This paper explores a neglected normative dimension of algorithmic opacity in the workplace and the labor market. It argues that explanations of algorithms and algorithmic decisions are of noninstrumental value. That is because explanations of the structure and function of parts of the social world form the basis for reflective clarification of our practical orientation toward the institutions that play a central role in our life. Using this account of the noninstrumental value of explanations, the paper diagnoses distinctive normative defects in the workplace and economic institutions which a reliance on AI can encourage, and which lead to alienation.

Skepticism about blameworthiness says that there is good reason to doubt that, in our world, humans are ever blameworthy for their deeds. A significant problem for the discussion of this view is that it is unclear how to understand the kind of blame that should be at issue. This paper makes a new proposal. The basic idea is that the kind of blame skeptics should be skeptical about is constituted by responses that can violate the targets’ claims and by the responders’ thought that the targets have forfeited this claim because of their morally objectionable actions and because of how they were when they performed them. This view identifies an important part of our everyday lives and frames discussions about skepticism about blameworthiness in a new way.

Separatists about grounding take explanations to be separate from their corresponding grounding-facts. Grounding-facts are supposed to underlie, or back, such explanations. However, the backing relation hasn’t received much attention in the literature. The aim of this paper is to provide an informative definition of backing. First, I examine two prominent proposals: backing as explaining (Kovacs 2017; 2019a) and backing as grounding (see Sjölin Wirling 2020). Finally, I put forward my own proposal. I argue that under plausible assumptions about the role of backing and the nature of explanation, backing should be understood as a form of truthmaking, minimally construed.

Algorithmic communications pose several challenges to democracy. The three phenomena of filtering, hypernudging, and microtargeting can have the effect of polarizing an electorate and thus undermine the deliberative potential of a democratic society. Algorithms can spread fake news throughout the society, undermining the epistemic potential that broad participation in democracy is meant to offer. They can pose a threat to political equality in that some people may have the means to make use of algorithmic communications and the sophistication to be immune from attempts at manipulation, while other people are vulnerable to manipulation by those who use these means. My concern here is with the danger that algorithmic communications can pose to political equality, which arises because most citizens must make decisions about what and who to support in democratic politics with only a sparse budget of time, money, and energy. Algorithmic communications such as hypernudging and microtargeting can be a threat to democratic participation when persons are operating in environments that do not conduce to political sophistication. This constitutes a deepening of political inequality. The political sophistication necessary to counter this vulnerability is rooted for many in economic life and it can and ought to be enhanced by changing the terms of economic life.

Various theorists have endorsed the “communication argument”: communicative capacities are necessary for morally responsible agency because blame aims at a distinctive kind of moral communication. I contend that existing versions of the argument, including those defended by Gary Watson and Coleen Macnamara, face a pluralist challenge: they do not seem to sit well with the plausible view that blame has multiple aims. I then examine three possible rejoinders to the challenge, suggesting that a context-specific, function-based approach constitutes the most promising modification of the communication argument.

Machine learning algorithms are increasingly used to shape high-stake allocations, sparking research efforts to orient algorithm design towards ideals of justice and fairness. In this research on algorithmic fairness, normative theorizing has primarily focused on identification of “ideally fair” target states. In this paper, we argue that this preoccupation with target states in abstraction from the situated dynamics of deployment is misguided. We propose a framework that takes dynamic trajectories as direct objects of moral appraisal, highlighting three respects in which such trajectories can be subject to evaluation in relation to their (i) temporal dynamics, (ii) robustness, and (iii) representation.

Law-enforcement agencies are increasingly able to leverage crime statistics to make risk predictions for particular individuals, employing a form of inference that some condemn as violating the right to be “treated as an individual.” I suggest that the right encodes agents’ entitlement to a fair distribution of the burdens and benefits of the rule of law. Rather than precluding statistical prediction, it requires that citizens be able to anticipate which variables will be used as predictors and act intentionally to avoid them. Furthermore, it condemns reliance on various indexes of distributive injustice, or unchosen properties, as evidence of law-breaking.

Is an outcome where many people are saved and one person dies better than an outcome where the one is saved and the many die? According to the standard utilitarian justification, the former is better because it has a greater sum total of well-being. This justification involves a controversial form of moral aggregation, because it is based on a comparison between aggregates of different people’s well-being. Still, an alternative justification—the Argument for Best Outcomes—does not involve moral aggregation. I extend the Argument for Best Outcomes to show that any utilitarian evaluation can be justified without moral aggregation.

Automated Influence is the use of Artificial Intelligence (AI) to collect, integrate, and analyse people’s data in order to deliver targeted interventions that shape their behaviour. We consider three central objections against Automated Influence, focusing on privacy, exploitation, and manipulation, showing in each case how a structural version of that objection has more purchase than its interactional counterpart. By rejecting the interactional focus of “AI Ethics” in favour of a more structural, political philosophy of AI, we show that the real problem with Automated Influence is the crisis of legitimacy that it precipitates.

A number of philosophers have held that we visually experience objects’ occluded parts, such as the out-of-view exterior of a voluminous, opaque object. That idea is supposed to be what best explains the fact that we see objects as whole or complete despite having only a part of them in view at any given moment. Yet, the claim doesn’t express a phenomenological datum and the reasons for thinking we do experience objects’ occluded parts, I argue, aren’t compelling. Additionally, I anticipate and reply to attempts to salvage the idea by appeal to perceptual expectation and amodal completion. Lastly, I address potential concerns that the only way to capture the phenomenal character of perceiving voluminous objects is to say experience outstrips what’s in view, providing a description of such experience without any implication of that idea.

This paper warns of two threats to moral responsibility that arise when accounting for omissions, given some plausible assumptions about how abilities are related to responsibility. The first problem threatens the legitimacy of our being responsible by expanding the preexisting tension that luck famously raises for moral responsibility. The second threat to moral responsibility challenges the legitimacy of our practices of holding responsible. Holding others responsible for their omissions requires us to bridge an epistemic gap that does not arise when holding others responsible for their actions—one that we might often fail to cross.

Rudolf Carnap’s principle of tolerance states that there is no need to justify the adoption of a logic by philosophical means. Carnap uses the freedom provided by this principle in his philosophy of mathematics: he wants to capture the idea that mathematical truth is a matter of linguistic rules by relying on a strong metalanguage with infinitary inference rules. In this paper, I give a new interpretation of an argument by E. W. Beth, which shows that the principle of tolerance does not suffice to remove all obstacles to the employment of infinitary rules.

It is becoming more common that the decision-makers in private and public institutions are predictive algorithmic systems, not humans. This article argues that relying on algorithmic systems is procedurally unjust in contexts involving background conditions of structural injustice. Under such nonideal conditions, algorithmic systems, if left to their own devices, cannot meet a necessary condition of procedural justice, because they fail to provide a sufficiently nuanced model of which cases count as relevantly similar. Resolving this problem requires deliberative capacities uniquely available to human agents. After exploring the limitations of existing formal algorithmic fairness strategies, the article argues that procedural justice requires that human agents relying wholly or in part on algorithmic systems proceed with caution: by avoiding doxastic negligence about algorithmic outputs, by exercising deliberative capacities when making similarity judgments, and by suspending belief and gathering additional information in light of higher-order uncertainty.

More than a decade of exacting scientific research involving paleontological fragments and ancient DNA has lately produced a series of pronouncements about a purportedly novel population of archaic hominins dubbed “the Denisova.” The science involved in these matters is both technically stunning and, socially, at times a bit reckless. Here I discuss the responsibilities which scientists incur when they make inductively risky pronouncements about the different relative contributions by Denisovans to genomes of members of apparent subpopulations of current humans (i.e., the so-called “races”). This science is sensational: it is science which empirically speculates, to the public delight’s and entertainment, about scintillating topics such as when humans evolved, where we came from, and who else we were having sex with during our early hominin history. An initial characterization of sensational science emerges from my discussion of the case, as well as a diagnosis of an interactive phenomenon termed amplified inductive risk.

Humeans identify the laws of nature with universal generalizations that systematize rather than govern the particular matters of fact. Humeanism is frequently accused of circularity: laws explain their instances, but Humean laws are, in turn, grounded by those instances. Unfortunately, this argument trades on controversial assumptions about grounding and explanation that Humeans routinely reject. However, recently an ostensibly semantic circularity objection has been offered, which seeks to avoid reading such assumptions into the Humean view. This paper argues that the new semantic version tacitly relies on the familiar metaphysical one and, therefore, it ultimately brings nothing new to the table.

We can find in the passages that set out the Master Argument a precursor to the paradox of knowability. That paradox shows that if all truths are knowable, all truths are known. Similarly, Berkeley might be read as proposing that if all sensible objects are (distinctly) conceivable, then all sensible objects are conceived.

This article discusses why it is the case that we refuse to accept strange evaluative claims as being true in fictions, even though we are happy to go along with other types of absurdities in such contexts. For instance, we would refuse to accept the following statement as true, even in the context of a fiction: (i) In killing her baby, Giselda did the right thing; after all, it was a girl.

This article offers a sensibilist diagnosis of this puzzle, inspired by an observation first made by David Hume. According to sensibilism, the way we feel about things settles their evaluative properties. Thus, when confronted with a fictional scenario where the configuration of non-evaluative facts and properties is relevantly similar to the actual world, we refuse to go along with evaluative properties being instantiated according to a different pattern. It is the attitudes we hold in the actual world that fix the extension of evaluative terms, even in nonactual worlds. When engaging with a fiction, we (to some extent) leave our beliefs about what the world is like behind, while taking our emotional attitudes with us into the fiction.

To substantiate this diagnosis, this paper outlines a sensibilist semantics for evaluative terms based on recent discussion regarding predicates of personal taste, and explains how, together with standard assumptions about the nature of fictional discourse, it makes the relevant predictions with respect to engagement with fictions.

Organizations have neither a right to vote nor a right to life. But we can owe them to keep our promises or show them gratitude. So we owe some things to organizations, but not everything we owe to people. What explains this? Individualistic views explain it just in terms of features of organizations’ individual members. Collectivistic views explain it just in terms of features of those organizations. Neither view works. Instead, we need to synthesize these approaches. Some individual interests are distinctively collective. Individuals have an interest in participating in successful collective action. This explains organizations’ apparently fragmented moral status.

The paper articulates a novel strategy against external world skepticism. It shows that a modal assumption of the skeptical argument cannot be justified.

A version of the principle of alternate possibilities (PAP) claims that one is only blameworthy for actions which one was able to avoid. Much of the discussion about PAP concerns Frankfurt’s counterexamples to it. After fifty years of refined debates, progress might seem hopeless. Yet, we can make headway by asking: “what’s our reason for believing PAP?” The best answer is this: lacking eligible alternatives—alternatives whose cost is not too high to reasonably opt for—is a good excuse. Yet, this principle is subject to straightforward counterexamples, unless it is given an epistemicised reading. And in this latter case, it does not support PAP. So, PAP is unsupported, at least for now.