I advance a novel argument for an infallibilist theory of knowledge, according to which we know all and only those propositions that are certain for us. I argue that this theory lets us reconcile major extant theories of knowledge, in the following sense: for any of these theories, if we require that its central condition (evidential support, reliability, safety, etc.) obtains to a maximal degree, we get a theory of knowledge extensionally equivalent to infallibilism. As such, the infallibilist can affirm (...) that, when their conditions are suitably interpreted, most post-Gettier theories of knowledge offer necessary and sufficient conditions for knowledge. The infallibilist can thus reconcile major theories of knowledge, and is in a better position to explain the intuitive appeal of these theories than the fallibilist who only accepts one of them, and rejects the rest. (shrink)

In “How to Be an Infallibilist,” Julien Dutant (2016, 149) presents a simple and seemingly plausible argument that knowledge requires infallible belief—roughly, belief that could not be mistaken. As Dutant recognizes, infallibilism is almost universally dismissed, in large part because it seems to rule out any knowledge of the physical world. He seeks to show how we can be an Infallibilist without being a skeptic, based on the assumption that knowledge has a safety condition. I critically examine each line of (...) Dutant’s argument, showing that the argument is unsound on any plausible interpretation. I also question the idea that knowledge cannot be the conjunction of true belief and a nonfactive condition, that any belief about the physical world could not be false, and that any nonskeptical alternative to infallibilism would have to allow knowledge of chancy outcomes. I briefly suggest that a fallibilist account can be strict enough to satisfy the Infallibilist’s quest for certainty. (shrink)

People can be treated unjustly with respect to the level of credibility others accord to their testimony. This is the core idea of the philosophical idea of epistemic justice. It should be of utmost interest to criminal law which extensively deals with normative issues of evidence and testimony. It may reconstruct some of the long-standing criticisms of criminal law regarding credibility assessments and the treatment of witnesses, especially in sexual assault cases. However, philosophical discussions often overlook the intricate complexities of (...) real procedural law and its underlying considerations. In its present form, the philosophical notion of epistemic injustice provides limited insights into legal discourse; it necessitates translation and adaptation. This study contributes to this endeavor by examining the contentious issue of testimony from witnesses who have undergone trauma-focused psychotherapy. Since the 1980s, courts worldwide were troubled with cases of false accusations based on false memories generated by suggestive therapeutic interventions. As a result, such post-therapy testimonies are discounted in one way or another in many jurisdictions. However, courts are still confronted with such testimonies, and the modi vivendi legal systems have established to deal with them continue to give rise to concerns about unjust treatment of witnesses. The question is thus whether legal rules or established practices of evaluating testimony based on memories which resurfaced after psychotherapy are epistemically and legally just. The paper presents seven ways in which courts may assess such testimonies and examines them in light of epistemic and procedural justice. Some of them prima facie constitute a form of epistemic injustice because they discount testimonies to an unwarranted degree. But these injustices might be justified by overriding principles favoring defendants. Nonetheless, the idea of epistemic justice, more broadly understood, inspires two principles that may serve as a foundation for a future conception of epistemic justice adapted to the law. (shrink)

Certain characteristics make machine learning a powerful tool for processing large amounts of data, and also particularly unsuitable for explanatory purposes. There are worries that its increasing use in science may sideline the explanatory goals of research. We analyze the key characteristics of ML that might have implications for the future directions in scientific research: epistemic opacity and the ‘theory-agnostic’ modeling. These characteristics are further analyzed in a comparison of ML with the traditional statistical methods, in order to demonstrate what (...) it is specifically that makes ML methodology substantially unsuitable for reaching explanations. The analysis is given broader philosophical context by connecting it with the views on the role of prediction and explanation in science, their relationship, and the value of explanation. We proceed to show, first, that ML disrupts the relationship between prediction and explanation commonly understood as a functional relationship. Then we show that the value of explanation is not exhausted in purely predictive functions, but rather has a ubiquitously recognized value for both science and everyday life. We then invoke two hypothetical scenarios with different degrees of automatization of science, which help test our intuitions on the role of explanation in science. The main question we address is whether ML will reorient or otherwise impact our standard explanatory practice. We conclude with a prognosis that ML would diversify science into purely predictively oriented research based on ML-like techniques and, on the other hand, remaining faithful to anthropocentric research focused on the search for explanation. (shrink)

I will argue that internalism about justification entails the apparently absurd conclusion that it is possible to know specific facts about the external world—for example, that there is a tree in the quad—on the basis of introspection and a priori reflection. After a brief characterization of internalism (§1), I will set out the problem (§2). I will then discuss three replies: one that denies the form of doxastic voluntarism involved in the problem (§3), one that denies that knowledge of higher-order (...) facts about justification can justify corresponding first-order beliefs (§4), and, finally, one that involves biting the bullet (§5). I will argue that each reply fails. (shrink)

When we cook, by meticulously following a recipe, or adding a personal twist to it, we sometimes care not only to (re-)produce a taste that we can enjoy, but also to give our food a certain aftertaste. This is not surprising, given that we ordinarily take aftertaste to be an important part of the gustatory experience as a whole, one which we seek out, and through which we evaluate what we eat and drink—at least in many cases. What is surprising (...) is that aftertastes, from a psychological point of view, seem to be analogous to afterimages, and thus have little or no epistemic import. In this paper we tackle this puzzle, and argue that we are right in treating aftertastes seriously. The moral is that both from a metaphysical and an epistemic point of view aftertastes should be categorized differently from afterimages. (shrink)

The problem of perception is the problem of explaining how perceptual knowledge is possible. The skeptic has a simple solution: it is not possible. I analyze the weaknesses of one type of skeptical reasoning by making explicit a dynamic epistemic principle from dynamic epistemic logic that is implicitly used in debating the problem, with the aim of offering a novel diagnosis to this skeptical argument. I argue that prominent modest foundationalist responses to perceptual skepticism can be understood as rejecting the (...) dynamic assumption made by the skeptic, that there are independent reasons to doubt the truth of such a principle in the context of skeptical reasoning, and that making the dynamic principle explicit allows for a better understanding of at least one objection to modest foundationalism. (shrink)

In this chapter I introduce the thesis that perceptual appearances are cognitively penetrable and analyse cases made against phenomenal conservatism hinging on this thesis. In particular, I focus on objections coming from the externalist reliabilist camp and the internalist inferentialist camp. I conclude that cognitive penetrability doesn’t yield lethal or substantive difficulties for phenomenal conservatism.

Reliabilism is an intuitive and attractive view about epistemic justification. However, it has many well-known problems. I offer a novel condition on reliabilist theories of justification. This method coherence condition requires that a method be appropriately tested by appeal to a subject’s other belief-forming methods. Adding this condition to reliabilism provides a solution to epistemic circularity worries, including the bootstrapping problem.

Whether abduction is treated as an argument or as an inference, the mainstream view presupposes a tight connection between abduction and inference to the best explanation. This paper critically evaluates this link and supports a narrower view on abduction. Our main thesis is that merely the hypothesis-generative aspect, but not the evaluative aspect, is properly abductive in the sense introduced by C. S. Peirce. We show why equating abduction with IBE unnecessarily complicates argument evaluation by levelling the status of abduction (...) as a third reasoning mode. We also propose a scheme for abductive argument along with critical questions, and suggest retaining abduction alongside IBE as related but distinct categories. (shrink)

Censorship involves, inter alia, adopting a certain type of epistemic policy. While much has been written on the harms and benefits of free expression, and the associated rights thereof, the epistemic preconditions of justified censorship are relatively underexplored. In this paper, I argue that examining intrapersonal norms of how we ought to treat evidence that might come to us over time can shed light on interpersonal norms of evidence generation and sharing that are relevant in the context of censorship. The (...) upshot is that justified censorship requires the censor to meet a very high epistemic burden regarding the target proposition(s)—importantly, one that exceeds knowledge. (shrink)

This essay presents a unified account of safety, sensitivity, and severe testing. S’s belief is safe iff, roughly, S could not easily have falsely believed p, and S’s belief is sensitive iff were p false S would not believe p. These two conditions are typically viewed as rivals but, we argue, they instead play symbiotic roles. Safety and sensitivity are both valuable epistemic conditions, and the relevant alternatives framework provides the scaffolding for their mutually supportive roles. The relevant alternatives condition (...) holds that a belief is warranted only if the evidence rules out relevant error possibilities. The safety condition helps categorise relevant from irrelevant possibilities. The sensitivity condition captures ‘ruling out’. Safety, sensitivity, and the relevant alternatives condition are typically presented as conditions on warranted belief or knowledge. But these properties, once generalised, help characterise other epistemic phenomena, including warranted inference, legal verdicts, scientific claims, reaching conclusions, addressing questions, warranted assertion, and the epistemic force of corroborating evidence. We introduce and explain Mayo’s severe testing account of statistical inference. A hypothesis is severely tested to the extent it passes tests that probably would have found errors, were they present. We argue Mayo’s account is fruitfully understood using the resulting relevant alternatives framework. Recasting Mayo’s condition using the conceptual framework of contemporary epistemology helps forge fruitful connections between two research areas—philosophy of statistics and the analysis of knowledge—not currently in sufficient dialogue. The resulting union benefits both research areas. (shrink)

The generality problem is one of the most pressing challenges for reliabilism. The problem begins with this question: of all the process types exemplified by a given process token, which types are the relevant ones for determining whether the resultant belief counts as knowledge? As philosophers like Earl Conee and Richard Feldman have argued, extant responses to the generality problem have failed, and it looks as if no solution is forthcoming. In this paper, I present a new response to the (...) generality problem that illuminates the nature of knowledge-enabling reliability. My response builds upon the insights of Juan Comesaña’s well-founded solution to the generality problem, according to which relevant types are content–evidence pairs, i.e., descriptions of both the target belief’s content and the evidence on which the belief was based. While most responses to the generality problem, including Comesaña’s, only posit one relevant type for any given process token, I argue that knowledge-enabling reliability requires a process token to be reliable with respect to multiple content–evidence pairs, each with varying degrees of descriptive specificity. I call this solution multi-type evidential reliabilism. After offering a clear formulation of MTE, I conclude by arguing that MTE is sufficiently informative to rebut Conee and Feldman’s generality problem objection to a reliability condition on knowledge. (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.

The paper argues that (i) the notion of epistemic reliability, as it is standardly defined within mainstream epistemology, is a multidimensional concept, and that (ii) paying attention to reliability’s multidimensional nature can significantly expand reliabilism’s purview both in the theoretical and practical domain. Reliabilist theories of knowledge and justification agree that a process is reliable just in case it leads to a ‘sufficiently high preponderance of true over false beliefs.’ Given this straightforward definition, reliability appears to be, and so far (...) has been implicitly treated by epistemologists as, a mono-dimensional concept. Nevertheless, the terminology and formulas that statisticians and computers scientists have developed for assessing the reliability of binary classification processes reveal that for any belief-forming process whose aim is to identify the presence or the absence of some quality ι, there are at least four different interpretations of the above definition. These interpretations correspond to four separate dimensions of reliability: Precision, Negative Predictive Value (NPV), Recall and Specificity. Though epistemologists do not explicitly specify how they interpret the above definition of reliability, due to their main interest in ‘know-that’, their mono-dimensional focus seems to have so far been mainly directed at Precision alone (or, at best, to both Precision and NPV—but without distinguishing between the two). As it transpires, however, the (entirely) neglected dimensions of Recall and Specificity can guide the satisfaction of another important epistemic goal, best captured by the locution ‘know-most/all.’ The upshot is that different epistemic goals call for high levels of reliability along different dimensions (and, often, their combinations too). This observation invites upgrading reliabilism into dimensional reliabilism, according to which assessing whether a process is knowledge-conducive depends on whether it manifests high reliability along the appropriate dimension(s), given the epistemic goal at hand. (shrink)

What justifies our beliefs about what other people say? According to epistemic inferentialism​, the justification of comprehension-based beliefs depends on the justification of other beliefs, e.g., beliefs about what words the speaker uttered or even what sounds they produced. According to epistemic non-inferentialism, the justification of comprehension-based beliefs ​does not depend on the justification of other beliefs. This paper offers a new defense of epistemic non-inferentialism. First, I discuss three counterexamples to epistemic non-inferentialism provided recently by Brendan Balcerak Jackson. I (...) argue that only one of Balcerak Jackson’s counterexamples is effective, and that it is effective against only one version of epistemic non-inferentialism, viz. language comprehension dogmatism. Second, I propose an alternative version of epistemic non-inferentialism, viz. comprehension-process reliabilism, which is immune to these counterexamples. I conclude that we should follow Balcerak Jackson in his rejection of language comprehension dogmatism but not all the way to the endorsement of epistemic inferentialism. Comprehension-process reliabilism is superior to both these alternatives. (shrink)

Advanced medical imaging, such as CT, fMRI and PET, has undergone enormous progress in recent years, both in accuracy and utilization. Such techniques often bring with them an illusion of immediacy, the idea that the body and its diseases can be directly inspected. In this paper we target this illusion and address the issue of the reliability of advanced imaging tests as knowledge procedures, taking positron emission tomography in oncology as paradigmatic case study. After individuating a suitable notion of reliability, (...) we argue that PET is a highly theory-laden and non-immediate knowledge procedure, in spite of the photographic-like quality of the images it delivers; the diagnostic conclusions based on the interpretation of PET images are population-dependent; PET images require interpretation, which is inherently observer-dependent and therefore variable. We conclude with a three-step methodological proposal for enhancing the reliability of advanced medical imaging. (shrink)

The author investigates whether the model prevalent today of an “humble reason” - based on fallibilism and epistemic humility - is the most appropriate to express the theological truth, even in the light of the debate within the contemporary theism (rational theology). To answer this question it is necessary to examine the epistemological status of “human truth” and the “truth of faith”, in order to develop a common approach to sciences, philosophy and theology. Finally, the author shows how the communitarian (...) dimension and the “choice” (a call for freedom) are inherent to theology understood as “critical faith”, whose peculiarity is to posit itself without nostalgia in the middle between fideism, rationalism and a certain relativism. From the path, finally, it emerges the proposal of a new “cristeologia”. Il rapporto tra fede e ragione è la questione meta-teologica per eccellenza. L’autore si propone dunque d’indagare se il modello oggi prevalente di “ragione umile” – basato su fallibilismo e umiltà epistemica – sia il più adeguato per esprimere le verità teologiche, anche alla luce del dibattito interno al teismo contemporaneo (teologia razionale). Per rispondere a questa domanda è necessario esaminare lo statuto epistemologico della verità umana e della verità di fede, per poter elaborare un metodo comune alle discipline scientifiche, filosofiche e teologiche, capace di trovare un equilibrio tra fideismo e positivismo teologico. Dopo una breve panoramica storica dei rapporti tra fede e ragione nel pensiero occidentale (con particolare attenzione alla dottrina dei praeambula fidei), l’articolo cercherà di mostrare come, nel paradigma della verità relazionale e del fallibilismo sia necessario integrare il problema delle fonti, dell’autorità, del concetto di ispirazione e della nozione di “rivelazione”, giacché contraddistinguono la teologia rispetto alla filosofia. Si mostrerà infine come la dimensione comunitaria e della scelta (appello alla libertà) siano connaturate alla teologia intesa come fede critica, la cui peculiarità è di collocarsi senza nostalgie nel mezzo tra fideismo, razionalismo e un certo relativismo. Dal percorso emergerà infine la proposta di una nuova "cristeologia". (shrink)

The generality problem is commonly considered to be a critical difficulty for reliabilism. In this paper, we present a dynamical perspective on the problem in the spirit of naturalized epistemology. According to this outlook, it is worth investigating how token belief-forming processes instantiate specific types in the biological agent’s cognitive architecture and background experience, consisting in the process of attractor-guided neural activation. While our discussion of the generality problem assigns “scientific types” to token processes, it represents a unified account in (...) the sense that it incorporates contextual and common sense features emphasized by other authors. (shrink)

In Weaving the Web, Berners-Lee defines Social Machines as biotechnologically hybrid Web-processes on the basis of which, “high-level activities, which have occurred just within one human’s brain, will occur among even larger more interconnected groups of people acting as if the shared a larger intuitive brain”. The analysis and design of Social Machines has already started attracting considerable attention both within the industry and academia. Web science, however, is still missing a clear definition of what a Social Machine is, which (...) has in turn resulted in several calls for a “philosophical engineering” ; Halpin et al. 2010). This paper is a first attempt to respond to this call, by combining contemporary philosophy of mind and cognitive science with epistemology. The idea of philosophical engineering implies that a sufficiently good conception of Social Machines should be of both theoretical and practical advantage. To demonstrate how the present approach can satisfy both objectives it will be used in order to address one of Wikipedia’s most worrying concerns—i.e., the current and ongoing decline in the number of its active contributors. (shrink)

This paper develops an account of the opacity problem in medical machine learning (ML). Guided by pragmatist assumptions, I argue that opacity in ML models is problematic insofar as it potentially undermines the achievement of two key purposes: ensuring generalizability and optimizing clinician–machine decision-making. Three opacity amelioration strategies are examined, with explainable artificial intelligence (XAI) as the predominant approach, challenged by two revisionary strategies in the form of reliabilism and the interpretability by design. Comparing the three strategies, I argue that (...) interpretability by design is most promising to overcome opacity in medical ML. Looking beyond the individual opacity amelioration strategies, the paper also contributes to a deeper understanding of the problem space and the solution space regarding opacity in medical ML. (shrink)

ABSTRACTEvidence, beliefs, and knowledge are words that are currently used in the intelligence community. The primary goal of intelligence is to provide knowledge and foreknowledge of an enemy’s in...