The responsibility gap, commonly described as a core challenge for the effective governance of, and trust in, AI and autonomous systems (AI/AS), is traditionally associated with a failure of the epistemic and/or the control condition of moral responsibility: the ability to know what we are doing and exercise competent control over this doing. Yet these two conditions are a red herring when it comes to understanding the responsibility challenges presented by AI/AS, since evidence from the cognitive sciences shows that individual (...) humans face very similar responsibility challenges with regard to these two conditions. While the problems of epistemic opacity and attenuated behaviour control are not unique to AI/AS technologies (though they can be exacerbated by them), we show that we can learn important lessons for AI/AS development and governance from how philosophers have recently revised the traditional concept of moral responsibility in response to these challenges to responsible human agency from the cognitive sciences. The resulting instrumentalist views of responsibility, which emphasize the forward-looking and flexible role of agency cultivation, hold considerable promise for integrating AI/AS into a healthy moral ecology. We note that there nevertheless is a gap in AI/AS responsibility that has yet to be extensively studied and addressed, one grounded in a relational asymmetry of vulnerability between human agents and sociotechnical systems like AI/AS. In the conclusion of this paper we note that attention to this vulnerability gap must inform and enable future attempts to construct trustworthy AI/AS systems and preserve the conditions for responsible human agency. (shrink)

Do we have free will? Understanding free will as the ability to act freely, and free actions as exercises of this ability, I maintain that the default answer to this question is “yes.” I maintain that free actions are a natural kind, by relying on the influential idea that kinds are homeostatic property clusters. The resulting position builds on the view that agents are a natural kind and yields an attractive alternative to recent revisionist accounts of free action. My view (...) also overcomes difficulties confronted by previous views according to which free actions might be a natural kind. On my view, free actions exist and we often act freely, as long as we possess various features that are related in the right sorts of ways to each other and to the world. In turn, we acquire and retain the concept as long as most of us possess enough of those features. (shrink)

Galen Strawson’s Basic Argument is that because self-creation is required to be truly morally responsible and self-creation is impossible, it is impossible to be truly morally responsible for anything. I contend that the Basic Argument is unpersuasive and unsound. First, I argue that the moral luck debate shows that the self-creation requirement appears to be contradicted and supported by various parts of our commonsense ideas about moral responsibility, and that this ambivalence undermines the only reason that Strawson gives for the (...) self-creation requirement. Second, I argue that the self-creation requirement is so demanding that either it is an implausible requirement for a species of true moral responsibility that we take ourselves to have or it is a plausible requirement of a species of true moral responsibility that we have never taken ourselves to have. Third, I explain that Strawson overgeneralizes from instances of constitutive luck that obviously undermine moral responsibility to all kinds of constitutive luck. (shrink)

Against its prominent compatiblist and libertarian opponents, I defend Galen Strawson’s Basic Argument for the impossibility of moral responsibility. Against John Martin Fischer, I argue that the Basic Argument does not rely on the premise that an agent can be responsible for an action only if he is responsible for every factor contributing to that action. Against Alfred Mele and Randolph Clarke, I argue that it is absurd to believe that an agent can be responsible for an action when no (...) factor contributing to that action is up to that agent. Against Derk Pereboom and Clarke, I argue that the versions of agent-causal libertarianism they claim can immunize the agent to the Basic Argument actually fail to do so. Against Robert Kane, I argue that the Basic Argument does not rely on the premise that simply the presence of indeterministic factors in the process of bringing an action about is itself what rules out the agent’s chance for being responsible for that action. (shrink)

Skepticism about moral responsibility, or what is more commonly referred to as moral responsibility skepticism, refers to a family of views that all take seriously the possibility that human beings are never morally responsible for their actions in a particular but pervasive sense. This sense is typically set apart by the notion of basic desert and is defined in terms of the control in action needed for an agent to be truly deserving of blame and praise. Some moral responsibility skeptics (...) wholly reject this notion of moral responsibility because they believe it to be incoherent or impossible. Others maintain that, though possible, our best philosophical and scientific theories about the world provide strong and compelling reasons for adopting skepticism about moral responsibility. What all varieties of moral responsibility skepticism share, however, is the belief that the justification needed to ground basic desert moral responsibility and the practices associated with it—such as backward-looking praise and blame, punishment and reward (including retributive punishment), and the reactive attitudes of resentment and indignation—is not met. Versions of moral responsibility skepticism have historically been defended by Spinoza, Voltaire, Diderot, d’Holbach, Priestley, Schopenhauer, Nietzsche, Clarence Darrow, B.F. Skinner, and Paul Edwards, and more recently by Galen Strawson, Derk Pereboom, Bruce Waller, Neil Levy, Tamler Sommers, and Gregg D. Caruso. -/- Critics of these views tend to focus both on the arguments for skepticism about moral responsibility and on the implications of such views. They worry that adopting such a view would have dire consequences for our interpersonal relationships, society, morality, meaning, and the law. They fear, for instance, that relinquishing belief in moral responsibility would undermine morality, leave us unable to adequately deal with criminal behavior, increase anti-social conduct, and destroy meaning in life. Optimistic skeptics, however, respond by arguing that life without free will and basic desert moral responsibility would not be as destructive as many people believe. These optimistic skeptics argue that prospects of finding meaning in life or of sustaining good interpersonal relationships, for instance, would not be threatened. They further maintain that morality and moral judgments would remain intact. And although retributivism and severe punishment, such as the death penalty, would be ruled out, they argue that the imposition of sanctions could serve purposes other than the punishment of the guilty—e.g., it can also be justified by its role in incapacitating, rehabilitating, and deterring offenders. (shrink)

Galen Strawson has published several versions of an argument to the effect that moral responsibility is impossible, whether determinism is true or not. Few philosophers have been persuaded by the argument, which Strawson remarks is often dismissed “as wrong, or irrelevant, or fatuous, or too rapid, or an expression of metaphysical megalomania.” I offer here a two-part explanation of why Strawson’s argument has impressed so few. First, as he usually states it, the argument is lacking at least one key premise. (...) The premise in question concerns the very point on which Strawson and many of his contemporary opponents disagree. Strawson will persuade these opponents only when he convinces them of the truth of this crucial premise. Second, Strawson employs a striking conception of responsibility that has not generally been remarked upon. Many might accept that responsibility so conceived is impossible. But there is very different conception of responsibility that appears of great importance to us. To be thoroughly convincing, Strawson will have to show either that this second conception is inadequate or that his argument goes through even given this alternative conception. (shrink)

I’ve been told that in the good old days of the 1970s, when Quine’s desert landscapes were regarded as ideal real estate and David Lewis and John Rawls had not yet left a legion of influential students rewriting the terrain of metaphysics and ethics respectively, compatibilism was still compatibilism about free will. And, of course, incompatibilism was still incompatibilism about free will. That is, compatibilism was the view that free will was compatible with determinism. Incompatibilism was the view that free (...) will was incompatible with determinism.1 What philosophers argued about was whether free will was compatible with determinism. Mostly, this was an argument about how to understand claims that one could do otherwise. You needn’t have bothered to talk about moral responsibility, because it was just obvious that you couldn’t have moral responsibility without free will. The literature was a temple of clarity. Then, somehow, things began to go horribly wrong. To be sure, there had been some activity in the 1960s that would have struck some observers as ominous. Still, it was not until the 1980s that those initial warning signs gave way to real trouble. The meanings of terms twisted. (shrink)

In this paper I give a novel argument for this view that the AGENT concept has an externalist semantics. The argument argues the conclusion from two premises: first, that our first relationships to agents is through a subpersonal mechanism which requires for its function an agential proto-concept which refers directly; and second, that there is a continuity of reference between this proto-concept and the mature concept AGENT. I argue the first on the basis of results in the developmental psychology of (...) social cognition. I argue the second on the basis of a process of elimination, by considering three possibilities for the relationship between the two concepts. On the basis of these two premises the conclusion is drawn that AGENT is a concept that refers directly. That has the following consequences for the philosophy of action: first, that “action” is not an appropriate term for reductive analysis, as the causal theory of action assumes; and second, that we should be looking to the appropriate empirical disciplines for an elucidation of the concept. (shrink)

There are two standard theoretical responses to putative errors in ordinary thinking about some given target property: eliminativism or revisionism. Roughly, eliminativism is the denial that the target property exists, and revisionism is the view that the property exists, but that people tend to have false beliefs about it. Recently, Shaun Nichols has proposed a third option: discretionism. Discretionism is the idea that some terms have multiple reference conventions, so that it may be true to say with eliminativists that the (...) property does not exist, and true to say with revisionists that the property does exist. This article explores the viability of discretionism, and argues that it faces serious difficulties. Even if the difficulties faced by discretionism can be overcome, it is unclear that discretionism secures anything beyond what is already available to standard revisionist views. The article concludes with some reflections about Nichols’ account of the bare retributive norm. (shrink)

Revisionism about moral responsibility is the view that we would do well to distinguish between what we think about moral responsibility and what we ought to think about it, that the former is in some important sense implausible and conflicts with the latter, and so we should revise our concept accordingly. In this paper, I assess two related problems for revisionism and claim that focus on the first of these problems has thus far allowed the second to go largely unnoticed. (...) Here I develop this new objection to revisionism and argue that, while revisionists can successfully respond to the reference-anchoring problem, the normativity-anchoring problem poses a serious objection to the view. In particular, the methodological commitments used to motivate revisionism make it uniquely difficult for revisionists to justify our continued participation in the practice of moral praising and blaming. I conclude by briefly addressing a potential objection based on a common charge against revisionism: that there is no real difference between the view and its conventional competitors and thus the normativity-anchoring problem is of little interest in the broader dialectic. I argue that both of these claims are false. (shrink)

The present chapter is concerned with revisionism about free will. It begins by offering a new characterization of revisionist accounts and the way such accounts fit (or do not) in the familiar framework of compatibilism and incompatibilism. It then traces some of the recent history of the development of revisionist accounts, and concludes by remarking on some challenges for them.