This essay argues that current theories of action fail to explain agentive control because they have left out a psychological capacity central to control: attention. This makes it impossible to give a complete account of the mental antecedents that generate action. By investigating attention, and in particular the intention-attention nexus, we can characterize the functional role of intention in an illuminating way, explicate agentive control so that we have a uniform explanation of basic cases of (...) causal deviance in action as well as other defects of agency, explain cases of skilled agency and sharpen questions about the role of thought in agency. This provides for a different orientation in the theory of action. (shrink)

When performing a skilled action—whether something impressive like a double somersault or something mundane like reaching for a glass of water—you exercise control over your bodily movements. Specifically, you guide their course. In what does that control consist? In this dissertation, I argue that it consists in attending to what you are doing. More specifically, in attending, agents harness their perceptual and perceptuomotor states directly and practically in service of their goals and, in doing so, settle the fine-grained (...) manner in which their bodies will move—details an intention alone leaves unsettled. This requires, among other things, that we reject views on which agents’ control is identical with their practical rationality. When all goes well, agents attentionally prioritize what is motivationally relevant to them to the exclusion of what would otherwise distract them from achieving their goals. However, sometimes agents attend distractedly—i.e., without prioritizing. As the aim of attention is to avoid distraction, this entails the possibility of defective attention. Defective attention, in turn, casts light on scenarios in which agents lose control over what they are doing, as when a skilled practitioner ‘chokes under pressure’. A complaint sometimes levelled against accounts, like mine, that claim to reduce agents’ control of their behaviour to that of causally efficacious mental states or events is that these accounts invariably deprive agents themselves of their rightful role in the generation of behaviour. This is the “Disappearing Agent Problem” for “reductive” or “event-causal” theories of action. I argue that, correctly understood, extant reductive theories do face a genuine Disappearing Agent Problem. However, it is a problem we solve by recognizing the role that conscious attention plays in making an action the agent’s own. Accordingly, I develop and defend an attentional account of action ownership. On this view, allocating conscious attention in service of your goals is sufficient for a kind of conscious perspective (“motivational perspective”), which, when active in controlling your behaviour, constitutes the behaviour as your own doing. As I explain, such perspective also contributes to explaining the subjective structure of your perceptual awareness of the world around you. (shrink)

Mere capacity views hold that agents who can intervene in an unfolding movement are performing an agentially controlled action, regardless of whether they do intervene. I introduce a simple argument to show that the noncausal explanation offered by mere capacity views fails to explain both control and action. In cases where bodily subsystems, rather than the agent, generate control over a movement, agents can often intervene to override non-agential control. Yet, contrary to what capacity views suggest, in (...) these cases, this capacity to intervene does not amount to agential control or action. I illustrate this with a case study of how passive breathing, a mere behavior, is misclassified by mere capacity views. I end by revisiting the central alternative to mere capacity views: causal control views. Advances in our understanding of how agents exert control over unfolding movements indicate that the nature of control is characterized by ubiquitous, small-scale causal interventions. (shrink)

Agent-regret seems to give rise to a philosophical puzzle. If we grant that we are not morally responsible for consequences outside our control (the ‘Standard View’), then agent-regret—which involves self-reproach and a desire to make amends for consequences outside one’s control—appears rationally indefensible. But despite its apparent indefensibility, agent-regret still seems like a reasonable response to bad moral luck. I argue here that the puzzle can be resolved if we appreciate the role that agent-regret plays in a larger (...) social practice that helps us deal with bad moral luck. That agent-regret is a component in a social practice limits the questions that we can reasonably ask about it. While we can ask whether an experience of agent-regret is rational given the norms of this practice, we cannot ask the question that motivates the puzzle of agent-regret, viz. whether agent-regret is rationally defensible according to the Standard View. (shrink)

My primary aim is to defend a nonreductive solution to the problem of action. I argue that when you are performing an overt bodily action, you are playing an irreducible causal role in bringing about, sustaining, and controlling the movements of your body, a causal role best understood as an instance of agent causation. Thus, the solution that I defend employs a notion of agent causation, though emphatically not in defence of an account of free will, as most theories of (...) agent causation are. Rather, I argue that the notion of agent causation introduced here best explains how it is that you are making your body move during an action, thereby providing a satisfactory solution to the problem of action. (shrink)

The book, Autonomous Agents: From Self-Control to Autonomy (1995), by Alfred R. Mele, deals primarily with two main concepts, “self-control” and “individual autonomy,” and the relationship between them. The book is divided into two parts: (1) a view of self-control, the self-controlled person, and behaviour manifesting self-control, and (2) a view of personal autonomy, the autonomous person, and autonomous behaviour. Mele (Ibid.) defines self-control as the opposite of the Aristotelian concept of akrasia, or the contrary (...) of akrasia, which implies weakness of will, incontinence, or lack of self-control—the state of mind in which one acts against one’s better judgement. According to Mele, the concept of self-control can be approached from two perspectives: (a) how self-control affects human behaviour, and (b) how self-control-associated behavior can enhance our understanding of ‘personal autonomy’ and ‘autonomous behaviour’—personal autonomy requires self-control, and autonomous persons and autonomous behaviour are naturally found together. Therefore, I might say that self-control is essential to enhancing one’s autonomy. -/- In part I, we find an account of self-control where Mele argues that even an ideally self-controlled person might lack autonomy. In part II, Mele gives an explicit account of autonomy and explains what must be added to self-control to achieve autonomy. This is the pivotal claim made by Mele (dismantling the intuitively connected notions of self-control and autonomy). (shrink)

Sometimes you make a choice. Whether or not you made it was up to you. The choice was free. But how can this be? A scientific view of the world may leave no room for free choice. Free will literature continually explodes. Yet experts still focus on control or on a power to do otherwise. Sadly, they neglect another intuitive feature of free will: being an underived source or ultimate originator. When acting freely, one is a self-determined, self-directed, sole (...) author of change. Surprisingly, though the literature is replete with this pre-theoretic feature of free will, virtually no one makes rigorous use of it when constructing arguments. Equipped with fresh research, this book contends that agent-causation—currently an unpopular theory—deserves serious reconsideration. Only agent-causation accounts for ultimate origination, and popular objections fail under rigorous scrutiny. This book is tightly argued. It challenges and deepens recent advances in free will scholarship made by Peter van Inwagen, Robert Kane, Carl Ginet, Randy Clarke, Stewart Goetz, Tim O’Connor, and others. (shrink)

In a recent paper I argued that agent causation theorists should be compatibilists. In this paper, I argue that compatibilists should be agent causation theorists. I consider six of the main problems facing compatibilism: (i) the powerful intuition that one can't be responsible for actions that were somehow determined before one was born; (ii) Peter van Inwagen's modal argument, involving the inference rule (β); (iii) the objection to compatibilism that is based on claiming that the ability to do otherwise is (...) a necessary condition for freedom; (iv) "manipulation arguments," involving cases in which an agent is manipulated by some powerful being into doing something that he or she would not normally do, but in such a way that the compatibilist's favorite conditions for a free action are satisfied; (v) the problem of constitutive luck; and (vi) the claim that it is not fair to blame someone for an action if that person was determined by forces outside of his or her control to perform that action. And in the case of each of these problems, I argue that the compatibilist has a much more plausible response to that problem if she endorses the theory of agent causation than she does otherwise. (shrink)

In this paper I argue that many scholars involved in the contemporary free will debates have underappreciated the philosophical appeal of agent causation because the resources of contemporary emergentism have not been adequately introduced into the discussion. Whereas I agree that agent causation’s main problem has to do with its intelligibility, particularly with respect to the issue of how substances can be causally relevant, I argue that the notion of substance causation can be clearly articulated from an emergentist framework. According (...) to my proposal, a free agent is _a causally powerful substance that emerges in an anomic way from her constitutive mental events_, downwardly constraining, selecting and, in this way, having control on them. As we shall see, this particular concept of agent causation not only makes sense of the deep insight behind agent libertarianism, but it also provides us with the resources to solve some of the main objections that have been raised against it. It is true that here I cannot develop a complete defense of the evidential credentials of emergentism. Still, even if the considerations that follow do not serve to convince detractors of agent causal libertarian accounts of free will, they do suggest that libertarian agent causation is more promising than is typically acknowledged. (shrink)

Plausibly, agents act freely iff their actions are responses to reasons. But what sort of relationship between reason and action is required for the action to count as a response? The overwhelmingly dominant answer to this question is modalist. It holds that responses are actions that share a modally robust or secure relationship with the relevant reasons. This thesis offers a new alternative answer. It argues that responses are actions that can be explained by reasons in the right way. This (...) explanationist answer comes apart from the modalist answer. For it holds that actions are responses to reasons if they are explained by those reasons even if they don’t share a modally robust relationship. Explanationism thus offers a novel way of vindicating the intuition that alternative possibilities don’t matter to responding to reasons and (consequently) free agency. The key dialectical position the thesis develops is that both modalism and explanationism constitute attempts to capture he core type of relationship encoded by the notion of a response. Responses to reasons, at core, involve a non-accidental relationship between reason and action. We can either understand non-accidentality as a modal phenomenon – as a modally robust tracking between two facts. Or we can understand non-accidentality as an explanatory phenomenon – as a special explanatory relationship between two facts. According to my rival explanationist proposal, two facts share a non-accidental relationship iff we can give a unified explanation of why both obtain. Unified explanations are explanations of why [p&q] that cannot be decomposed into two (or more) separate independent explanations of p and of q. Consequently, according to explanationism about responding to reasons, actions are responses to reasons iff those reasons offer a rational explanation of the action that cannot be decomposed into separate independent components. (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)

Judgments of blame for others are typically sensitive to what an agent knows and desires. However, when people act negligently, they do not know what they are doing and do not desire the outcomes of their negligence. How, then, do people attribute blame for negligent wrongdoing? We propose that people attribute blame for negligent wrongdoing based on perceived mental control, or the degree to which an agent guides their thoughts and attention over time. To acquire information about others’ mental (...) control, people self-project their own perceived mental control to anchor third-personal judgments about mental control and concomitant responsibility for negligent wrongdoing. In four experiments (N = 841), we tested whether perceptions of mental control drive third-personal judgments of blame for negligent wrongdoing. Study 1 showed that the ease with which people can counterfactually imagine an individual being non-negligent mediated the relationship between judgments of control and blame. Studies 2a and 2b indicated that perceived mental control has a strong effect on judgments of blame for negligent wrongdoing and that first-personal judgments of mental control are moderately correlated with third-personal judgments of blame for negligent wrongdoing. Finally, we used an autobiographical memory manipulation in Study 3 to make personal episodes of forgetfulness salient. Participants for whom past personal episodes of forgetfulness were made salient judged negligent wrongdoers less harshly compared to a control group for whom past episodes of negligence were not salient. Collectively, these findings suggest that first-personal judgments of mental control drive third-personal judgments of blame for negligent wrongdoing and indicate a novel role for counterfactual thinking in the attribution of responsibility. (shrink)

The Shape of Agency offers interlinked explanations of the basic building blocks of agency, as well as its exemplary instances. The first part offers accounts of a collection of related phenomena that have long troubled philosophers of action: control over behaviour, non-deviant causation, and intentional action. These accounts build on earlier work in the causalist tradition, and undermine the claims made by many that causalism cannot offer a satisfying account of non-deviant causation, and therefore fails as an account of (...) intentional action. The second part turns to modes of agentive excellence—ways that agents display quality of form—providing a novel account of skill, including an account of the ways that agents display more or less skill. Shepherd discusses the role of knowledge in skill, and concludes that while knowledge is often important, it is inessential. This leads to a discussion of the way that knowledge of action and knowledge of how to act informs action execution. Knowledgeable action includes a unique epistemic underpinning: in knowledgeable action, the agent has authoritative knowledge of what she is doing and how she is doing it when and because she is poised to control her action by way of practical reasoning. (shrink)

According to some collectivists, purposive groups that lack decision-making procedures such as riot mobs, friends walking together, or the pro-life lobby can be morally responsible and have moral duties. I focus on plural subject- and we-mode-collectivism. I argue that purposive groups do not qualify as duty-bearers even if they qualify as agents on either view. To qualify as a duty-bearer, an agent must be morally competent. I develop the Update Argument. An agent is morally competent only if the agent has (...) sufficient positive and negative control over updating their goal-seeking states. Positive control involves the general ability to update one’s goal-seeking states, whereas negative control involves the absence of other agents with the capacity to arbitrarily interfere with updating one’s goal-seeking states. I argue that even if purposive groups qualify as plural subjects or we-mode group agents, these groups necessarily lack negative control over updating their goal-seeking states. This creates a cut-off point for groups as duty-bearers: Organized groups may qualify as duty-bearers, whereas purposive groups cannot qualify as duty-bearers. (shrink)

Researchers often claim that self-control is a skill. It is also often stated that self-control exertions are intentional actions. However, no account has yet been proposed of the skillful agency that makes self-control exertion possible, so our understanding of self-control remains incomplete. Here I propose the skill model of self-control, which accounts for skillful agency by tackling the guidance problem: how can agents transform their abstract and coarse-grained intentions into the highly context-sensitive, fine-grained control (...) processes required to select, revise and correct strategies during self-control exertion? The skill model borrows conceptual tools from ‘hierarchical models’ recently developed in the context of motor skills, and asserts that self-control crucially involves the ability to manage the implementation and monitoring of regulatory strategies as the self-control exercise unfolds. Skilled agents are able do this by means of flexible practical reasoning: a fast, context-sensitive type of deliberation that incorporates non-propositional representations into the formation and revision of the mixed-format intentions that structure self-control exertion. The literatures on implementation intentions and motivation framing offer corroborating evidence for the theory. As a surprising result, the skill of self-control that allows agents to overcome the contrary motivations they experience is self-effacing: instead of continuously honing this skill, expert agents replace it with a different one, which minimizes or prevents contrary motivations from arising in the first place. Thus, the more expert you are at self-control, the less likely you are to use it. (shrink)

It is tempting to think (though many deny) that epistemic agents exercise a distinctive kind of control over their belief‐like attitudes. My aim here is to sketch a “bottom‐up” model of epistemic agency, one that draws on an analogous model of practical agency, according to which an agent's conditional beliefs are reasons‐responsive planning states that initiate and sustain mental behavior so as to render controlled.

Mere capacity views hold that agents who can intervene in an unfolding movement are performing an agentially controlled action, regardless of whether they do intervene. I introduce a simple argument to show that the noncausal explanation offered by mere capacity views fails to explain both control and action. In cases where bodily subsystems, rather than the agent, generate control over a movement, agents can often intervene to override non-agential control. Yet, contrary to what capacity views suggest, in (...) these cases, this capacity to intervene does not amount to agential control or action. I illustrate this with a case study of how passive breathing, a mere behavior, is misclassified by mere capacity views. I end by revisiting the central alternative to mere capacity views: causal control views. Advances in our understanding of how agents exert control over unfolding movements indicate that the nature of control is characterized by ubiquitous, small-scale causal interventions. (shrink)

Are there are normative reasons to have – or refrain from having – certain emotions? The dominant view is that there are. I disagree. In this paper, I argue for Strong Eliminativism – the view that there are no reasons for emotions. My argument for this claim has two premises. The first premise is that there is a deliberative constraint on reasons: a reason for an agent to have an attitude must be able to feature in that agent’s deliberation to (...) that attitude. My argument for this premise is that in order to have reasons for an attitude, we need to be able to exhibit some relevant form of control over this attitude, and this relevant form of control is deliberative control. The second premise is that no one can deliberate to any emotion. My argument for this premise turns on the claim that there is no deliberative question that is settled by forming (or giving up) an emotion. I contend that this is so due to the well-known phenomenon of recalcitrant emotions: for any deliberative question that can be settled, there is no guarantee that the relevant emotional state will follow (or be revised). Strong Eliminativism follows from these two premises. (shrink)

In this paper I provide an exposition and critique of Johnson and Noorman’s (2014) three conceptualizations of the agential roles artificial systems can play. I argue that two of these conceptions are unproblematic: that of causally efficacious agency and “acting for” or surrogate agency. Their third conception, that of “autonomous agency,” however, is one I have reservations about. The authors point out that there are two ways in which the term “autonomy” can be used: there is, firstly, the engineering sense (...) of the term, which simply refers to the ability of some system to act independently of substantive human control. Secondly, there is the moral sense of the term, which has traditionally grounded many notions of human mor-al responsibility. I argue that the continued usage of “autonomy” in discussions of artificial agency complicates matters unnecessarily. This occurs in two ways: firstly, the condition of autonomy, even in its engineering sense, fails to accurately describe the way “autonomous” systems are developed in practice. Secondly, the continued usage of autonomy in the moral sense introduces unnecessary meta-physical baggage form the free will debate into discussions about moral agency. In order to understand the debate surrounding autonomy, we would therefore first need to settle many seemingly intractable metaphysical questions regarding the existence of free will in human beings. (shrink)

”Meaningful human control” is a term invented in the political and legal debate on autonomous weapons system, but it is nowadays also used in many other contexts. It is supposed to specify conditions under which an artificial system is under the right kind of control to avoid responsibility gaps: that is, situations in which no moral agent is responsible. Santoni de Sio and Van den Hoven have recently suggested a framework that can be used by system designers to (...) operationalize this kind of control. It is the purpose of this paper to facilitate further operationalization of ”meaningful human control”. -/- This paper consists of two parts. In the first part I resolve an ambiguity that plagues current operationalizations of MHC. One of the design conditions says that the system should track the reasons of the relevant agents. This condition is ambiguous between the kind of reasons involved. On one interpretation it says that a system should track motivating reasons, while it is concerned with normative reasons on the other. Current participants in the debate interpret the framework as being concerned with (something in the vicinity of) motivating reasons. I argue against this interpretation by showing that meaningful human control requires that a system tracks normative reasons. Moreover, I maintain that an operationalization of meaningful human control that fails to track the right kind of reasons is morally problematic.When this is properly understood, it can be shown that the framework of MHC is committed to the agent-relativity of reasons. More precisely, I argue in the second part of this paper that if the tracking condition of MHC plays an important role in responsibility attribution (as the proponents of the view maintain), then the framework is incompatible with first-order normative theories that hold that normative reasons are agent-neutral (such as many versions of consequentialism). In the final section I present three ways forward for the proponent of MHC as reason-responsiveness. (shrink)

This paper introduces a new family of cases where agents are jointly morally responsible for outcomes over which they have no individual control, a family that resists standard ways of understanding outcome responsibility. First, the agents in these cases do not individually facilitate the outcomes and would not seem individually responsible for them if the other agents were replaced by non-agential causes. This undermines attempts to understand joint responsibility as overlapping individual responsibility; the responsibility in question is essentially joint. (...) Second, the agents involved in these cases are not aware of each other's existence and do not form a social group. This undermines attempts to understand joint responsibility in terms of actual or possible joint action or joint intentions, or in terms of other social ties. Instead, it is argued that intuitions about joint responsibility are best understood given the Explanation Hypothesis, according to which a group of agents are seen as jointly responsible for outcomes that are suitably explained by their motivational structures: something bad happened because they didn’t care enough; something good happened because their dedication was extraordinary. One important consequence of the proposed account is that responsibility for outcomes of collective action is a deeply normative matter. (shrink)

A standard account of privacy says that it is essentially a kind of control over personal information. Many privacy scholars have argued against this claim by relying on so-called threatened loss cases. In these cases, personal information about an agent is easily available to another person, but not accessed. Critics contend that control accounts have the implausible implication that the privacy of the relevant agent is diminished in threatened loss cases. Recently, threatened loss cases have become important because (...) Edward Snowden’s revelation of how the NSA and GCHQ collected Internet and mobile phone data presents us with a gigantic, real-life threatened loss case. In this paper, I will defend the control account of privacy against the argument that is based on threatened loss cases. I will do so by developing a new version of the control account that implies that the agents’ privacy is not diminished in threatened loss cases. (shrink)

Agentive awareness is one's awareness of oneself as presently acting. Dominant accounts in cognitive science consider agentive awareness to be grounded in the states and processes underlying sensorimotor control. In this paper, I raise concerns for this approach and develop an alternative. Broadly, in the approach I defend, one is agentively aware in the virtue of intending to act. I further argue that agentive awareness is not constituted by intentions themselves but rather first-personal thoughts that are (...) formed on the basis of them. I develop this proposal, highlight some of its theoretical advantages, and show how it successfully meets various challenges. (shrink)

Necessarily, if S lacks the ability to exercise control, S is not an agent. If S is not an agent, S cannot act intentionally, responsibly, or rationally, nor can S possess or exercise free will. In spite of the obvious importance of control, however, no general account of control exists. In this paper I reflect on the nature of control itself. I develop accounts of control ’s exercise and control ’s possession that illuminate what (...) it is for degrees of control —that is, the degree of control an agent possesses or exercises in a given circumstance—to vary. Finally, I demonstrate the usefulness of the account on offer by showing how it generates a solution to a long-standing problem for causalist theories of action, namely, the problem of deviant causation. (shrink)

This dissertation is a defence of agent-regret and an exploration of its role in our lives. I argue that agent-regret shows that an agent takes seriously her status as an agent who impacts the world, but who only has fallible control over it. To accept responsibility for any outcomes, she must accept responsibility for unintended outcomes, too: agent-regret is part of being a human agent. In doing this, I try to defend and develop Williams’s own conception of agent-regret. -/- (...) In the first part, I explore the nature of agent-regret. Agent-regret is distinct from guilt because we can feel agent-regret without being at fault. I argue that several challenges that seek to reduce agent-regret to a form of guilt fail. I further argue that agent-regret takes as its object not only something one has done, but the fact that one did it; I discuss this in terms of what Bernard Williams called taking an “external” view on one’s own action. -/- I suggest that we can best understand the object of agent-regret as one’s responsibility for an outcome. I argue that this form of “responsibility” is conceptually separate from liability or answerability; it concerns whether the outcome can be ascribed to an agent. This is a restricted form of causal responsibility but retains its agential character. I suggest that we can be responsible for outcomes even when we did not intend to bring them about. -/- In the second part, I vindicate the propriety of agent-regret against the ideas that we are not responsible for unintended outcomes or that such responsibility is not important. I set out several challenges to this effect. I argue that we are responsible for unintended outcomes because in order to be responsible as agents at all, we must use fallible abilities. When we exercise these abilities but fail, we are responsible as agents for those unintended outcomes that arise. I then consider the importance of being responsible for particular unintended outcomes and analyse the idea that this affects our identities. I argue that responsibility for an outcome affects a form of identity, but that it does not involve essential features of a person. Instead, our responsibility for outcomes is a contingent feature that nonetheless plays an important role in our interpersonal interactions and self-conceptions. I argue that these reactions are appropriate, because our responsibility for particular outcomes is independently significant—but these reactions also lend this responsibility added significance as part of inescapable human practices. -/- Thus our responsibility for outcomes is important. Its importance means that it can serve as an appropriate object of agent-regret that vindicates the propriety of agent-regret. I end by considering several interesting features of agent-regret, including its expression and “pure” cases, which are cases where an agent feels agent-regret despite not regretting the result of her action. (shrink)

Where artificial agents are not liable to be ascribed true moral agency and responsibility in their own right, we can understand them as acting as proxies for human agents, as making decisions on their behalf. What I call the ‘Moral Proxy Problem’ arises because it is often not clear for whom a specific artificial agent is acting as a moral proxy. In particular, we need to decide whether artificial agents should be acting as proxies for low-level agents — e.g. individual (...) users of the artificial agents — or whether they should be moral proxies for high-level agents — e.g. designers, distributors or regulators, that is, those who can potentially control the choice behaviour of many artificial agents at once. Who we think an artificial agent is a moral proxy for determines from which agential perspective the choice problems artificial agents will be faced with should be framed: should we frame them like the individual choice scenarios previously faced by individual human agents? Or should we, rather, consider the expected aggregate effects of the many choices made by all the artificial agents of a particular type all at once? This paper looks at how artificial agents should be designed to make risky choices, and argues that the question of risky choice by artificial agents shows the moral proxy problem to be both practically relevant and difficult. (shrink)

Our focus here is on whether, when influenced by implicit biases, those behavioural dispositions should be understood as being a part of that person’s character: whether they are part of the agent that can be morally evaluated.[4] We frame this issue in terms of control. If a state, process, or behaviour is not something that the agent can, in the relevant sense, control, then it is not something that counts as part of her character. A number of theorists (...) have argued that individuals do not have control, in the relevant sense, over the operation of implicit bias. We will argue that this claim is mistaken. We articulate and develop a notion of control that individuals have with respect to implicit bias, and argue that this kind of control can ground character-based evaluation of such behavioural dispositions. (shrink)

From our everyday commuting to the gold medalist’s world-class performance, skillful actions are characterized by fine-grained, online agentive control. What is the proper explanation of such control? There are two traditional candidates: intellectualism explains skillful agentive control by reference to the agent’s propositional mental states; anti-intellectualism holds that propositional mental states or reflective processes are unnecessary since skillful action is fully accounted for by automatic coping processes. I examine the evidence for three psychological phenomena recently (...) held to support anti-intellectualism and argue that it supports neither traditional candidate, but an intermediate attention-control account, according to which the top-down, intention-directed control of attention is a necessary component of skillful action. Only this account recognizes both the role of automatic control in skilled action and the need for higher-order cognition to thread automatic processes together into a unified, skillful performance. This applies to bodily skillful action in general, from the world-class performance of experts to mundane, habitual action. The attention-control account stresses that, for intentions to play their role as top-down modulators of attention, agents must sustain the intention’s activation; hence, the need for reflecting throughout performance. (shrink)

Virtue ethics has many times been suggested as a promising recipe for the construction of artificial moral agents due to its emphasis on moral character and learning. However, given the complex nature of the theory, hardly any work has de facto attempted to implement the core tenets of virtue ethics in moral machines. The main goal of this paper is to demonstrate how virtue ethics can be taken all the way from theory to machine implementation. To achieve this goal, we (...) critically explore the possibilities and challenges for virtue ethics from a computational perspective. Drawing on previous conceptual and technical work, we outline a version of artificial virtue based on moral functionalism, connectionist bottom–up learning, and eudaimonic reward. We then describe how core features of the outlined theory can be interpreted in terms of functionality, which in turn informs the design of components necessary for virtuous cognition. Finally, we present a comprehensive framework for the technical development of artificial virtuous agents and discuss how they can be implemented in moral environments. (shrink)

ABSTRACTThe contemporary debate over responsibility for belief is divided over the issue of whether such responsibility requires doxastic control, and whether this control must be voluntary in nature. It has recently become popular to hold that responsibility for belief does not require voluntary doxastic control, or perhaps even any form of doxastic ‘control’ at all. However, Miriam McCormick has recently argued that doxastic responsibility does in fact require quasi-voluntary doxastic control: “guidance control,” a complex, (...) compatibilist form of control. In this paper, I pursue a negative and a positive task. First, I argue that grounding doxastic responsibility in guidance control requires too much for agents to be the proper targets for attributions of doxastic responsibility. I will focus my criticisms on three cases in which McCormick's account gives the intuitively wrong verdict. Second, I develop a modified conception of McCormick's notion of “ownership of belief,” which I call Weak Doxastic Ownership. I employ this conception to argue that responsibility for belief is possible even in the absence of guidance control. In doing so, I argue that the notion of doxastic ownership can do important normative work in grounding responsibility for belief without being subsumed under or analyzed in terms of the notion of doxastic control. (shrink)

Incompatibilists and compatibilists (mostly) agree that there is a strong intuition that a manipulated agent, i.e., an agent who is the victim of methods such as indoctrination or brainwashing, is unfree. They differ however on why exactly this intuition arises. Incompatibilists claim our intuitions in these cases are sensitive to the manipulated agent’s lack of ultimate control over her actions, while many compatibilists argue that our intuitions respond to damage inflicted by manipulation on the agent’s psychological and volitional capacities. (...) Much hangs on this issue because manipulation-based arguments are among the most important for defending incompatibilist views of free will. In this paper, I investigate this issue from a experimental perspective, using a set of statistical methods well suited for identifying the features of hypothetical cases people’s intuitions are responding to. Results strongly support the compatibilist view—subjects’ tendency to judge that a manipulated agent is unfree was found to depend on their judgments that the agent suffers impairments to certain psychological/volitional capacities that compatibilists say are the basis for free will. I discuss the significance of these results for the use of manipulation cases in the philosophical debate about free will. (shrink)

In philosophy of mind, zombies are imaginary creatures that are exact physical duplicates of conscious subjects but for whom there is no first-personal experience. Zombies are meant to show that physicalism—the theory that the universe is made up entirely out of physical components—is false. In this paper, I apply the zombie thought experiment to the realm of morality to assess whether moral agency is something independent from sentience. Algorithms, I argue, are a kind of functional moral zombie, such that thinking (...) about the latter can help us better understand and regulate the former. I contend that the main reason why algorithms can be neither autonomous nor accountable is that they lack sentience. Moral zombies and algorithms are incoherent as moral agents because they lack the necessary moral understanding to be morally responsible. To understand what it means to inflict pain on someone, it is necessary to have experiential knowledge of pain. At most, for an algorithm that feels nothing, ‘values’ will be items on a list, possibly prioritised in a certain way according to a number that represents weightiness. But entities that do not feel cannot value, and beings that do not value cannot act for moral reasons. (shrink)

Weighing competing obligations and achieving the “greatest balance” of right over wrong guides an individual, an agency or a country in determining what ought to be done in an ethically challenging situation. Conducting controlled human infection model (CHIM) studies in India is one such situation. The ethical challenge in conducting a CHIM study lies in completing the difficult task of introducing standardised, attenuated strains of micro-organisms into normal healthy volunteers, at the same time ensuring the safety of these healthy individuals (...) from potential and completely informed risks in a fashion that is transparent and accountable. The bar is further raised against the background of already fragile public confidence in biomedical research in India; especially when “deliberate” introduction of microbial agents into healthy individuals is involved, with the larger altruistic objective of gain to society as a whole. -/- This paper discusses the uses of CHIM studies with respect to the larger scientific Indian research enterprise of the 21st century. It further explores etic and emic perspectives in conducting such trials in India and seeks to generate an ethical coherence to the justification for conducting CHIM studies in India. The paper deliberates on ethical issues arising out of conducting CHIM studies and reflects on how developing the capacity for CHIM studies in India is likely to strengthen the health research and development sector in the country. -/- . (shrink)

An agent exercises instrumental rationality to the degree that she adopts appropriate means to achieving her ends. Adopting appropriate means to achieving one’s ends can, in turn, involve overcoming one’s strongest desires, that is, it can involve exercising synchronic self-control. However, contra prominent approaches, I deny that synchronic self-control is possible. Specifically, I draw on computational models and empirical evidence from cognitive neuroscience to describe a naturalistic, multi-system model of the mind. On this model, synchronic self-control is (...) impossible. Must we, then, give up on a meaningful conception of instrumental rationality? No. A multi-system view still permits something like synchronic self-control: an agent can control her very strong desires. Adopting a multi-system model of the mind thus places limitations on our conceptions of instrumental rationality, without requiring that we abandon the notion altogether. (shrink)

Due to intensive agriculture, rapid industrialization and anthropogenic activities have caused environmental pollution, land degradation and increased pressure on the natural resources and contributing to their adulteration. Bioremediation is the use of biological organisms to destroy, or reduce the hazardous wastes on a contaminated site. Bioremediation is the most potent management tool to control the environmental pollution and recover contaminated soil. Use of biological materials, coupled to other advanced processes is one of the most promising and inexpensive approaches for (...) removing environmental pollutants. Bioremediation technology is a beneficial alternative which leads to degrade of pollutants. This article presents the important biological organisms used in bioremediation technologies. (shrink)

Although virtue ethics has repeatedly been proposed as a suitable framework for the development of artificial moral agents, it has been proven difficult to approach from a computational perspective. In this work, we present the first technical implementation of artificial virtuous agents in moral simulations. First, we review previous conceptual and technical work in artificial virtue ethics and describe a functionalistic path to AVAs based on dispositional virtues, bottom-up learning, and top-down eudaimonic reward. We then provide the details of a (...) technical implementation in a moral simulation based on a tragedy of the commons scenario. The experimental results show how the AVAs learn to tackle cooperation problems while exhibiting core features of their theoretical counterpart, including moral character, dispositional virtues, learning from experience, and the pursuit of eudaimonia. Ultimately, we argue that virtue ethics provides a compelling path toward morally excellent machines and that our work provides an important starting point for such endeavors. (shrink)

This article provides an answer to the question: What is the function of cognition? By answering this question it becomes possible to investigate what are the simplest cognitive systems. It addresses the question by treating cognition as a solution to a design problem. It defines a nested sequence of design problems: (1) How can a system persist? (2) How can a system affect its environment to improve its persistence? (3) How can a system utilize better information from the environment to (...) select better actions? And, (4) How can a system reduce its inherent informational limitations to achieve more successful behavior? This provides a corresponding nested sequence of system classes: (1) autonomous systems, (2) (re)active autonomous systems, (3) informationally controlled autonomous systems (autonomous agents), and (4) cognitive systems. -/- This article provides the following characterization of cognition: The cognitive system is the set of mechanisms of an autonomous agent that (1) allow increase of the correlation and integration between the environment and the information system of the agent, so that (2) the agent can improve the selection of actions and thereby produce more successful behavior. -/- Finally, it shows that common cognitive capacities satisfy the characterization: learning, memory, representation, decision making, reasoning, attention, and communication. (shrink)

Two main approaches can be discerned in the literature on agentive self-awareness: a top-down approach, according to which agentive self-awareness is fundamentally holistic in nature and involves the operations of a central-systems narrator, and a bottom-up approach that sees agentive self-awareness as produced by lowlevel processes grounded in the very machinery responsible for motor production and control. Neither approach is entirely satisfactory if taken in isolation; however, the question of whether their combination would yield a full (...) account of agentive self-awareness remains very much open. In this paper, I contrast two disorders affecting the control of voluntary action: the anarchic hand syndrome and utilization behavior. Although in both conditions patients fail to inhibit actions that are elicited by objects in the environment but inappropriate with respect to the wider context, these actions are experienced in radically different ways by the two groups of patients. I discuss how top-down and bottom-up processes involved in the generation of agentive self-awareness would have to be related in order to account for these differences. (shrink)

An interesting aspect of Ernest Sosa’s (2017) recent thinking is that enhanced performances (e.g., the performance of an athlete under the influence of a performance-enhancing drug) fall short of aptness, and this is because such enhanced performances do not issue from genuine competences on the part of the agent. In this paper, I explore in some detail the implications of such thinking in Sosa’s wider virtue epistemology, with a focus on cases of cognitive enhancement. A certain puzzle is then highlighted, (...) and the solution proposed draws from both the recent moral responsibility literature on guidance control (e.g., Fischer and Ravizza 1998; Fischer 2012) as well as from work on cognitive integration in the philosophy of mind and cognitive science (e.g., Clark and Chalmers 1998; Clark 2008; Pritchard 2010; Palermos 2014; Carter 2017). (shrink)

A restrictive view of self-control identifies exercises of self-control with synchronic intrapsychic processes, and pictures diachronic and externally-scaffolded strategies not as proper instances of self-control, but as clever ways of avoiding the need to exercise that ability. In turn, defenders of an inclusive view of self-control typically argue that we should construe self-control as more than effortful inhibition, and that, on grounds of functional equivalence, all these diverse strategies might be properly described as instances of (...) self-control. In this paper, I take a fresh look at this debate by focusing on cases of addiction. I argue that addicted agents face a paradigmatic sort of self-control challenge, which makes addiction an important test case for theories of self-control. And I discuss evidence that highlights both the unreliability of synchronic intrapsychic strategies and the crucial role that is played by diachronic and externally-scaffolded strategies in successful attempts at achieving abstinence by addicted individuals. Abstaining addicts are a paradigmatic case of agents who successfully exercise self-control, and they mostly do so by relying on diachronic and externally-scaffolded strategies. This, I argue, lends further support to an inclusive view of self-control. (shrink)

Nearly all defences of the agent-causal theory of free will portray the theory as a distinctively libertarian one — a theory that only libertarians have reason to accept. According to what I call ‘the standard argument for the agent-causal theory of free will’, the reason to embrace agent-causal libertarianism is that libertarians can solve the problem of enhanced control only if they furnish agents with the agent-causal power. In this way it is assumed that there is only reason to (...) accept the agent-causal theory if there is reason to accept libertarianism. I aim to refute this claim. I will argue that the reasons we have for endorsing the agent-causal theory of free will are nonpartisan. The real reason for going agent-causal has nothing to do with determinism or indeterminism, but rather with avoiding reductionism about agency and the self. As we will see, if there is reason for libertarians to accept the agent-causal theory, there is just as much reason for compatibilists to accept it. It is in this sense that I contend that if anyone should be an agent-causalist, then everyone should be an agent-causalist. (shrink)

Whether we are free agents or not and to what extent depends on factors such as the necessary conditions for free will and our definition of human agency and identity. The present article, apart from possible alternatives and the causality of the agent regarding his actions, addresses the element of inclination as a necessary condition for free will. Therefore, an analysis of these conditions determines that even though in some circumstances the range of alternatives the agent can choose is very (...) limited or the agent chooses an action without having the inclination to choose it; however, this does not make the agent forced; rather, free will has levels and every individual possesses a certain extent of it in different conditions. Just the fact that the ultimate choice is made by the agent himself and has voluntarily performed the act himself ultimately is sufficient for being free. As a result, even though the influence of external factors on human actions cannot be disregarded; however, because the action is ultimately not outside his control, accordingly, there is neither such a thing as a compelled agent and nor are our actions left to luck. The character of the agent that is voluntarily formed through our actions plays a determinant role in our choices and actions due to the epistemic and motivational components they possess. (shrink)

I argue that when determining whether an agent ought to perform an act, we should not hold fixed the fact that she’s going to form certain attitudes (and, here, I’m concerned with only reasons-responsive attitudes such as beliefs, desires, and intentions). For, as I argue, agents have, in the relevant sense, just as much control over which attitudes they form as which acts they perform. This is important because what effect an act will have on the world depends not (...) only on which acts the agent will simultaneously and subsequently perform, but also on which attitudes she will simultaneously and subsequently form. And this all leads me to adopt a new type of practical theory, which I call rational possibilism. On this theory, we first evaluate the entire set of things over which the agent exerts control, where this includes the formation of certain attitudes as well as the performance of certain acts. And, then, we evaluate individual acts as being permissible if and only if, and because, there is such a set that is itself permissible and that includes that act as a proper part. Importantly, this theory has two unusual features. First, it is not exclusively act-orientated, for it requires more from us than just the performance of certain voluntary acts. It requires, in addition, that we involuntarily form certain attitudes. Second, it is attitude-dependent in that it holds that which acts we’re required to perform depends on which attitudes we’re required to form. I then show how these two features can help us both to address certain puzzling cases of rational choice and to understand why most typical practical theories (utilitarianism, virtue ethics, rational egoism, Rossian deontology, etc.) are problematic. (shrink)

A compelling, and popular, thought is that ability entails control: S’s being able to φ entails that φ be, in some sense, in S’s control. But this intuition is inconsistent with a different prima facie compelling thought: that S’s succeeding in φ-ing entails that S is able to φ. In this paper, I introduce a new form of evidence to help adjudicate between these two theses: namely, probability judgments about ability ascriptions. I argue that these judgments provide evidence (...) in favor of the intuition that success entails ability, and against the intuition that ability requires control. Moreover, I argue that these judgments support one particular analysis which vindicates the success intuition, namely, the analysis of ability in terms of conditionals. (shrink)

Interactions between an intelligent software agent and a human user are ubiquitous in everyday situations such as access to information, entertainment, and purchases. In such interactions, the ISA mediates the user’s access to the content, or controls some other aspect of the user experience, and is not designed to be neutral about outcomes of user choices. Like human users, ISAs are driven by goals, make autonomous decisions, and can learn from experience. Using ideas from bounded rationality, we frame these interactions (...) as instances of an ISA whose reward depends on actions performed by the user. Such agents benefit by steering the user’s behaviour towards outcomes that maximise the ISA’s utility, which may or may not be aligned with that of the user. Video games, news recommendation aggregation engines, and fitness trackers can all be instances of this general case. Our analysis facilitates distinguishing various subcases of interaction, as well as second-order effects that might include the possibility for adaptive interfaces to induce behavioural addiction, and/or change in user belief. We present these types of interaction within a conceptual framework, and review current examples of persuasive technologies and the issues that arise from their use. We argue that the nature of the feedback commonly used by learning agents to update their models and subsequent decisions could steer the behaviour of human users away from what benefits them, and in a direction that can undermine autonomy and cause further disparity between actions and goals as exemplified by addictive and compulsive behaviour. We discuss some of the ethical, social and legal implications of this technology and argue that it can sometimes exploit and reinforce weaknesses in human beings. (shrink)

Some of the things we do intentionally we do halfheartedly. I develop and defend an account of halfheartedness with respect to action on which one is halfhearted with respect to an action A if one’s overall motivation to A is weak. This requires getting clear on what it is to have some level of overall motivation with respect to an action, and on what it means to say one’s overall motivation is weak or strong. After developing this account, I defend (...) the claim that one key functional expression of halfhearted action is the possession of impaired control over the action in question. Finally, I elucidate a puzzle that sometimes arises with respect to halfhearted action. The puzzle arises when an agent’s commitment in acting conflicts with an agent’s acceptance of poor performance. (shrink)

The Metaphysics of Free Will provides a through statement of the major grounds for skepticism about the reality of free will and moral responsibility. The author identifies and explains the sort of control that is associated with personhood and accountability, and shows how it is consistent with causal determinism. In so doing, out view of ourselves as morally responsible agents is protected against the disturbing changes posed by science and religion.

Does self-control require willpower? The question cuts to the heart of a debate about whether self-control is identical with some psychological process internal to the agents or not. Noticeably absent from these debates is systematic evidence about the folk-psychological category of self-control. Here, we present the results of two behavioral studies (N = 296) that indicate the structure of everyday use of the concept. In Study 1, participants rated the degree to which different strategies to respond to (...) motivational conflict exemplify self-control. Participants distinguished between intra-psychic and externally-scaffolded strategies and judged that the former exemplified self-control more than the latter. In Study 2, participants provided various solutions to manage motivational conflict and rated their proposals on effectiveness. Participants produced substantially more intra-psychic strategies, rated them as more effective, and advised them at a higher rate than externally-scaffolded strategies. Taken together, these results suggest that while people recognize a plurality of strategies as genuine instances of self-control, purely internal exercises of self-control are considered more prototypical than their externally-scaffolded counterparts. This implies a hierarchical structure for the folk psychological category of self-control. The concept encompasses a variety of regulatory strategies and organizes these strategies along a hierarchical continuum, with purely intra-psychic strategies at the center and scaffolded strategies in the periphery. (shrink)

Michael Moore’s recent book Mechanical Choices: The Responsibility of the Human Machine is full of rich, insightful discussion of many important issues related to free will and moral responsibility. I will focus on one particular issue raised by Moore: the question of whether we can have control over the past. Moore defends a compatibilist account of moral responsibility on which there are some possible cases in which agents do have such control. But Moore seeks to avoid positing too (...) much control over the past. I will argue that Moore’s attempt is not successful. Furthermore, I will try to show that compatibilists in general face a serious challenge regarding control over the past. In sections 1 and 2 I will explore Moore’s views. In section 3 and 4 I will offer a general argument against compatibilism about moral responsibility and determinism. (shrink)

This paper presents a new trilemma with respect to resolving the control and alignment problems in machine ethics. Section 1 outlines three possible types of artificial moral agents (AMAs): (1) 'Inhuman AMAs' programmed to learn or execute moral rules or principles without understanding them in anything like the way that we do; (2) 'Better-Human AMAs' programmed to learn, execute, and understand moral rules or principles somewhat like we do, but correcting for various sources of human moral error; and (3) (...) 'Human-Like AMAs' programmed to understand and apply moral values in broadly the same way that we do, with a human-like moral psychology. Sections 2–4 then argue that each type of AMA generates unique control and alignment problems that have not been fully appreciated. Section 2 argues that Inhuman AMAs are likely to behave in inhumane ways that pose serious existential risks. Section 3 then contends that Better-Human AMAs run a serious risk of magnifying some sources of human moral error by reducing or eliminating others. Section 4 then argues that Human-Like AMAs would not only likely reproduce human moral failures, but also plausibly be highly intelligent, conscious beings with interests and wills of their own who should therefore be entitled to similar moral rights and freedoms as us. This generates what I call the New Control Problem: ensuring that humans and Human-Like AMAs exert a morally appropriate amount of control over each other. Finally, Section 5 argues that resolving the New Control Problem would, at a minimum, plausibly require ensuring what Hume and Rawls term ‘circumstances of justice’ between humans and Human-Like AMAs. But, I argue, there are grounds for thinking this will be profoundly difficult to achieve. I thus conclude on a skeptical note. Different approaches to developing ‘safe, ethical AI’ generate subtly different control and alignment problems that we do not currently know how to adequately resolve, and which may or may not be ultimately surmountable. (shrink)