Robyn Waller

In this paper we make the case for the emergence of novel kind of bias with the use of algorithmic decision-making systems. We argue that the distinctive generative process of feature creation, characteristic of machine learning (ML), contorts feature parameters in ways that can lead to emerging feature spaces that encode novel algorithmic bias involving already marginalized groups. We term this bias _assembled bias._ Moreover, assembled biases are distinct from the much-discussed algorithmic bias, both in source (training data versus feature creation) and in content (mimics of extant societal bias versus reconfigured categories). As such, this problem is distinct from issues arising from bias-encoding training feature sets or proxy features. Assembled bias is not epistemically transparent in source or content. Hence, when these ML models are used as a basis for decision-making in social contexts, algorithmic fairness concerns are compounded.

This chapter examines the control capacities of individuals with certain mental disorders and how, specifically, their reasons-responsiveness improves with treatment. Successful talk therapy, in particular, can bring individuals with disorders of agency closer to full-blown agency. The discussion focuses, first, on Agoraphobia and Exposure Therapy and, second, on Borderline Personality Disorder and Dialectical Behavior Therapy. We can see effective techniques of talk therapy, such as gradual exposure or radical acceptance exercises, as operating on the ability of patients to respond appropriately to reasons. In this way, reasons-responsiveness is a lens through which both to better appreciate the flourishing of patients, via improved control in both moral and nonmoral contexts. The chapter bridges philosophical concepts of reasons-responsiveness with key psychiatric constructs, such as distress tolerance and experiential avoidance.

Neuroscientists have located brain activity that prepares or encodes action plans before agents are aware of intending to act. On the basis of these findings and broader agency research, activity in these regions has been proposed as the neural realizers of practical intention. My aim in this paper is to evaluate the case for taking these neural states to be neural representations of intention. I draw on work in philosophy of action on the role and nature of practical intentions to construct a framework of the functional profile of intentions fit for empirical investigation. With this framework, I turn to the broader empirical neuroscience literature on agency to assess these proposed neural representations of intention. I argue that while these neural states in some respects satisfy the functions of intention in planning agency prospective of action, their fit with the role of intention in action execution is not well supported. I close by offering a sketch of which experimental task features could aid in the search for the neural realizer of intention in action.

Many philosophical theories of causation are egalitarian, rejecting a distinction between causes and mere causal conditions. We sought to determine the extent to which people's causal judgments discriminate, selecting as causes counternormal events—those that violate norms of some kind—while rejecting non-violators. We found significant selectivity of this sort. Moreover, priming that encouraged more egalitarian judgments had little effect on subjects. We also found that omissions are as likely as actions to be judged as causes, and that counternormative selectivity appears to apply equally to actions and omissions.

Informed consent has passed through three stages. The first paternalistic stage lasted for many centuries: The doctor's diagnosis and healing arts were kept secret, and informing patients was regarded as professionally and ethically wrong. Second came the legal stage, when the right of patients to make informed decisions concerning their own treatment was imposed by the courts and reluctantly tolerated by medical professionals. The third informed consent stage emerged more recently: the general therapy stage. The therapeutic benefits of informed consent have been well established, and informed consent is widely recognized as an important element in sound medical practice. When patients are effectively informed and can exert knowledgeable control over their own treatment decisions and therapy processes, that enhances recovery, strengthens the immune system, promotes better pain tolerance, prevents depression, and encourages patient cooperation and fortitude in treatment, rehabilitation, and preventative procedures. As the medical community has absorbed greater knowledge of this research, informed consent has been recognized as both ethically essential and therapeutically sound: the hallmark of the current general therapy stage of informed consent.

In this article the author makes the case for a hybrid sourcehood–leeway compatibilist account of free will. To do so, she draws upon Lehrer’s writing on free will, including his preference-based compatibilist account and Frankfurt-style cases from the perspective of the cognizant agent. The author explores what distinguishes kinds of intentional influence in manipulation cases and applies this distinction to a new perspectival variant of Frankfurt cases, those from the perspective of the counterfactual intervenor. She argues that it matters what kind of intentional influence is at issue in the counterfactual intervention and, further, that our judgments about desert of praise are affected by occupying the POV of the counterfactual intervenor. The author concludes that such attention to perspectival variants of Frankfurt cases supports the view that compatibilist sourcehood accounts of moral responsibility require an additional compatibilist could-have-done-otherwise condition to capture a more robust sense of moral responsibility.

In Free Will and Luck, Mele presents a case of an agent Ernie, whose zygote was intentionally designed so that Ernie A-s in 30 years, bringing about a certain event E. Mele uses this case of original design to outline the zygote argument against compatibilism. In this paper I criticize the zygote argument. Unlike other compatibilists who have responded to the zygote argument, I contend that it is open to the compatibilist to accept premise one, that Ernie does not act freely and is not morally responsible for anything he does. I argue that compatibilists should deny premise two. Diana’s effective intention to create Ernie’s zygote such that Ernie A-s in 30 years and her intervention to bring about his A-ing mark a significant difference between Ernie and normal agents in a deterministic universe with regard to how their zygotes were created that affects whether those agents act freely and are morally responsible for so acting.

What are the types of action at issue in the free will and moral responsibility debate? Are the neuroscientists who make claims about free will and moral responsibility studying those types of action? If not, can the existing paradigm in the field be modified to study those types of action? This paper outlines some claims made by neuroscientists about the inefficacy of conscious intentions and the implications of this inefficacy for the existence of free will. It argues that, typically, the types of actions at issue in the philosophical literature require proximal or distal conscious decisions and have the right kind of connection to reasons. It points out that neuroscientists are not studying this class of actions, as their studies focus on simple commanded actions (e.g., finger flex) and simple Buridan choices (e.g., push the left or right button). Finally, it argues that neuroscience already has the resources to study the types of action relevant for free will and moral responsibility and outlines two experiments which focus on skilled actions and moral choices that could be run using the available technology.

The aim of this paper is to challenge libertarian accounts of free will. It is argued that there is an irreconcilable tension between the way in which philosophers motivate the incompatibilist ability to do otherwise and the way in which they formally express it. Potential incompatibilist responses in the face of this tension are canvassed, and it is argued that each response is problematic. It is not claimed that incompatibilist accounts in general are incoherent, but rather that any incompatibilist account that requires that an agent have alternative possibilities at the point of a free action fails

In Responsibility and Control: A Theory of Moral Responsibility John Martin Fischer and Mark Ravizza propose an account of moral responsibility according to which an agent is morally responsible for an action just when that action is the product of her own moderately reasons-responsive mechanism, where reasons-responsiveness is explained in terms of the mechanism’s regular reasons-receptivity and weak reasons-reactivity. In a review of Fischer and Ravizza’s book Mele contends that their weakly reasons-reactivity condition is inadequate, constructing a case in which, according to their theory, an extreme agoraphobic is morally responsible for his staying in his home. In this paper I modify Fischer and Ravizza’s account of moral responsibility in light of Mele’s problematic example, suggesting a refinement of their weakly reasons-reactivity requirement via a distinction between weakly sufficient reasons and strongly sufficient reasons

The aim of this dissertation is to respond to a collection of conceptual and scientific threats to compatibilist accounts of free will, particularly reasons-responsive views. Compatibilists hold that free will is compatible with the truth of determinism. Some compatibilists also claim that some actual agent at least sometimes acts freely, where it is true that she acts freely in virtue of her satisfying a specific set of control and epistemic conditions. These conditions often include the possession of certain capacities, such as the capacity to recognize reasons for acting and to act on the basis of those reasons. First, I address two pressing philosophical objections to the adequacy of compatibilist accounts of free will: Are reasons-responsive accounts of free will too permissive in their class of free agents? Must reasons-responsive theorists - and compatibilists more generally - concede that some agents who are manipulated or intentionally designed by others are none the less acting freely and morally responsible for their actions? I then survey a widely proposed scientific threat to compatibilist accounts of free will - and to any account of free will that requires the same basic capacities as compatibilists do: Even if the compatibilist account accurately captures the basic capacities that an agent requires to act freely, does neuroscientific work on practical decision-making and action production limit - or worse, speak against -- our actual exercise of free will, as compatibilists conceive of it? I argue that the compatibilist need not yield to these philosophical challenges to the adequacy of their accounts of free will and to their claim that some actual agents act freely, sketching plausible compatibilist-friendly reasons to doubt the potency of each challenge. Finally, I argue that the extant incompatibilist accounts of free will suffer from a previously undiagnosed, serious problem: that there is an irreconcilable tension between the way in which philosophers motivate the incompatibilist ability to do otherwise and the way in which they formally express it. I conclude by offering a diagnosis for the future of the free will debate given the arguments advanced