Statistical fairness criteria are widely used for diagnosing and ameliorating algorithmic bias. However, these fairness criteria are controversial as their use raises several difficult questions. I argue that the major problems for statistical algorithmic fairness criteria stem from an incorrect understanding of their nature. These criteria are primarily used for two purposes: first, evaluating AI systems for bias, and second constraining machine learning optimization problems in order to ameliorate such bias. The first purpose typically involves treating each criterion as a (...) necessary condition for fairness. The second use involves treating criteria as sufficient conditions for fairness. Since the criteria are used for both roles, some researchers have treated them as both necessary and sufficient conditions, i.e., as definitions of algorithmic fairness. However, serious problems have been raised for the use of these fairness criteria. Under ordinary circumstances, it is impossible to satisfy multiple criteria at the same time. Moreover, there are counterexamples to both the sufficiency and necessity for fairness of each criterion. I argue that we should instead understand fairness criteria as merely providing evidence of fairness. In other words, satisfaction (or violation) of these criteria should be understood as potential evidence of fairness (or bias). Whether a criterion counts as evidence in a particular case will depend on stakeholders' background knowledge and the specific features of the system's task. This evidence account of fairness conditions provides guidance for recognizing both the appropriate uses and the limitations of fairness criteria. (shrink)

Sometimes inquirers may rationally pursue a theory even when the available evidence does not favor that theory over others. Features of a theory that favor pursuing it are known as considerations of promise or pursuitworthiness. Examples of such reasons include that a theory is testable, that it has a useful associated analogy, and that it suggests new research and experiments. These reasons need not be evidence in favor of the theory. This raises the question: what kinds of reasons are provided (...) by pursuitworthiness considerations? Are they epistemic reasons or practical reasons? I argue that pursuitworthiness considerations are a kind of non-evidential epistemic reason, which I call an inquisitive reason. In support of this, I first point out two important similarities between the traditional pursuitworthiness considerations discussed in philosophy of science, which I call promise reasons, and certain social epistemic reasons that I call social inquisitive reasons. Specifically, both kinds of reason (1) favor pursuing a theory in a non-evidential way, and (2) concern promoting successful inquiry. I then propose recognition of a new category of normative reason: inquisitive reasons. This category contains both promise and social inquisitive reasons. Finally, I argue that inquisitive reasons share three essential features with previously recognized epistemic reasons: a connection to epistemic aims, explanatory independence, and the presence of a specific right-kind/wrong-kind reasons distinction. Each of these features have been used to argue that evidence should be treated as part of a distinct, independent domain of epistemic normativity. Since inquisitive reasons share these features, they too should be considered part of this independent epistemic domain. Thus, inquisitive reasons, including pursuitworthiness considerations, are epistemic reasons. (shrink)

I argue that recognizing a distinct doxastic attitude called endorsement, along with the epistemic norms governing it, solves the self-undermining problem for conciliationism about disagreement. I provide a novel account of how the self-undermining problem works by pointing out the auxiliary assumptions the objection relies on. These assumptions include commitment to certain epistemic principles linking belief in a theory to following prescriptions of that theory. I then argue that we have independent reason to recognize the attitude of endorsement. Endorsement is (...) the attitude of resilient and committed advocacy which is appropriate for researchers to have toward their own theory. Recognizing the importance of endorsement, and of its resiliency, gives us reason to deny the epistemic principles that serve as auxiliary assumptions in the self-undermining objection. This defuses the objection, and provides additional support for the theory of endorsement. (shrink)

[Wilfrid Hodges] During the last forty or so years it has become popular to offer explanations of logical notions in terms of games. There is no doubt that many people find games helpful for understanding various logical phenomena. But we ask whether anything is really 'explained' by these accounts, and we analyse Paul Lorenzen's dialogue foundations for constructive logic as an example. The conclusion is that the value of games lies in their ability to provide helpful metaphors and representations, (...) rather than in any true conceptual analysis. In fact some of the standard explanations of logical notions in terms of competitive games simply don't work. /// [Erik C. W. Krabbe] In an attempt to redeem the Lorenzen-type dialogues from their detractors, it is perhaps best first to provide a survey of the various benefits these dialogues have been supposed to yield. This will be done in Section I. It will not be possible, within the confines of this paper, to scrutinize them all, but in Section II we shall delve deeper into the capacity of this type of dialogue to yield a model for the immanent criticism of philosophical positions. Section III will extend the concept of a dialogue in such a way as to conform better with our intuitive conceptions of what a rational discussion of a position should contain. This will be followed up by a concept of 'winning a dialogue' that takes position midway between the old conception of 'winning one play' and that of the full-fledged presentation of a winning strategy. Concepts of 'rational discussion' are thus shown to be, plausibly, more fundamental than those of proof. In Section IV, I shall discuss the specific problems about dialogical foundations put forward by Wilfrid Hodges. (shrink)

This paper presents the epistemic practices account, a theory about the nature of epistemic normativity. The account aims to explain how the pursuit of epistemic values such as truth and knowledge can give rise to epistemic norms. On this account, epistemic norms are the internal rules of epistemic social practices. The account explains four crucial features of epistemic normativity while dissolving some apparent tensions between them. The account also provides a unified theory of epistemic and zetetic normativity.

Scientists, philosophers, and other researchers commonly assert their theories. This is surprising, as there are good reasons for skepticism about theories in cutting-edge research. I propose a new account of assertion in research contexts that vindicates these assertions. This account appeals to a distinct propositional attitude called endorsement, which is the rational attitude of committed advocacy researchers have to their theories. The account also appeals to a theory of conversational pragmatics known as the Question Under Discussion model, or QUD. Hence, (...) I call the theory the EQUD model. Motivating this account is a recognition that the speech act of assertion has two roles to play in research contexts. The first is the advocacy role, in which researchers assert a theory in order to advocate for it. The second is the evidential role, which is used to add to the common stock of information available to a field of inquiry. The EQUD model provides an account of warranted assertion for both these roles in research contexts. This success provides support for the theory of endorsement. It also provides support for information updating accounts of assertion. (shrink)

It is valuable for inquiry to have researchers who are committed advocates of their own theories. However, in light of pervasive disagreement, such a commitment is not well explained by the idea that researchers believe their theories. Instead, this commitment, the rational attitude to take toward one’s favored theory during the course of inquiry, is what I call endorsement. Endorsement is a doxastic attitude, but one which is governed by a different type of epistemic rationality. This inclusive epistemic rationality is (...) sensitive to reasons beyond those to think the particular proposition in question is true. Instead, it includes extrinsic epistemic reasons, which concern the health of inquiry more generally. Such extrinsic reasons include the distribution of cognitive labor that a researcher will contribute to by endorsing a particular theory. Recognizing endorsement and inclusive epistemic rationality thus allows us to smooth a tension between individual rationality and collective rationality. It does so by showing how it can be epistemically rational to endorse a theory on the basis of the way this endorsement will benefit collective inquiry. I provide a decision theoretic treatment for inclusive epistemic rationality and endorsement which illustrates how this can be accomplished. (shrink)

Thought: A Journal of Philosophy, EarlyView.

Many AI systems that make important decisions are black boxes: how they function is opaque even to their developers. This is due to their high complexity and to the fact that they are trained rather than programmed. Efforts to alleviate the opacity of black box systems are typically discussed in terms of transparency, interpretability, and explainability. However, there is little agreement about what these key concepts mean, which makes it difficult to adjudicate the success or promise of opacity alleviation methods. (...) I argue for a unified account of these key concepts that treats the concept of understanding as fundamental. This allows resources from the philosophy of science and the epistemology of understanding to help guide opacity alleviation efforts. A first significant benefit of this understanding account is that it defuses one of the primary, in-principle objections to post hoc explainable AI (XAI) methods. This “rationalization objection” argues that XAI methods provide mere rationalizations rather than genuine explanations. This is because XAI methods involve using a separate “explanation” system to approximate the original black box system. These explanation systems function in a completely different way than the original system, yet XAI methods make inferences about the original system based on the behavior of the explanation system. I argue that, if we conceive of XAI methods as idealized scientific models, this rationalization worry is dissolved. Idealized scientific models misrepresent their target phenomena, yet are capable of providing significant and genuine understanding of their targets. (shrink)

Discussion of epistemic responsibility typically focuses on belief formation and actions leading to it. Similarly, accounts of collective epistemic responsibility have addressed the issue of collective belief formation and associated actions. However, there has been little discussion of collective responsibility for preventing epistemic harms, particularly those preventable only by the collective action of an unorganized group. We propose an account of collective epistemic responsibility which fills this gap. Building on Hindriks' (2019) account of collective moral responsibility, we introduce the Epistemic (...) Duty to Join Forces. Our theory provides an account of the responsibilities of scientists to prevent epistemic harms during inquiry. (shrink)

One of the main lines of research in algorithmic fairness involves individual fairness (IF) methods. Individual fairness is motivated by an intuitive principle, similar treatment, which requires that similar individuals be treated similarly. IF offers a precise account of this principle using distance metrics to evaluate the similarity of individuals. Proponents of individual fairness have argued that it gives the correct definition of algorithmic fairness, and that it should therefore be preferred to other methods for determining fairness. I argue that (...) individual fairness cannot serve as a definition of fairness. Moreover, IF methods should not be given priority over other fairness methods, nor used in isolation from them. To support these conclusions, I describe four in-principle problems for individual fairness as a definition and as a method for ensuring fairness: (1) counterexamples show that similar treatment (and therefore IF) are insufficient to guarantee fairness; (2) IF methods for learning similarity metrics are at risk of encoding human implicit bias; (3) IF requires prior moral judgments, limiting its usefulness as a guide for fairness and undermining its claim to define fairness; and (4) the incommensurability of relevant moral values makes similarity metrics impossible for many tasks. In light of these limitations, I suggest that individual fairness cannot be a definition of fairness, and instead should be seen as one tool among several for ameliorating algorithmic bias. (shrink)

Inference and meaning -- Some reflections on language games -- Language as thought and as communication -- Meaning as functional classification : a perspective on the relation of syntax to semantics -- Naming and saying -- Grammar and existence : a preface to ontology -- Abstract entities -- Being and being known -- The lever of Archimedes -- Some reflections on thoughts and things -- Mental events -- Phenomenalism -- The identity approach to the mind-body problem -- Philosophy and the (...) scientific image of man -- "...this I or he or it (the thing) which thinks..." -- Some remarks on Kant's theory of experience -- The role of imagination in Kant's theory of experience. (shrink)

Intellectual courage requires acting to promote epistemic goods despite significant risk of harm. Courage is distinguished from recklessness and cowardice because the expected epistemic benefit of a courageous action outweighs (in some sense) the threatened harm. Sometimes, however, inquirers pursue theories that are not best supported by their current evidence. For these inquirers, the expected epistemic benefit of their actions cannot be explained by appeal to their evidence alone. The probability of pursuing the true theory cannot contribute enough to the (...) expected epistemic benefit for the action to count as courageous rather than reckless. Thus, there must be some other epistemic consideration which favors their action, besides evidence for their theory. I argue that the proper account of intellectual courage requires recognition of inquisitive reasons: a distinct category of epistemic reasons which concern successful inquiry. This category includes reasons to think a theory itself is promising, e.g., that the theory suggests potentially fruitful new research. It also includes social epistemic reasons, e.g., that pursuing a theory will improve the distribution of cognitive labor. Inquisitive reasons help explain why researchers who pursue improbable theories count as intellectually courageous, rather than reckless. The expected epistemic benefit of an action is partially determined by the inquisitive reasons in its favor. On my account, intellectually courageous inquiry requires acting in a way that is sensitive to inquisitive reasons. (shrink)

1. The lever in question is, of course, that with which, provided that an appropriate fulcrum could be found, Archimedes could move the world. In the analogy I have in mind, the fulcrum is the given, by virtue of which the mind gets leverage on the world of knowledge.

Virtue epistemology has been divided into two camps: reliabilists and responsibilists. This division has been attributed in part to a focus on different types of virtues, viz., faculty virtues and character virtues. I will argue that this distinction is unhelpful, and that we should carve up the theoretical terrain differently. Making several better distinctions among virtues will show us two important things. First, that responsibilists and reliabilists are actually engaged in different, complementary projects; and second, that certain responsibilist critiques of (...) reliabilism miss the mark. With these distinctions on the table, we can see that the virtue reliabilist project is in some ways more fundamental than the responsibilist project, since the latter importantly depends on the former. I argue that the distinctively epistemic value of the responsibilist’s character virtues is derived from their connections to the reliabilist’s constitutive virtues. While this will give us a unified account of the epistemic value of intellectual virtues, it is not a reduction of the responsibilist project to the reliabilist one; rather, it as a way of securing the separate importance of each project by clarifying how they relate to one another. (shrink)

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.

Bayesian confirmation theory is our best formal framework for describing inductive reasoning. The problem of old evidence is a particularly difficult one for confirmation theory, because it suggests that this framework fails to account for central and important cases of inductive reasoning and scientific inference. I show that we can appeal to the fragmentation of doxastic states to solve this problem for confirmation theory. This fragmentation solution is independently well-motivated because of the success of fragmentation in solving other problems. I (...) also argue that the fragmentation solution is preferable to other solutions to the problem of old evidence. These other solutions are already committed to something like fragmentation, but suffer from difficulties due to their additional commitments. If these arguments are successful, Bayesian confirmation theory is saved from the problem of old evidence, and the argument for fragmentation is bolstered by its ability to solve yet another problem. (shrink)

Formal implication is usually represented by symbolization such as ‘ φx ⊃ Ψx,’ which may be read, “for all values of ‘x’, φx implies Ψx.” If the values of the variable ‘x’, in ‘φx’ and ‘Ψx’ be ‘x1’ ‘x2’ ‘x3’, etc., then … ‘φx’ formally implies ‘Ψx’ if and only if, whatever values of ‘x’, ‘xn’, be chosen, ‘φxn’ materially implies ‘Ψxn’ …However, this still leaves it doubtful which of two possible interpretations of expressions having the form ‘ φx ⊃ (...) Ψx’ is to be taken as correct. … It means one thing to say, “Every existent having the property φ … has also the property Ψ,” and it means quite a different thing to say, “Every thinkable thing which should have the property φ must also have the property Ψ.” The second of these holds only when having the properly φ logically entails having the property Ψ; when ‘Ψx’ is deductible from ‘φx’. … The first of them, however, holds not only in such cases … but also in every case where among existent things, one property is universally accompanied by another. (shrink)

It is often irrational to believe philosophical claims because they are subject to systematic disagreement, under-determination, and pessimistic induction. Along with some other authors in this volume, I argue that many philosophers should (and do) have a different attitude to their own philosophical commitments. On my account, this attitude is a form of epistemic acceptance called endorsement. However, several objections have been raised to this view and others like it. One worry is that endorsement is spineless: that people who merely (...) endorse their theories fail to have a stable, global commitment to their views. A second worry is that endorsement, and the assertions it licenses, are objectionably insincere. Here, I defend the theory of endorsement from these objections. I suggest that endorsement can be more resilient than belief, while better supporting both intellectual courage and humility. Then, I argue that endorsement can be perfectly sincere, since it is not deceptive. (shrink)

In 2003 Zia Movahed pointed to a passage of Avicenna, written probably in 1022, which Movahed claimed anticipated the modal formula of Barcan (that ‘For every |$x$| necessarily |$\phi $|’ entails ‘Necessarily for every |$x$||$\phi $|’), and its converse. Since 2003, examination of early logical writings of Avicenna has clarified how he understood entailments between modal sentences, using his own new temporal language to provide a kind of semantics. In the light of that, Movahed’s claim for the Barcan formula needs (...) some tidying up but is basically correct. But by his semantics Avicenna should not have claimed the converse Barcan formula and probably didn’t claim it. (shrink)

A collection of some of Sellars' lectures and articles from 1951 to 1962.

The contemporary canon of what constitutes ancient political thought was fixed in the course of the nineteenth century by the then newly reigning discipline of the philosophy of history. It made little difference whether this discipline was positivistically or dialectically inclined. Whatever the methodological commitment there was general agreement that the sources of ancient wisdom on the nature and ends of social and political life were to be found in the political and ethical writings of Plato and Aristotle and, to (...) a lesser extent, Cicero. Here Cicero was read, in part, as a repository of a Stoic political wisdom that had not survived in the original texts. According to this tradition, genuine, if not final, political wisdom is to be found in the reflections of these ancient thinkers -- not a totally unexpected conclusion to come from philosophers. For these nineteenth-century practitioners of Geistesgeschichte the essence of ancient experience is distilled in its inner life, its art, literature, religion, and above all in its philosophical speculation, but not, for instance, in its political experience as recorded in its history. (shrink)

We consider two formalisations of the notion of a compositionalsemantics for a language, and find some equivalent statements in termsof substitutions. We prove a theorem stating necessary and sufficientconditions for the existence of a canonical compositional semanticsextending a given partial semantics, after discussing what features onewould want such an extension to have. The theorem involves someassumptions about semantical categories in the spirit of Husserl andTarski.

We trace two logical ideas further back than they have previously been traced. One is the idea of using diagrams to prove that certain logical premises do—or don’t—have certain logical consequences. This idea is usually credited to Venn, and before him Euler, and before him Leibniz. We find the idea correctly and vigorously used by Abū al-Barakāt in 12th century Baghdad. The second is the idea that in formal logic, P logically entails Q if and only if every model of (...) P is a model of Q. This idea is usually credited to Tarski, and before him Bolzano. But again we find Abū al-Barakāt already exploiting the idea for logical calculations. Abū al-Barakāt’s work follows on from related but inchoate research of Ibn Sīnā in eleventh century Persia. We briefly trace the notion of model-theoretical consequence back through Paul the Persian and in some form back to Aristotle himself. (shrink)

The most important work by one of America's greatest twentieth-century philosophers, Empiricism and the Philosophy of Mind is both the epitome of Wilfrid Sellars' entire philosophical system and a key document in the history of philosophy. First published in essay form in 1956, it helped bring about a sea change in analytic philosophy. It broke the link, which had bound Russell and Ayer to Locke and Hume--the doctrine of "knowledge by acquaintance." Sellars' attack on the Myth of the Given (...) in Empiricism and the Philosophy of Mind was a decisive move in turning analytic philosophy away from the foundationalist motives of the logical empiricists and raised doubts about the very idea of "epistemology." With an introduction by Richard Rorty to situate the work within the history of recent philosophy, and with a study guide by Robert Brandom, this publication of Empiricism and the Philosophy of Mind makes a difficult but indisputably significant figure in the development of analytic philosophy clear and comprehensible to anyone who would understand that philosophy or its history. (shrink)

§1. Introduction. I dedicate this essay to the two-dozen-odd people whose refutations of Cantor's diagonal argument have come to me either as referee or as editor in the last twenty years or so. Sadly these submissions were all quite unpublishable; I sent them back with what I hope were helpful comments. A few years ago it occurred to me to wonder why so many people devote so much energy to refuting this harmless little argument—what had it done to make them (...) angry with it? So I started to keep notes of these papers, in the hope that some pattern would emerge.These pages report the results. They might be useful for editors faced with similar problem papers, or even for the authors of the papers themselves. But the main message to reach me is that there are several points of basic elementary logic that we usually teach and explain very badly, or not at all.In 1995 an engineer named William Dilworth, who had published a refutation of Cantor's argument in the Transactions of the Wisconsin Academy of Sciences, Arts and Letters, sued for libel a mathematician named Underwood Dudley who had called him a crank. (shrink)

The author's primary objective is to show that versions of legal positivism, according to which legal validity sometimes depends on moral validity (Inclusive Legal Positivism), are theoretically preferable to those forms of positivism (Exclusive Legal Positivism) which deny this possibility. The author attempts to substantiate this conclusion by demonstrating that Inclusive Legal Positivism provides a better theoretical account of challenges to legal validity based on a document like the Canadian Charter of Rights and Freedoms. His secondary aim is to show (...) that the choice between Inclusive and Exclusive Legal Positivism can have important consequences for legal practice. (shrink)