is a relatively neglected element in contemporary theories of justice. This has left several crucial questions about agency and justice undertheorized, such as why some particular agents are responsible for realizing justice, how their contribution towards realizing justice should be understood, and what role agents such as activists and community leaders play in realizing justice. We aim to contribute towards a better understanding of the landscape of these (...) kinds of questions. First, we argue that theorists should distinguish between (i) agents who are responsible for realizing justice, but not committed to it, (ii) agents who are not responsible for realizing justice, but who are committed to it, and (iii) agents who are both responsible for and committed to realizing justice. Second, we discuss how to incorporate agents of justice more robustly into theorizing about justice. (shrink)

Recently, it has been a part of the so-called consequentializing project to attempt to construct versions of consequentialism that can support agent-relative moral constraints. Mark Schroeder has argued that such views are bound to fail because they cannot make sense of the agent relative value on which they need to rely. In this paper, I provide a fitting-attitude account of both agent-relative and agent-neutral values that can together be used to consequentialize agent- (...) class='Hi'>relative constraints. (shrink)

The purpose of this paper is to challenge what is often called the “Uniqueness” thesis. According to this thesis, given one’s total evidence, there is a unique rational doxastic attitude that one can take to any proposition. It is sensible for defenders of Uniqueness to commit to an accompanying principle that: when some agent A has equal epistemic reason both to believe that p and to believe that not p, the unique epistemically rational doxastic attitude for A to adopt (...) with respect to whether p is the suspension of judgment. In this paper, I offer a case wherein the agent has equal epistemic reason both to believe that p and to believe that not p, but the agent is not epistemically required to suspend judgment about whether p. Furthermore, the case is such that there seems to be no uniquely rational attitude for the agent to adopt. (shrink)

Are we ever morally permitted to do what is morally wrong? It seems intuitive that we are, but evidence for dissociations among judgment of permissibility and wrongness are relatively scarce. Across 4 experiments (N = 1,438), we show that people judge that some behaviors can be morally wrong and permissible. The dissociations arise because these judgments track different morally relevant aspects of everyday moral encounters. Judgments of individual rights predicted permissibility but not wrongness, while character assessment predicted wrongness but not (...) permissibility. These findings suggest a picture in which moral evaluation is granular enough to express reasoning about different types of normative considerations, notably the possibility that people can exercise their rights in morally problematic ways. (shrink)

Agent-based models have played a prominent role in recent debates about the merits of democracy. In particular, the formal model of Lu Hong and Scott Page and the associated “diversity trumps ability” result has typically been seen to support the epistemic virtues of democracy over epistocracy (i.e., governance by experts). In this paper we first identify the modeling choices embodied in the original formal model and then critique the application of the Hong-Page results to philosophical debates on the (...) class='Hi'>relative merits of democracy. In particular we argue that the “best-performing agents” in Hong-Page model should not be interpreted as experts. We next explore a closely related model in which best-performing agents are more plausibly seen as experts and show that the diversity trumps ability result fails to hold. However, with changes in other parameters (such as the deliberation dynamic) the diversity trumps ability result is restored. The sensitivity of this result to parameter choices illustrates the complexity of the link between formal modeling and more general philosophical claims; we use this debate as a platform for a more general discussion of when and how agent-based models can contribute to philosophical discussions. (shrink)

Contemporary philosophers and scientists have done much to expand our understanding of the structure and neural mechanisms of joint action. But the phenomenology of joint action has only recently become a live topic for research. One method of clarifying what is unique about the phenomenology of joint action is by considering the alternative perspective of agents subsumed in group action. By group action we mean instances of individual agents acting while embedded within a group agent, instead of with individual (...) coordination. Paradigm examples are educational bureaucracies, corporations, and nation states. There is a phenomenological difference between agents whose actions are subsumed within a group action (“the university did X”) as compared to agents who act jointly (“we did x”). Attending to this difference clarifies what is phenomenologically distinctive about joint action. Appealing to an Aristotelian account of agency and to the metaphysical concept of weak emergence, we argue that what makes paradigmatic group action distinctive is the relative inaccessibility, un-revisability, and evaluative simplicity of the group agent’s goal(s) from the perspective of individual agents. This suggests that a distinctive feature of joint agency is the maintenance of a greater sense of individual agency. Put simply, joint agency is often experienced as an enhancement of the individuals’ agency precisely because our paradigmatic agential powers are extended intersubjectively as we act together. In contrast, group agency often involves a loss of the sense of agency, precisely because it is the emergent group agent that maintains the agential powers. (shrink)

According to the standard account of forgiveness, you forgive your wrongdoer by overcoming your resentment towards them. But how exactly must you do so? And when is such overcoming fitting? The aim of this paper is to introduce a novel version of the standard account to answer these questions. Its core idea is that the reactive attitudes are a fitting response not just to someone’s blameworthiness, but to their blameworthiness being significant for you, or worthy of your caring, in virtue (...) of your relationship to it. Someone’s blameworthiness is significant for you to the extent you’re bound up with what grounds it––e.g. with the wrongdoer’s being a participant in human relationships, with their attitudes, or with the victim’s being a source of demands. So you may fittingly not care about someone’s blameworthiness if it’s sufficiently insignificant for you in this manner––e.g. if their wrong happened far off in place and time. And forgiveness revolves around this. You forgive your wrongdoer if and only if, partly out of goodwill towards them, you cease to care about their blameworthiness––a bit as if their wrong had happened far off. If I’m right, this agent-relativity-based account can resolve the apparent ‘paradoxy of forgiveness’, satisfies a number of desiderata, and is plausible on an intuitive level. (shrink)

In this essay, we explore an issue of moral uncertainty: what we are permitted to do when we are unsure about which moral principles are correct. We develop a novel approach to this issue that incorporates important insights from previous work on moral uncertainty, while avoiding some of the difficulties that beset existing alternative approaches. Our approach is based on evaluating and choosing between option sets rather than particular conduct options. We show how our approach is particularly well-suited to address (...) this issue of moral uncertainty with respect to agents that have credence in moral theories that are not fully consequentialist. (shrink)

Although there have been efforts to integrate Semantic Web technologies and artificial agents related AI research approaches, they remain relatively isolated from each other. Herein, we introduce a new ontology framework designed to support the knowledge representation of artificial agents’ actions within the context of the actions of other autonomous agents and inspired by standard cognitive architectures. The framework consists of four parts: 1) an event ontology for information pertaining to actions and events; 2) an epistemic ontology containing facts about (...) knowledge, beliefs, perceptions and communication; 3) an ontology concerning future intentions, desires, and aversions; and, finally, 4) a deontic ontology for modeling obligations and prohibitions which limit agents’ actions. The architecture of the ontology framework is inspired by deontic cognitive event calculus as well as epistemic and deontic logic. We also describe a case study in which the proposed DCEO ontology supports autonomous vehicle navigation. (shrink)

Many people presume that you can permissibly support the good features of a symbol, person, activity, or work of art while simultaneously denouncing its bad features. This chapter refines and assesses this commonsense (but undertheorized) moral justification for supporting problematic people, projects, and political symbols, and proposes an analogue of the Doctrine of Double Effect called the Doctrine of Double Endorsement (DDN). DDN proposes that when certain conditions are met, it is morally permissible to directly endorse some object in virtue (...) of its positive properties while standing against its negative properties, even though it would be morally impermissible to directly endorse those negative properties themselves. These conditions include separability (the good and bad features must not be inextricably linked), proportionality (the positive value of the good features must be significantly greater than the negative value of the bad features), and constrained choice (there must not be other things that the agent could endorse instead that share the same positive features but are not saddled with the negative ones). The chapter applies these constraints to a number of practical issues, including (among others) voting for morally troubling candidates, supporting Confederate monuments, and consuming sexist art. (shrink)

Artificial Social Agents (ASAs), which are AI software driven entities programmed with rules and preferences to act autonomously and socially with humans, are increasingly playing roles in society. As their sophistication grows, humans will share greater amounts of personal information, thoughts, and feelings with ASAs, which has significant ethical implications. We conducted a study to investigate what ethical principles are of relative importance when people engage with ASAs and whether there is a relationship between people’s values and the ethical (...) principles they prioritise. The study uses scenarios, embedded with the five AI4People ethical principles (Beneficence, Non-maleficence, Autonomy, Justice, and Explicability), involving ASAs taking on roles traditionally played by humans to understand whether these roles and behaviours (including dialogues) are seen as acceptable or unacceptable. Results from 268 participants reveal the greatest sensitivity to ASA behaviours that relate to Autonomy, Justice, Explicability, and the privacy of their personal data. Models were created using Schwartz’s Refined Values as a possible indicator of how stakeholders discern and prioritise the different AI4People ethical principles when interacting with ASAs. Our findings raise issues around the ethical acceptability of ASAs for nudging and changing behaviour due to participants’ desire for autonomy and their concerns over deceptive ASA behaviours such as pretending to have memories and emotions. (shrink)

The existence of group agents is relatively widely accepted. Examples are corporations, courts, NGOs, and even entire states. But should we also accept that there is such a thing as group consciousness? I give an overview of some of the key issues in this debate and sketch a tentative argument for the view that group agents lack phenomenal consciousness. In developing my argument, I draw on integrated information theory, a much-discussed theory of consciousness. I conclude by pointing out an implication (...) of my argument for the normative status of group agents. (shrink)

While philosophers hold that it is patently absurd to blame robots or hold them morally responsible [1], a series of recent empirical studies suggest that people do ascribe blame to AI systems and robots in certain contexts [2]. This is disconcerting: Blame might be shifted from the owners, users or designers of AI systems to the systems themselves, leading to the diminished accountability of the responsible human agents [3]. In this paper, we explore one of the potential underlying reasons for (...) robot blame, namely the folk's willingness to ascribe inculpating mental states or "mens rea" to robots. In a vignette-based experiment (N=513), we presented participants with a situation in which an agent knowingly runs the risk of bringing about substantial harm. We manipulated agent type (human v. group agent v. AI-driven robot) and outcome (neutral v. bad), and measured both moral judgment (wrongness of the action and blameworthiness of the agent) and mental states attributed to the agent (recklessness and the desire to inflict harm). We found that (i) judgments of wrongness and blame were relatively similar across agent types, possibly because (ii) attributions of mental states were, as suspected, similar across agent types. This raised the question - also explored in the experiment - whether people attribute knowledge and desire to robots in a merely metaphorical way (e.g., the robot "knew" rather than really knew). However, (iii), according to our data people were unwilling to downgrade to mens rea in a merely metaphorical sense when given the chance. Finally, (iv), we report a surprising and novel finding, which we call the inverse outcome effect on robot blame: People were less willing to blame artificial agents for bad outcomes than for neutral outcomes. This suggests that they are implicitly aware of the dangers of overattributing blame to robots when harm comes to pass, such as inappropriately letting the responsible human agent off the moral hook. (shrink)

I shall argue that in some wars both sides are (as a collective) justified, that is, they can both satisfy valid jus ad bellum requirements. Moreover, in some wars – but not in all – the individual soldiers on the unjustified side (that is, on the side without jus ad bellum) may nevertheless kill soldiers (and also civilians as a side-effect) on the justified side, even if the enemy soldiers always abide by jus in bello constraints. Traditional just war theory (...) and self-proclaimed “revisionist” just war theory think otherwise since the former focuses on the law enforcement or public authority justification for inflicting harm and the latter on the self-defense justification. These are both intrinsically asymmetrical justifications: there is no justified self-defense (properly understood) against justified self-defense, nor is there justified law-enforcement against justified law-enforcement. However, there can, as I will show, be justified self-defense against force that is justified by a necessity justification, and there can be force justified by a necessity justification being used against force that is also justified by a necessity justification. The necessity justification is not intrinsically asymmetrical, and it is an indispensable justification in the context of war. Moreover, with regard to some forms of inflicting harm on others one may give special weight to one’s own interests and the interests of those to whom one has special responsibilities when assessing the proportionality of those acts. That is, the proportionality calculation may be agent-relative. This is in particular so in the case of foreseeably preventing innocent and non-threatening people from being saved (for instance, by shooting down a tactical bomber who would have saved them by destroying an ammunitions factory) but less so in the case of the intentional or foreseeable direct harming of innocent and non-threatening people (dropping bombs on people standing near an ammunitions factory). In the light of these considerations, I will then answer the question as to when soldiers may justifiably participate in war (and when not). (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)

Philosophers have often noted a contrast between practical and theoretical reasons when it comes to cases involving equally balanced reasons. When there are strong practical reasons for A-ing, and equally strong practical reasons for some incompatible option, B-ing, the agent is permitted to make an arbitrary choice between them, having sufficient reason to A and sufficient reason to B. But when there is strong evidence for P and equally strong evidence for ~ P, one isn’t permitted to simply believe (...) one or the other. Instead, one must withhold belief, neither believing that P nor believing that ~ P. This paper examines what explains this contrast, focusing in particular on a proposal recently developed by Mark Schroeder across several papers. Schroeder aims to explain the contrast by an appeal to non-evidential, epistemic reasons against belief. But, I argue, it’s not clear exactly what those reasons are, nor how those reasons are to be weighed against evidential reasons. Despite these challenges, I argue that there are grounds for optimism that the contrast can be explained within the broad framework Schroeder provides, and I aim to provide resources to meet the aforementioned challenges. (shrink)

Are obligations of collective agents—such as states, businesses, and non-profits—ever overdemanding? I argue they are not. I consider two seemingly attractive routes to collective overdemandingness: that an obligation is overdemanding on a collective just if the performance would be overdemanding for members; and that an obligation is overdemanding on a collective just if the performance would frustrate the collective’s permissible deep preferences. I reject these. Instead, collective overdemandingness complaints should be reinterpreted as complaints about inability or third-party costs. These are (...) not the same as overdemandingness. Accordingly, we can ask an awful lot of collective agents. (shrink)

According to agent relativism, each person's moral requirements are relative to her desires or interests. That is, whether a person morally ought to ø depends on what interests or desires she has. Some philosophers charge that the main argument for agent relativism trades on an ambiguity in the term "reason," "reason for action,'' or the like. This paper shows that although the argument for agent relativism may indeed harbor an ambiguity, the ambiguity is no Achilles’ heel. (...) To remove it is not to discredit the argument, but to uncover the strongest form of it, the form that makes agent relativism tenacious. (shrink)

This paper explores the idea that a semantics for ‘ought’ should be neutral between different ways of deciding what an agent ought to do in a situation. While the idea is, I argue, well-motivated, taking it seriously leads to surprising, even paradoxical, problems for theorizing about the meaning of ‘ought’. This paper describes and defends one strategy—a form of Expressivism for the modal ‘ought’—for navigating these problems.

A number of formal epistemologists have argued that perfect rationality requires probabilistic coherence, a requirement that they often claim applies only to ideal agents. However, in “Rationality as an Absolute Concept,” Roy Sorensen contends that ‘rational’ is an absolute term. Just as Peter Unger argued that being flat requires that a surface be completely free of bumps and blemishes, Sorensen claims that being rational requires being perfectly rational. When we combine these two views, though, they lead to counterintuitive results. If (...) being rational requires being perfectly rational, and only the probabilistically coherent are perfectly rational, then this indicts all ordinary agents as irrational. In this paper, I will attempt to resolve this conflict by arguing that Sorensen is only partly correct. One important sense of ‘rational’, the sanctioning sense of ‘rational’, is an absolute term, but another important sense of ‘rational’, the sense in which someone can have rational capacities, is not. I will, then, show that this distinction has important consequences for theorizing about ideal rationality, developing an account of the relationship between ordinary and ideal rationality. Because the sanctioning sense of ‘rational’ is absolute, it is rationally required to adopt the most rational attitude available, but which attitude is most rational can change depending on whether we are dealing with ideal agents or people more like ourselves. (shrink)