Explainable artificial intelligence (XAI) seeks to produce explanations for those machine learning methods which are deemed opaque. However, there is considerable disagreement about what this means and how to achieve it. Authors disagree on what should be explained (topic), to whom something should be explained (stakeholder), how something should be explained (instrument), and why something should be explained (goal). In this paper, I employ insights from means-end epistemology to structure the field. According to means-end epistemology, different means ought to be (...) rationally adopted to achieve different epistemic ends. Applied to XAI, different topics, stakeholders, and goals thus require different instruments. I call this the means-end account of XAI. The means-end account has a descriptive and a normative component: on the one hand, I show how the specific means-end relations give rise to a taxonomy of existing contributions to the field of XAI; on the other hand, I argue that the suitability of XAI methods can be assessed by analyzing whether they are prescribed by a given topic, stakeholder, and goal. (shrink)

According to a widely held view epistemic justification is a normative notion. According to another widely held assumption, epistemic justification comes in degrees. Given that gradability requires a context-sensitivity that normativity seems to lack, these two assumptions stand in tension. Giving up the assumption of gradability of justification represents a lesser theoretical cost.

Scientists imagine for epistemic reasons, and these imaginings can be better or worse. But what does it mean for an imagining to be epistemically better or worse? There are at least three metaepistemological frameworks that present different answers to this question: epistemological consequentialism, deontic epistemology, and virtue epistemology. This paper presents empirical evidence that scientists adopt each of these different epistemic frameworks with respect to imagination, but argues that the way they do this is best explained if scientists are fundamentally (...) epistemic consequentialists about imagination. (shrink)

While it is commonly accepted that justified beliefs must be strongly supported by evidence and that support comes in degrees, the question of how much evidential support one needs in order to have a justified belief remains. In this paper, I consider how the question about degrees of evidential support connects with recent debates between consequentialist and deontological explanations of epistemic norms. I argue that explaining why strong, but not conclusive, evidential support is required for justification should be one explanandum (...) that such theories seek to explain. Furthermore, I argue that foundational theories that appeal to the promotion of epistemic value (especially consequentialism, but perhaps also some versions of epistemic deontology) are better suited to provide such an explanation. (shrink)

Sometimes, we face choices between actions most likely to lead to valuable outcomes, and actions which put us in a better position to learn. These choices exemplify what is called the exploration/exploitation trade-off. In computer science and psychology, this trade-off has fruitfully been applied to modulating the way agents or systems make choices over time. This article extends the trade-off to belief. We can be torn between two ways of believing, one of which is expected to be more accurate in (...) light of current evidence, whereas the other is expected to lead to more learning opportunity and accuracy in the long run. Further, it is sometimes rationally permissible to choose the latter. The article breaks down the features of action which give rise to the trade-off, and then argues that each feature applies equally well to belief. This conclusion is an instance of a systematic, foreseeable way in which what is rational to believe now depends on what one expects to be doing in the future. That is, epistemic rationality fundamentally concerns time. (shrink)

Epistemic consequentialism has been charged with ignoring the epistemic separateness of propositions and with (thereby) allowing trade-offs between propositions. Here, I do two things. First, I investigate the metaphor of the epistemic separateness of propositions. I argue that either (i) the metaphor is meaningfully unpacked in a way that is modeled on the moral separateness of persons, in which case it doesn’t support a ban on trade-offs or (ii) it isn’t meaningfully unpacked, in which case it really doesn’t support a (...) ban on trade-offs. Second, I consider the strategy of arguing against the trade-off permitting conception of epistemic rationality that is central to epistemic consequentialism on the basis of our intuitive verdicts about the permissibility of trade-offs in cases. I argue the execution of this strategy suffers a methodological mistake that, once corrected, vitiates the probative value of those intuitive verdicts. Hence the conclusion: the separateness of propositions provides no support for a ban on trade-offs, and an influential independent argument for such a ban is flawed. (shrink)

Some propositions are not likely to be true overall, but are likely to be true if you believe them. Appealing to the platitude that belief aims at truth, it has become increasingly popular to defend the view that such propositions are epistemically rational to believe. However, I argue that this view runs into trouble when we consider the connection between what’s epistemically rational to believe and what’s practically rational to do. I conclude by discussing how rejecting the view bears on (...) three other epistemological issues. First, we’re able to uncover a flaw in a common argument for permissivism. Second, we can generate a problem for prominent versions of epistemic consequentialism. Finally, we can better understand the connection between epistemic rationality and truth: epistemic rationality is a guide to true propositions rather than true beliefs. (shrink)

Selim Berker argues that epistemic consequentialism is pervasive in epistemology and that epistemic consequentialism is structurally flawed. is incorrect, however. I distinguish between epistemic consequentialism and epistemic instrumentalism and argue that most putative consequentialists should be considered instrumentalists. I also identify the structural problem of epistemic consequentialism Berker attempts to pinpoint and show that epistemic instrumentalism does not have the consequentialist problem.

Sometimes epistemologists theorize about belief, a tripartite attitude on which one can believe, withhold belief, or disbelieve a proposition. In other cases, epistemologists theorize about credence, a fine-grained attitude that represents one’s subjective probability or confidence level toward a proposition. How do these two attitudes relate to each other? This article explores the relationship between belief and credence in two categories: descriptive and normative. It then explains the broader significance of the belief-credence connection and concludes with general lessons from the (...) debate thus far. (shrink)

Elgin offers an influential and far-reaching challenge to veritism. She takes scientific understanding to be non-factive and maintains that there are epistemically useful falsehoods that figure ineliminably in scientific understanding and whose falsehood is no epistemic defect. Veritism, she argues, cannot account for these facts. This paper argues that while Elgin rightly draws attention to several features of epistemic practices frequently neglected by veritists, veritists have numerous plausible ways of responding to her arguments. In particular, it is not clear that (...) false propositional commitments figure ineliminably in understanding in the manner supposed by Elgin. Moreover, even if scientific understanding were non-factive and false propositional commitments did figure ineliminably in understanding, the veritist can account for this in several ways without thereby abandoning veritism. (shrink)

A theory of what we should believe should include a theory of what we should believe when we are uncertain about what we should believe and/or uncertain about the factors that determine what we should believe. In this paper, I present a novel theory of what we should believe that gives normative externalists a way of responding to a suite of objections having to do with various kinds of error, ignorance, and uncertainty. This theory is inspired by recent work in (...) ethical theory in which non-consequentialists 'consequentialize' their theories and then use the tools of decision-theory to give us an account of what we ought (in some sense) to do when we're uncertain about what we ought (in some primary sense) to do. On my proposal, because what we ought to do is acquire knowledge and avoid ignorance, we ought to believe iff the probability of coming to know is sufficiently high. This view has a number of important virtues. Among them, it gives us a unified story about how defeaters defeat (a theory developed with Julien Dutant), explains puzzling intuitions about the differences between lottery cases, preface cases, and cases of perceptual knowledge, and provides externalists (and internalists!) with a general framework for thinking about subjective normativity. It's also relatively brief. (shrink)

Philosophy and Phenomenological Research, EarlyView.

Despite the recent backlash against epistemic consequentialism, an explicit systematic alternative has yet to emerge. This paper articulates and defends a novel alternative, Epistemic Kantianism, which rests on a requirement of respect for the truth. §1 tackles some preliminaries concerning the proper formulation of the epistemic consequentialism / non-consequentialism divide, explains where Epistemic Kantianism falls in the dialectical landscape, and shows how it can capture what seems attractive about epistemic consequentialism while yielding predictions that are harder for the latter to (...) secure in a principled way. §2 presents Epistemic Kantianism. §3 argues that it is uniquely poised to satisfy the desiderata set out in §1 on an ideal theory of epistemic justification. §4 gives three further arguments, suggesting that it (i) best explains the objective normative significance of the subject's perspective in epistemology, (ii) follows from the kind of axiology needed to solve the swamping problem together with modest assumptions about the relation between the evaluative and the deontic, and (iii) illuminates certain asymmetries in epistemic value and obligation. §5 takes stock and reassesses the score in the debate. (shrink)

It is typically thought that some epistemic states are valuable—knowing, truly or accurately believing, understanding. These are states it’s thought good to be in and it’s also said that we aim or want to be in them. It is then sometimes claimed that these sorts of thoughts about epistemic goods or values ground or explain our epistemic norms. For instance, we think subjects should follow their evidence when they form their beliefs. But why should they? Why not believe against the (...) evidence or ignore it completely in deciding what to believe? Here’s a compelling sort of answer: because epistemic subjects are or ought to be trying to know more and following their evidence is a means to that end or to fulfilling that obligation. In this paper I argue that this compelling thought cannot be right. Subjects who are trying to know more will regularly fail to conform to some of our most familiar epistemic norms. (shrink)

Could it be right to convict and punish defendants using only statistical evidence? In this paper, I argue that it is not and explain why it would be wrong. This is difficult to do because there is a powerful argument for thinking that we should convict and punish defendants using statistical evidence. It looks as if the relevant cases are cases of decision under risk and it seems we know what we should do in such cases (i.e., maximize expected value). (...) Given some standard assumptions about the values at stake, the case for convicting and punishing using statistical evidence seems solid. In trying to show where this argument goes wrong, I shall argue (against Lockeans, reliabilists, and others) that beliefs supported only by statistical evidence are epistemically defective and (against Enoch, Fisher, and Spectre) that these epistemic considerations should matter to the law. To solve the puzzle about the role of statistical evidence in the law, we need to revise some commonly held assumptions about epistemic value and defend the relevance of epistemology to this practical question. (shrink)

Vice epistemology is the philosophical study of the nature, identity, and epistemological significance of intellectual vices. Such vices include gullibility, dogmatism, prejudice, closed-mindedness, and negligence. These are intellectual character vices, that is, intellectual vices that are also character traits. I ask how the notion of an intellectual character vice should be understood, whether such vices exist, and how they might be epistemologically significant. The proposal is that intellectual character vices are intellectual character traits that impede effective and responsible inquiry. I (...) argue that situationist critiques of virtue epistemology pose no significant threat to this proposal. Studies by social psychologists of belief in conspiracy theories suggest that it is sometimes appropriate to explain questionable beliefs by reference to intellectual character vices. Neither ‘regulative’ nor ‘analytic’ epistemology has any good reason to question the epistemological significance of such vices. (shrink)

Addressing the ‘virtue conflation’ problem requires the preservation of intuitive distinctions between virtue types, that is, between intellectual and moral virtues. According to one influential attempt to avoid this problem proposed by Julia Driver, moral virtues produce benefits to others—in particular, they promote the well-being of others—while the intellectual virtues, as such, produce epistemic good for the agent. We show that Driver's demarcation of intellectual virtue, by adverting to the self-/other distinction, leads to a reductio, and ultimately, that the prospects (...) for resolving the virtue conflation problem look dim within an epistemic consequentialist approach to the epistemic right and the epistemic good. (shrink)

The paper discusses some challenges to veritism, the view that the fundamental epistemic good is knowledge. It looks like the best way to meet these challenges might be to appeal to some of Sosa's ideas about the value of achievements, but I argue that the performance normativity framework only gives us part of what we want. What we need is a more radical break with the veritist approach. We need to embrace gnosticism, the view that knowledge is the fundamental epistemic (...) good. (shrink)

Direct epistemic consequentialism is the idea that X is epistemically permissible iff X maximizes epistemic value. It has received lots of attention in recent years and is widely accepted by philosophers to have counterintuitive implications. There are various reasons one might suspect that the relevant intuitions will not be widely shared among non-philosophers. This paper presents an initial empirical study of ordinary intuitions. The results of two experiments demonstrate that the counterintuitiveness of epistemic consequentialism is more than a philosophers' worry---the (...) folk seem to agree! (shrink)

This is a critical discussion of the accuracy-first approach to epistemic norms. If you think of accuracy (gradational or categorical) as the fundamental epistemic good and think of epistemic goods as things that call for promotion, you might think that we should use broadly consequentialist reasoning to determine which norms govern partial and full belief. After presenting consequentialist arguments for probabilism and the normative Lockean view, I shall argue that the consequentialist framework isn't nearly as promising as it might first (...) appear. (shrink)

In this essay, I aim to motivate and defend a pluralistic view of epistemic rationality. At the core of the view is the notion that epistemic rationality is essentially a species of practical rationality put in the service of various epistemic goals. I begin by sketching some closely related views that have appeared in the literature. I then present my preferred version of the view and sketch some of its benefits. Thomas Kelly has raised challenging objections to a part of (...) the view, which I then present and address. I then raise some distinct worries that might linger and sketch some possible ways one might address them. (shrink)

There has been considerable discussion recently of consequentialist justifications of epistemic norms. In this paper, I shall argue that these justifications are not justifications. The consequentialist needs a value theory, a theory of the epistemic good. The standard theory treats accuracy as the fundamental epistemic good and assumes that it is a good that calls for promotion. Both claims are mistaken. The fundamental epistemic good involves accuracy, but it involves more than just that. The fundamental epistemic good is knowledge, not (...) mere true belief, because the goodness of an epistemic state is connected to that state's ability to give us reasons. If I'm right about the value theory, this has a number of significant implications for the consequentialist project. First, the good-making features that attach to valuable full beliefs are not features of partial belief. The resulting value theory does not give us the values we need to give consequentialist justifications of credal norms. Second, the relevant kind of good does not call for promotion. It is good to know, but the rational standing of a belief is not determined by the belief's location in a ranked set of options. In the paper's final section, I explain why the present view is a kind of teleological non-consequentialism. There is a kind of good that is prior to the right, but as the relevant kind of good does not call for promotion the value theory shows us what is wrong with the consequentialist project. (shrink)

Critics have recently argued that reliabilists face trade-off problems, forcing them to condone intuitively unjustified beliefs when they generate lots of true belief further downstream. What these critics overlook is that reliabilism entails that there areside-constraintson belief-formation, on account of which there are some things you should not believe, even if doing so would have very good epistemic consequences. However, we argue that by embracing side-constraints the reliabilist faces a dilemma: she can either hold on to reliabilism, and with it (...) aforementioned side-constraints, but then needs to explain why we should allow the pursuit of justification to get in the way of the acquisition of true belief; or she can deny that there are side-constraints – and in effect give up on reliabilism. We'll suggest that anyone moved by the considerations that likely attract people to reliabilism in the first place – the idea the true belief is good, and as such should be promoted – should go for the second horn, and instead pursue a form ofepistemic utilitarianism. (shrink)

Recent empirical work on non-philosophers’ intuitions about epistemic normativity reveals patterns that cannot be fully accounted for by direct epistemic consequentialism. On the basis of these results, one might picture participants as “epistemic deontologists.” We present the results of two new experiments that support a more nuanced picture. We examine intuitions about guesses and hypotheses, and about beliefs. Our results suggest a two-factor model of intuitions, wherein both consequentialist and non-consequentialist considerations affect participants’ judgments about epistemic permissibility.

A discussion of L.A. Paul's 'Transformative Experience' from an Author Meets Critics session at the 2015 Pacific APA.

To what extent can we rely on others for information without such reliance becoming epistemically problematic? In this paper, this question is addressed in terms of a specific form of reliance: cognitive outsourcing. Cognitive outsourcing involves handing over (outsourcing) one’s information collection and processing (the cognitive) to others. The specific question that will be asked about such outsourcing is if there is an epistemic problem about cognitive outsourcing as such. To ask if there is an epistemic problem with x for (...) S is to ask if x is a problem for S’s ability to acquire true belief and avoid false belief. To ask if there is a problem for S with x as such is to ask if it is impossible to solve the problem for S while leaving x as is. I argue that, if we consider the five most plausible candidate epistemic problems raised by cognitive outsourcing—i.e., unreliability, gullibility, irrationality, dependency, and lack of epistemic autonomy—we see for each candidate that it is either not an epistemic problem, or not a problem about cognitive outsourcing as such. (shrink)

I present an argument against the thesis of Uniqueness and in favour of Permissivism. Counterexamples to Uniqueness are provided, based on ‘Safespot’ propositions – i.e. a proposition that is guaranteed to be true provided the subject adopts a certain attitude towards it. The argument relies on a plausible principle: (roughly stated) If S knows that her believing p would be a true belief, then it is rationally permitted for S to believe p. One motivation for denying this principle – viz. (...) opposition to ‘epistemic consequentialism’ – is briefly discussed. The principle is extended to cover degrees of belief and compared with a couple of other well-known constraints on rational degrees of belief. (shrink)

On an evidentialist position, it is epistemically rational for us to believe propositions that are (stably) supported by our total evidence. We are epistemically permitted to believe such propositions, and perhaps even ought to do so. Epistemic rationality is normative. One popular way to explain the normativity appeals to epistemic teleology. The primary aim of this paper is to argue that appeals to epistemic teleology do not support that we ought to believe what is rational to believe, only that we (...) are permitted to do so. In arguing for that, I defend an epistemic teleological position that is radical in nature. It involves no commitment to aiming at the truth. I conclude by dispelling some worries that have been raised about my position. (shrink)

Reliabilism—the view that a belief is justified iff it is produced by a reliable process—is often characterized as a form of consequentialism. Recently, critics of reliabilism have suggested that since it is a form of consequentialism, reliabilism condones a variety of problematic trade-offs involving cases where someone forms an epistemically deficient belief now that will lead her to more epistemic value later. In the present paper, we argue that the relevant argument against reliabilism fails because it equivocates. While there is (...) a sense in which reliabilism is a kind of consequentialism, it is not of a kind on which we should expect problematic trade-offs. (shrink)

Hay dos supuestos sobre el valor epistémico que guían las discusiones más recientes sobre éste. El primero es que hay algo bueno con respecto a la creencia verdadera. El segundo supuesto es que es posible que dos creencias difieran en su valor incluso si ambas creencias son igualmente correctas. El veritista tiene fácil explicar el primer supuesto, pero tiene más difícil explicar el segundo. Para explicarlo, el veritista tiene que mostrar que las creencias verdaderas pueden diferir en su valor porque (...) encarnan diferentes valores de bienes derivativos. El veritista tiene que proporcionarnos una historia de cómo es posible que pares de creencias verdaderas difieran en su valor epistémico no-instrumental. La forma más prometedora que una explicación de este tipo puede tomar, en mi opinión, apela al trabajo de Sosa (2007) sobre el valor epistémico y la naturaleza del conocimiento. Según la postura de Sosa, el conocimiento es un tipo de logro, un tipo de éxito que es atribuible al sujeto y a sus habilidades. Estos logros, nos sugiere, son más valiosos que un caso de éxito atribuible a la suerte y que no constituye un logro. Presentaré algunas dificultades potenciales para su postura y consideraré si tiene los recursos para lidiar con ellas. Voy a delinear los detalles de una postura alternativa que evade estas dificultades. Tal postura retiene el análisis del conocimiento de Sosa, pero requiere que revisemos algunas propuestas veritistas sobre el valor epistémico. La postura revisada nos proporciona algunas herramientas que serán útiles al tratar de ofrecer una teoría de la creencia racional o diestra. (shrink)

In this article we discuss what we call the deliberative division of epistemic labor. We present evidence that the human tendency to engage in motivated reasoning in defense of our beliefs can facilitate the occurrence of divisions of epistemic labor in deliberations among people who disagree. We further present evidence that these divisions of epistemic labor tend to promote beliefs that are better supported by the evidence. We show that promotion of these epistemic benefits stands in tension with what extant (...) theories in epistemology take rationality to require in cases of disagreement. We argue that the epistemic benefits that result from the deliberative division of epistemic labor can provide epistemic reason to maintain confidence in cases of disagreement. We then show that the deliberative division of epistemic labor constitutes a distinct kind of epistemic dependence. (shrink)

Epistemic consequentialists think that epistemic norms are about believing the truth and avoiding error. Recently, a number of authors have rejected epistemic consequentialism on the basis that it incorrectly sanctions tradeoffs of epistemic goodness. Here, I argue that epistemic consequentialists should borrow two lessons from ethical consequentialists to respond to these worries. Epistemic consequentialists should construe their view as an account of right belief, which they distinguish from other notions like rational and justified belief. Epistemic consequentialists should also make their (...) view ‘sophisticated,’ in the sense of Railton. Epistemic consequentialism, I conclude, is best construed as sharing much of its structure with prominent act-consequentialist views in ethics. Epistemic consequentialism has an advantage over its ethical counterpart though: the key claim of the view is practically universally accepted, which gives us an additional reason to think it’s true. (shrink)

Direct Epistemic Consequentialism faces the Truth Fairy. Indirect Epistemic Consequentialism promises to avoid this issue. But there are worse fairies than the Truth Fairy. There is the Worse Fairy. The case of the Worse Fairy helps demonstrate that epistemic consequentialists who would solve problems like the Truth Fairy by ‘going indirect’ face a dilemma.