Philosophy of science was once a much more socially engaged endeavor, and can be so again. After a look back at philosophy of science in the 1930s-1950s, I turn to discuss the current potential for returning to a more engaged philosophy of science. Although philosophers of science have much to offer scientists and the public, I am skeptical that much can be gained by philosophers importing off-the-shelf discussions from philosophy of science to science and society. Such efforts will likely look (...) like efforts to do applied ethics by merely applying ethical theories to particular contexts and problems. While some insight can be gained by these kinds of endeavors, the most interesting and pressing problems for the actual practitioners and users of science are rarely addressed. Instead, I recommend that philosophers of science engage seriously and regularly with scientists and/or the users of science in order to gain an understanding of the conceptual issues on the ground. From such engagement, flaws in the traditional philosophical frameworks, and how such flaws can be remedied, become apparent. Serious engagement with the contexts of science thus provides the most fruit for philosophy of science per se and for the practitioners whom the philosophers aim to assist. And if one focuses on contexts where science has its most social relevance, these efforts can help to provide the thing that philosophy of science now lacks: a full-bodied philosophy of science in society. (shrink)

Traditionally, cognitive values have been thought of as a collective pool of considerations in science that frequently trade against each other. I argue here that a finer-grained account of the value of cognitive values can help reduce such tensions. I separate the values into groups, minimal epistemic criteria, pragmatic considerations, and genuine epistemic assurance, based in part on the distinction between values that describe theories per se and values that describe theory-evidence relationships. This allows us to clarify why these values (...) are central to science and what role they should play, while reducing the tensions among them. (shrink)

The doctrinal paradox shows that aggregating individual judgments by taking a majority vote does not always yield a consistent set of collective judgments. Philip Pettit, Luc Bovens, and Wlodek Rabinowicz have recently argued for the epistemic superiority of an aggregation procedure that always yields a consistent set of judgments. This paper identifies several additional epistemic advantages of their consistency maintaining procedure. However, this paper also shows that there are some circumstances where the majority vote procedure is epistemically superior. The epistemic (...) value of maintaining consistency does not always outweigh the epistemic value of making true judgments. (shrink)

Although epistemic values have become widely accepted as part of scientific reasoning, non-epistemic values have been largely relegated to the "external" parts of science (the selection of hypotheses, restrictions on methodologies, and the use of scientific technologies). I argue that because of inductive risk, or the risk of error, non-epistemic values are required in science wherever non-epistemic consequences of error should be considered. I use examples from dioxin studies to illustrate how non-epistemic consequences of error can and should be considered (...) in the internal stages of science: choice of methodology, characterization of data, and interpretation of results. (shrink)

Two goals - improving replicability and accountability of Machine Learning research respectively, have accrued much attention from the AI ethics and the Machine Learning community. Despite sharing the measures of improving transparency, the two goals are discussed in different registers - replicability registers with scientific reasoning whereas accountability registers with ethical reasoning. Given the existing challenge of the Responsibility Gap - holding Machine Learning scientists accountable for Machine Learning harms due to them being far from sites of application, this paper (...) posits that reconceptualizing replicability can help bridge the gap. Through a shift from model performance replicability to claim replicability, Machine Learning scientists can be held accountable for producing non-replicable claims that are prone to eliciting harm due to misuse and misinterpretation. In this paper, I make the following contributions. First, I define and distinguish two forms of replicability for ML research that can aid constructive conversations around replicability. Second, I formulate an argument for claim-replicability's advantage over model performance replicability in justifying assigning accountability to Machine Learning scientists for producing non-replicable claims and show how it enacts a sense of responsibility that is actionable. In addition, I characterize the implementation of claim replicability as more of a social project than a technical one by discussing its competing epistemological principles, practical implications on Circulating Reference, Interpretative Labor, and research communication. (shrink)

In this paper I shall defend the idea that there is an abstract and general core meaning of objectivity, and what is seen as a variety of concepts or conceptions of objectivity are in fact criteria of, or means to achieve, objectivity. I shall then discuss the ideal of value-free science and its relation to the objectivity of science; its status can be at best a criterion of, or means for, objectivity. Given this analysis, we can then turn to the (...) problem of inductive risk. Do the value judgements regarding inductive risk really pose a threat to the objectivity of science? I claim that this is not the case because they do not lower the thresholds scientifically postulated for objectivity. I shall conclude the paper with a discussion of under-appreciated influences of values on science, which indeed pose a serious threat to the objectivity of some scientific disciplines. (shrink)

We analyse the issue of using prior information in frequentist statistical inference. For that purpose, we scrutinise different kinds of sampling designs in Jerzy Neyman’s theory to reveal a variety of ways to explicitly and objectively engage with prior information. Further, we turn to the debate on sampling paradigms (design-based vs. model-based approaches) to argue that Neyman’s theory supports an argument for the intermediate approach in the frequentism vs. Bayesianism debate. We also demonstrate that Neyman’s theory, by allowing non-epistemic values (...) to influence evidence collection and formulation of statistical conclusions, does not compromise the epistemic reliability of the procedures and may improve it. This undermines the value-free ideal of scientific inference. (shrink)

The argument from inductive risk, as developed by Rudner and others, famously concludes that the scientist qua scientist makes value judgments. The paper aims to show that trust in the soundness of the argument is overrated – that philosophers who endorse its conclusion fail to refute two of the most important objections that have been raised to its soundness: Jeffrey’s objection that the genuine task of the scientist is to assign probabilities to hypotheses, and Levi’s objection that the argument is (...) ambiguous about decisions about how to act and decisions about what to believe, that only the former presuppose value judgments, and that qua scientist, the scientist only needs to decide what to believe. (shrink)

This article develops a constructive criticism of methodological conventionalism. Methodological conventionalism asserts that standards of inductive risk ought to be justified in virtue of their ability to facilitate coordination in a research community. On that view, industry bias occurs when conventional methodological standards are violated to foster industry preferences. The underlying account of scientific conventionality, however, is problematically incomplete. Conventions may be justified in virtue of their coordinative functions, but often qualify for posterior empirical criticism as research advances. Accordingly, industry (...) bias does not only threaten existing conventions but may impede their empirically warranted improvement if they align with industry preferences. My empiricist account of standards of inductive risk avoids such a problem by asserting that conventional justification can be pragmatically warranted but has, in principle, only a provisional status. Methodological conventions, therefore, should not only be defended from preference-based infringements on their coordinative function but ought to be subjected to empirical criticism. (shrink)

This paper articulates and defends three interconnected claims: first, that the debate on the role of values for science misses a crucial dimension, the institutional one; second, that institutions occupy the intermediate level between scientific activities and values and that they are to be systematically integrated into the analysis; third, that the appraisal of the institutions of science with respect to values should be undertaken within the premises of a comparative approach rather than an ideal approach. Hence, I defend the (...) view that the issue be framed in reference to the following question: "What kind of institutional rules should be in place in order for the scientific process to unfold in such a way that the values that we deem more important come to the fore?" Addressing this concern is equivalent to conducting a debate on institutions and their role for science. (shrink)

Deepfakes are realistic videos created using new machine learning techniques rather than traditional photographic means. They tend to depict people saying and doing things that they did not actually say or do. In the news media and the blogosphere, the worry has been raised that, as a result of deepfakes, we are heading toward an “infopocalypse” where we cannot tell what is real from what is not. Several philosophers have now issued similar warnings. In this paper, I offer an analysis (...) of why deepfakes are such a serious threat to knowledge. Utilizing the account of information carrying recently developed by Brian Skyrms, I argue that deepfakes reduce the amount of information that videos carry to viewers. I conclude by drawing some implications of this analysis for addressing the epistemic threat of deepfakes. (shrink)

I propose a general alethic theory of epistemic risk according to which the riskiness of an agent’s credence function encodes her relative sensitivity to different types of graded error. After motivating and mathematically developing this approach, I show that the epistemic risk function is a scaled reflection of expected inaccuracy. This duality between risk and information enables us to explore the relationship between attitudes to epistemic risk, the choice of scoring rules in epistemic utility theory, and the selection of priors (...) in Bayesian epistemology more generally. (shrink)

In this paper, I discuss some of the recent developments in the political turn of Social Epistemology, focusing on the notions of epistemic injustice and epistemic oppression. In the first part of the work, I introduce Kristie Dotson’s characterization of the epistemic injustices presented by Miranda Fricker, through the understanding of systematic ways of violating epistemic agency in terms of oppressions. In the second part, I discuss Dotson’s critique of Fricker on the grounds that there is an important kind of (...) epistemic injustice that is not dealt with in Fricker’s proposal, and which could help us to perceive a particular, pervasive kind of violation of the epistemic agency. (shrink)

The debate on the epistemology of disagreement has so far focused almost exclusively on cases of disagreement between individual persons. Yet, many social epistemologists agree that at least certain kinds of groups are equally capable of having beliefs that are open to epistemic evaluation. If so, we should expect a comprehensive epistemology of disagreement to accommodate cases of disagreement between group agents, such as juries, governments, companies, and the like. However, this raises a number of fundamental questions concerning what it (...) means for groups to be epistemic peers and to disagree with each other. In this paper, we explore what group peer disagreement amounts to given that we think of group belief in terms of List and Pettit’s ‘belief aggregation model’. We then discuss how the so-called ‘equal weight view’ of peer disagreement is best accommodated within this framework. The account that seems most promising to us says, roughly, that the parties to a group peer disagreement should adopt the belief that results from applying the most suitable belief aggregation function for the combined group on all members of the combined group. To motivate this view, we test it against various intuitive cases, derive some of its notable implications, and discuss how it relates to the equal weight view of individual peer disagreement. (shrink)

This essay raises new objections to the two dominant approaches to understanding the justification of group beliefs—_inflationary_ views, where groups are treated as entities that can float freely from the epistemic status of their members’ beliefs, and _deflationary_ views, where justified group belief is understood as nothing more than the aggregation of the justified beliefs of the group's members. If this essay is right, we need to look in an altogether different place for an adequate account of justified group belief. (...) From these objections emerges the skeleton of the positive view that this essay goes on to develop and defend, called the _group epistemic agent account_: groups are epistemic agents in their own right, with justified beliefs that respond to both evidence and normative requirements that arise only at the group level but that are nonetheless importantly constrained by the epistemic status of the beliefs of their individual members. (shrink)

When interests and preferences of researchers or their sponsors cause bias in experimental design, data interpretation or dissemination of research results, we normally think of it as an epistemic shortcoming. But as a result of the debate on science and values, the idea that all extra-scientific influences on research could be singled out and separated from pure science is now widely believed to be an illusion. I argue that nonetheless, there are cases in which research is rightfully regarded as epistemologically (...) deficient due to the influence of preferences on its outcomes. I present examples from biomedical research and offer an analysis in terms of social epistemology. (shrink)

One of the most important issues in the philosophy of science in recent decades is to assess the permissibility of the involvement of background and non-scientific factors in science and to place them next to evidence or to involve them in the process of weighing evidence. Proponents of the value-free science ideal have considered any intervention of this kind as a blow to the objectivity of science and slipping in the process of science. One of the important arguments in criticizing (...) this ideal is that of inductive risk, according to which any scientific statement or hypothesis and theory based on the common scientific method,is subject to possible shortcomings that are sometimes so detrimental that it is essential to prevent them and repair the process of hypothesis or scientific theorizing. Therefore, due to the possibility of errors in non-perceptual consequences, non-perceptual factors, including moral, social, and political values, should be involved in the process of science and these factors determine what assumptions or theories are accepted to avoid those consequences. For the first, Hempel presents the argument as below: Based on certain evidence as well as the scientific rules governing the research question, the probable results are: The hypothesis is accepted according to scientific rules and is, in fact, true. The hypothesis is rejected on the basis of scientific rules and is, in fact, false. The hypothesis is accepted according to scientific rules, but it should, in fact, be false. The hypothesis is rejected onthe basis of scientific rules, but it should, in fact, be true. The first two hypotheses are the results of scientific processes, but the last two hypotheses are probabilities that induction will occur.So both epistemologically and practically, we may have unpleasant consequences that must be remedied by reconstructing the rules of accepting or rejecting scientific assumptions.Hempel’s solution is to involve values in the process of science, so that, although values lack a logical connection with hypotheses, their role in the rules of accepting hypotheses to avoid scientific errors and scientific consequences is justified. The argument is expressed today in a new form as follows: It is a common method in induction science. There is a possibility of error in induction. Scientific error leads to unfortunate individual and social consequences in the practical field. The possible consequences of this can be overcome with unscientific values. The hypothesis must be organized in such a way that it results in the least error. After the alternative assumptions, a case should be selected that has the least adverse consequences. So values can affect the process of science in a permissible and reasonable way. Some of the most important drawbacks of this argument are: Interference of two scientific and practical fields: In this argument, the position of practice has been used for the field of science and the criterion of applying theory and hypothesis has been included in its epistemic justification level. Lack of guarantee of objectivity of theories: If the criterion of objectivity is determined outside the position of opinion, there will be no guarantee that the theories will reveal reality. Basically, the meaning of objectivity is that it should not rely on any personal desires or uses. (shrink)

This paper presents arguments that challenge what I call the fact/value separability thesis: the idea, roughly, that factual judgements can be made independently of judgements of value. I will look at arguments to the effect that facts and values are entangled in the following areas of the scientific process in economics: theory development, economic concept formation, economic modelling, hypothesis testing, and hypothesis acceptance.

For any normative domain, we can theorize about what is good in that domain. Such theories include utilitarianism, a view about what is good morally. But there are many domains other than the moral; these include the prudential, the aesthetic, and the intellectual or epistemic. In this last domain, it is good to be knowledgeable and bad to ignore evidence, quite apart from the morality, prudence, and aesthetics of these things. This dissertation builds a theory that stands to the epistemic (...) domain as utilitarianism stands to the moral domain. It builds an epistemic value theory. (shrink)

These preprints were automatically compiled into a PDF from the collection of papers deposited in PhilSci-Archive in conjunction with the PSA 2012.

In order to guide the decisions of real people who want to bring about good epistemic outcomes for themselves and others, we need to understand our epistemic values. In Knowledge in a Social World, Alvin Goldman has proposed an epistemic value theory that allows us to say whether one outcome is epistemically better than another. However, it has been suggested that Goldman's theory is not really an epistemic value theory at all because whether one outcome is epistemically better than another (...) partly depends on our non-epistemic interests. In this paper, I argue that an epistemic value theory that serves the purposes of social epistemology must incorporate non- epistemic interests in much the way that Goldman's theory does. In fact, I argue that Goldman's theory does not go far enough in this direction. In particular, the epistemic value of having a particular true belief should actually be weighted by how interested we are in the topic. (shrink)

The few extant arguments concerning the autonomy of science in the rational acceptance of hypotheses are examined. It is concluded that science is not morally autonomous, and that the attendant notion of rationality in science decisionmaking is inadequate. A more comprehensive notion of scientific rationality, which encompasses the old one, is proposed as a replacement. The general idea is that scientists qua scientist ought, in their acceptance decisions, to take into account the ethical consequences of acceptance as well as the (...) consequences with regard to the attainment of "purely scientific" or "epistemic" objectives. The result constitutes an argument for a (presumably cooperative) game theoretic treatment of inductive logic. (shrink)

I use a theorem from machine learning, called the “No Free Lunch” theorem to support the claim that non-epistemic values are essential to theory choice. I argue that NFL entails that predictive accuracy is insufficient to favor a given theory over others, and that NFL challenges our ability to give a purely epistemic justification for using other traditional epistemic virtues in theory choice. In addition, I argue that the natural way to overcome NFL’s challenge is to use non-epistemic values. If (...) my argument holds, non-epistemic values are entangled in theory choice regardless of human limitations and regardless of the subject matter. Thereby, my argument overcomes objections to the main lines of argument revealing the role of values in theory choice. At the end of the paper, I argue that, contrary to common conception, the epistemic challenge arising from NFL is distinct from Hume’s problem of induction and other forms of underdetermination. (shrink)

Although a strict dichotomy between facts and values is no longer accepted, less attention has been paid to the roles values should play in our acceptance of factual statements, or scientific descriptive claims. This paper argues that values, whether cognitive or ethical, should never preclude or direct belief on their own. Our wanting something to be true will not make it so. Instead, values should only be used to consider whether the available evidence provides sufficient warrant for a claim. This (...) argument is made for all relevant values, including cognitive, ethical, and social values. The rational integrity of science depends not on excluding some values and including others in the reasoning process, but of constraining all values to their proper role in belief acceptance. (shrink)

The argument from inductive risk attempts to show that practical and ethical costs of errors should influence standards of evidence for accepting scientific claims. A common objection charges that this argument presupposes a behavioral theory of acceptance that is inappropriate for science. I respond by showing that the argument from inductive risk is supported by a nonbehavioral theory of acceptance developed by Cohen, which defines acceptance in terms of premising. Moreover, I argue that theories designed to explain how acceptance can (...) be guided exclusively by epistemic values suffer from difficulties that do not afflict Cohen’s theory. (shrink)

This paper seeks to defend the following conclusions: The program advanced by Carnap and other necessarians for probability logic has little to recommend it except for one important point. Credal probability judgments ought to be adapted to changes in evidence or states of full belief in a principled manner in conformity with the inquirer’s confirmational commitments—except when the inquirer has good reason to modify his or her confirmational commitment. Probability logic ought to spell out the constraints on rationally coherent confirmational (...) commitments. In the case where credal judgments are numerically determinate confirmational commitments correspond to Carnap’s credibility functions mathematically represented by so—called confirmation functions. Serious investigation of the conditions under which confirmational commitments should be changed ought to be a prime target for critical reflection. The necessarians were mistaken in thinking that confirmational commitments are immune to legitimate modification altogether. But their personalist or subjectivist critics went too far in suggesting that we might dispense with confirmational commitments. There is room for serious reflection on conditions under which changes in confirmational commitments may be brought under critical control. Undertaking such reflection need not become embroiled in the anti inductivism that has characterized the work of Popper, Carnap and Jeffrey and narrowed the focus of students of logical and methodological issues pertaining to inquiry. (shrink)

In pure science, the standard approach to non-epistemic values is to exclude them as far as possible from scientific deliberations. When science is applied to practical decisions, non-epistemic values cannot be excluded. Instead, they have to be combined with scientific information in a way that leads to practically optimal decisions. A normative model is proposed for the processing of information in both pure and applied science. A general-purpose corpus of scientific knowledge, with high entry requirements, has a central role in (...) this model. Due to its high entry requirements, the information that it contains is sufficiently reliable for the vast majority of practical purposes. However, for some purposes, the corpus needs to be supplemented with additional information, such as scientific indications of danger that do not satisfy the entry requirements for the corpus. The role of non-epistemic values in the evaluation of scientific information should, as far as possible, be limited to determining the level of evidence required for various types of practical decisions. (shrink)

Three of the major issues in information ethics – intellectual property, speech regulation, and privacy – concern the morality of restricting people’s access to certain information. Consequently, policies in these areas have a significant impact on the amount and types of knowledge that people acquire. As a result, epistemic considerations are critical to the ethics of information policy decisions (cf. Mill, 1978 [1859]). The fact that information ethics is a part of the philosophy of information highlights this important connection with (...) epistemology. In this paper, I illustrate how a value-theoretic approach to epistemology can help to clarify these major issues in information ethics. However, I also identify several open questions about epistemic values that need to be answered before we will be able to evaluate the epistemic consequences of many information policies. (shrink)

We all pursue epistemic goals as individuals. But we also pursue collective epistemic goals. In the case of many groups to which we belong, we want each member of the group - and sometimes even the group itself - to have as many true beliefs as possible and as few false beliefs as possible. In this paper, I respond to the main objections to the very idea of such collective epistemic goals. Furthermore, I describe the various ways that our collective (...) epistemic goals can come into conflict with each other. And I argue that we must appeal to pragmatic considerations in order to resolve such conflicts. (shrink)

Amidst long-running debates within the field, high energy physics has adopted a statistical methodology that primarily employs standard frequentist techniques such as significance testing and confidence interval estimation, but incorporates Bayesian methods for limited purposes. The discovery of the Higgs boson has drawn increased attention to the statistical methods employed within HEP. Here I argue that the warrant for the practice in HEP of relying primarily on frequentist methods can best be understood as pragmatic, in the sense that statistical methods (...) are chosen for their suitability for the practical demands of experimental HEP, rather than reflecting a commitment to a particular epistemic framework. In particular, I argue that understanding the statistical methodology of HEP through the perspective of pragmatism clarifies the role of and rationale for significance testing in the search for new phenomena such as the Higgs boson. (shrink)

Several different Bayesian models of epistemic utilities (see, e. g., [37], [24], [40], [46]) have been used to explain why it is rational for scientists to perform experiments. In this paper, I argue that a model-suggested independently by Patrick Maher [40] and Graham Oddie [46]-that assigns epistemic utility to degrees of belief in hypotheses provides the most comprehensive explanation. This is because this proper scoring rule (PSR) model captures a wider range of scientifically acceptable attitudes toward epistemic risk than the (...) other Bayesian models that have been proposed. I also argue, however, that even the PSR model places unreasonably tight restrictions on a scientist's attitude toward epistemic risk. As a result, such Bayesian models of epistemic utilities fail as normative accounts-not just as descriptive accounts (see, e. g., [31], [14])-of scientific inquiry. (shrink)

To resolve the impasse between skeptic, idealist and positivist as to whether or not historical inquiry can be objective, an affirmative answer is argued by exposing, clarifying and challenging the common presupposition: the thesis of scientific value neutrality. The argument applies a more explicit version of the Braithwaite— Churchman-Rudner position to history and thus challenges the prevalent claim that history, unlike the law, has but one goal, the establishment of truth about the past. The important yet neglected residual issue concerns (...) what 'objectivity' means when inquiry (historical or scientific) is construed as a goal-directed activity controlled by epistemic and pragmatic utilities. (shrink)

Many philosophers have challenged the ideal of value-free science on the grounds that social or moral values are relevant to inferential thresholds. But given this view, how precisely and to what extent should scientists adjust their inferential thresholds in light of nonepistemic values? We suggest that signal detection theory provides a useful framework for addressing this question. Moreover, this approach opens up further avenues for philosophical inquiry and has important implications for philosophical debates concerning inductive risk. For example, the signal (...) detection theory framework entails that considerations of inductive risk and inferential-threshold placement cannot be conducted in isolation from base-rate information. (shrink)

A large number of legal rules create an obligation to prevent, repair or otherwise mitigate damage to human health or the environment. Many of these rules require that a legally relevant causal relation between human behaviour and the damage at issue is established, and in the establishment of causal relations of this kind scientific information is often pressed into service. This thesis examines this specifically legal use of scientific information. It shows that many legally relevant causal relations cannot be established (...) in this way. It also shows that the legal strategy for dealing with epistemic difficulties by relaxing the standard of proof is of limited value in tackling this problem. Consequently, the legal rules focused upon are less efficient tools for combating damage to human health and the environment than their appearances suggest. Other legal means to achieve this aim should therefore be considered. The thesis begins by presenting a framework which can be used to solve some important problems with communication and the division of labour that arise when scientific information is used to answer legal questions. This framework recognizes both epistemological and ontological differences between law and science. The latter are differences between the kinds of entity which are relevant in distinct theoretical contexts; the former, by contrast, are differences between applicable standards of proof. In subsequent parts of the thesis this framework shapes the discussion of the question: To what extent can legally relevant causal relations, involving legally significant entities, be established through the application of scientific information about other, scientifically relevant entities and legal standards of proof. The thesis concludes with a discussion of some of the ways in which the legal regulation could, potentially, be adapted, so as to counteract damage to human health and the environment more efficiently. (shrink)

A rule for the acceptance of scientific hypotheses called 'the principle of cost-benefit dominance' is shown to be more effective and efficient than the well-known principle of the maximization of expected utility. Harvey 's defense of his theory of the circulation of blood in animals is examined as a historical paradigm case of a successful defense of a scientific hypothesis and as an implicit application of the cost-benefit dominance rule advocated here. Finally, various concepts of 'dominance' are considered by means (...) of which the effectiveness of our rule may be increased. (shrink)

A rule for the acceptance of scientific hypotheses called ‘the principle of cost-benefit dominance’ is shown to be more effective and efficient than the well-known principle of the maximization of expected (epistemic) utility. Harvey's defense of his theory of the circulation of blood in animals is examined as a historical paradigm case of a successful defense of a scientific hypothesis and as an implicit application of the cost-benefit dominance rule advocated here. Finally, various concepts of ‘dominance’ are considered by means (...) of which the effectiveness of our rule may be increased. (shrink)

This paper challenges the appeal to theory virtues in theory choice as well as the appeal to the intellectual and moral virtues of an agent as determining unique choices between empirically equivalent theories. After arguing that theoretical virtues do not determine the choice of one theory at the expense of another theory, I argue that nor does the appeal to intellectual and moral virtues single out one agent, who defends a particular theory, and exclude another agent defending an alternative theory. (...) I analyse Duhem’s concept of good sense and its recent interpretation in terms of virtue epistemology. I argue that the virtue epistemological interpretation does not show how good sense leads to conclusive choices and scientific progress. (shrink)