The debate about the ethical implications of Artificial Intelligence dates from the 1960s :741–742, 1960; Wiener in Cybernetics: or control and communication in the animal and the machine, MIT Press, New York, 1961). However, in recent years symbolic AI has been complemented and sometimes replaced by Neural Networks and Machine Learning techniques. This has vastly increased its potential utility and impact on society, with the consequence that the ethical debate has gone mainstream. Such a debate has primarily focused on principles—the (...) ‘what’ of AI ethics —rather than on practices, the ‘how.’ Awareness of the potential issues is increasing at a fast rate, but the AI community’s ability to take action to mitigate the associated risks is still at its infancy. Our intention in presenting this research is to contribute to closing the gap between principles and practices by constructing a typology that may help practically-minded developers apply ethics at each stage of the Machine Learning development pipeline, and to signal to researchers where further work is needed. The focus is exclusively on Machine Learning, but it is hoped that the results of this research may be easily applicable to other branches of AI. The article outlines the research method for creating this typology, the initial findings, and provides a summary of future research needs. (shrink)

Technologies to rapidly alert people when they have been in contact with someone carrying the coronavirus SARS-CoV-2 are part of a strategy to bring the pandemic under control. Currently, at least 47 contact-tracing apps are available globally. They are already in use in Australia, South Korea and Singapore, for instance. And many other governments are testing or considering them. Here we set out 16 questions to assess whether — and to what extent — a contact-tracing app is ethically justifiable.

As the range of potential uses for Artificial Intelligence, in particular machine learning, has increased, so has awareness of the associated ethical issues. This increased awareness has led to the realisation that existing legislation and regulation provides insufficient protection to individuals, groups, society, and the environment from AI harms. In response to this realisation, there has been a proliferation of principle-based ethics codes, guidelines and frameworks. However, it has become increasingly clear that a significant gap exists between the theory of (...) AI ethics principles and the practical design of AI systems. In previous work, we analysed whether it is possible to close this gap between the ‘what’ and the ‘how’ of AI ethics through the use of tools and methods designed to help AI developers, engineers, and designers translate principles into practice. We concluded that this method of closure is currently ineffective as almost all existing translational tools and methods are either too flexible or too strict. This raised the question: if, even with technical guidance, AI ethics is challenging to embed in the process of algorithmic design, is the entire pro-ethical design endeavour rendered futile? And, if no, then how can AI ethics be made useful for AI practitioners? This is the question we seek to address here by exploring why principles and technical translational tools are still needed even if they are limited, and how these limitations can be potentially overcome by providing theoretical grounding of a concept that has been termed ‘Ethics as a Service.’. (shrink)

did plato and aristotle have anything to say about belief? The answer to this question might seem blindingly obvious: of course they did. Plato distinguishes belief from knowledge in the Meno, Republic, and Theaetetus, and Aristotle does so in the Posterior Analytics. Plato distinguishes belief from perception in the Theaetetus, and Aristotle does so in the De anima. They talk about the distinction between true and false beliefs, and the ways in which belief can mislead and the ways in which (...) it can steer us aright. Indeed, they make belief a central component of their epistemologies.The view underlying these claims—one so widespread these days as to remain largely unquestioned—is that when Plato and Aristotle talk... (shrink)

Assertion is a fundamental feature of language. This volume will be the place to look for anyone interested in current work on the topic.

A persuasive monograph that answers the keyepistemological arguments against anti-individualism in thephilosophy of mind.

By mid-2019 there were more than 80 AI ethics guides available in the public domain. Despite this, 2020 saw numerous news stories break related to ethically questionable uses of AI. In part, this is because AI ethics theory remains highly abstract, and of limited practical applicability to those actually responsible for designing algorithms and AI systems. Our previous research sought to start closing this gap between the ‘what’ and the ‘how’ of AI ethics through the creation of a searchable typology (...) of tools and methods designed to translate between the five most common AI ethics principles and implementable design practices. Whilst a useful starting point, that research rested on the assumption that _all_ AI practitioners are aware of the ethical implications of AI, understand their importance, and are actively seeking to respond to them. In reality, it is unclear whether this is the case. It is this limitation that we seek to overcome here by conducting a mixed-methods qualitative analysis to answer the following four questions: what do AI practitioners understand about the need to translate ethical principles into practice? What motivates AI practitioners to embed ethical principles into design practices? What barriers do AI practitioners face when attempting to translate ethical principles into practice? And finally, what assistance do AI practitioners want and need when translating ethical principles into practice? (shrink)

Something Aristotle calls ‘right logos’ plays a crucial role in his theory of virtue. But the meaning of ‘logos’ in this context is notoriously contested. I argue against the standard translation ‘reason’, and—drawing on parallels with Plato’s work, especially the Laws—in favor of its being used to denote what transforms an inferior epistemic state into a superior one: an explanatory account. Thus Aristotelian phronēsis, like his and Plato’s technē and epistēmē, is a matter of grasping explanatory accounts: in this case, (...) accounts that identify the right action and say why it is right. Arguably, Aristotelian rationality is a matter of being able to grasp accounts in general. (shrink)

This article highlights the limitations of the tendency to frame health- and wellbeing-related digital tools (mHealth technologies) as empowering devices, especially as they play an increasingly important role in the National Health Service (NHS) in the UK. It argues that mHealth technologies should instead be framed as digital companions. This shift from empowerment to companionship is advocated by showing the conceptual, ethical, and methodological issues challenging the narrative of empowerment, and by arguing that such challenges, as well as the risk (...) of medical paternalism, can be overcome by focusing on the potential for mHealth tools to mediate the relationship between recipients of clinical advice and givers of clinical advice, in ways that allow for contextual flexibility in the balance between patiency and agency. The article concludes by stressing that reframing the narrative cannot be the only means for avoiding harm caused to the NHS as a healthcare system by the introduction of mHealth tools. Future discussion will be needed on the overarching role of responsible design. (shrink)

Contextualists such as Cohen and DeRose claim that the truth conditions of knowledge attributions vary contextually, in particular that the strength of epistemic position required for one to be truly ascribed knowledge depends on features of the attributor's context. Contextualists support their view by appeal to our intuitions about when it's correct (or incorrect) to ascribe knowledge. Someone might argue that some of these intuitions merely reflect when it is conversationally appropriate to ascribe knowledge, not when knowledge is truly ascribed, (...) and so try to accommodate these intuitions even on an invariantist view. DeRose (Blackwell Guide to Epistemology, 1998; Philosophical Review, 2002) argues that any such 'warranted assertibility manoeuvre', or 'WAM', against contextualism is unlikely to succeed. Here, I argue that his objections to a WAM against contextualism are not persuasive and offer a pragmatic account of the data about ascriptions of knowledge. (shrink)

As the range of potential uses for Artificial Intelligence (AI), in particular machine learning (ML), has increased, so has awareness of the associated ethical issues. This increased awareness has led to the realisation that existing legislation and regulation provides insufficient protection to individuals, groups, society, and the environment from AI harms. In response to this realisation, there has been a proliferation of principle-based ethics codes, guidelines and frameworks. However, it has become increasingly clear that a significant gap exists between the (...) theory of AI ethics principles and the practical design of AI systems. In previous work , we analysed whether it is possible to close this gap between the ‘what’ and the ‘how’ of AI ethics through the use of tools and methods designed to help AI developers, engineers, and designers translate principles into practice. We concluded that this method of closure is currently ineffective as almost all existing translational tools and methods are either too flexible (and thus vulnerable to ethics washing) or too strict (unresponsive to context). This raised the question: if, even with technical guidance, AI ethics is challenging to embed in the process of algorithmic design, is the entire pro-ethical design endeavour rendered futile? And, if no, then how can AI ethics be made useful for AI practitioners? This is the question we seek to address here by exploring why principles and technical translational tools are still needed even if they are limited, and how these limitations can be potentially overcome by providing theoretical grounding of a concept that has been termed ‘Ethics as a Service’. (shrink)

Aristotle repeatedly claims that character-virtue “makes the goal right“, while Phronesis is responsible for working out how to achieve the goal. Many argue that these claims are misleading: it must be intellect that tells us what ends to pursue. I argue that Aristotle means just what he seems to say: despite putative textual evidence to the contrary, virtue is (a) a wholly non-intellectual state, and (b) responsible for literally supplying the contents of our goals. Furthermore, there are no good textual (...) or philosophical reasons to reject this straightforward interpretation. Contrary to widespread opinion, Aristotle does not characterize Phronesis as supplying ends. Instead, its ethical import lies wholly in its ability to “determine the mean“. Moreover, because character involves non-rational cognition of the end as good, Aristotle can restrict practical intellect to deliberation without abandoning his anti-Humean view that we desire our ends because we find them good. (shrink)

An essentialist theory of modality claims that the source of possibility and necessity lies in essence, where essence is then not to be defined in terms of necessity. Hence such theories owe us an account of why it is that the essences of things give rise to necessities in the way required. A new approach to understanding essence in terms of the notion of generalized identity promises to answer this challenge by appeal to the necessity of identity. I explore the (...) prospects for this approach, and argue that it fails. If one favours an account of essence in terms of generalized identity, then one will not, I argue, be able to satisfactorily defend an essentialist theory of modality against the challenge; if one wishes to defend an essentialist theory of modality, and thereby to give an explanation of how necessity arises from essence, one should not understand essence in terms of generalized identity. (shrink)

When subjects violate epistemic standards or norms, we sometimes judge them blameworthy rather than blameless. For instance, we might judge a subject blameworthy for dogmatically continuing to believe a claim even after receiving evidence which undermines it. Indeed, the idea that one may be blameworthy for belief is appealed to throughout the contemporary epistemic literature. In some cases, a subject seems blameworthy for believing as she does even though it seems prima facie implausible that she is morally blameworthy or professionally (...) blameworthy. Such cases raise the question of whether one can be blameworthy for a belief in a specifically epistemic sense rather than in some already recognised sense, such as being morally or professionally blameworthy. A number of authors have recently argued that there is a moral or social sense in which one ought to conform one’s beliefs to the evidence. In this paper, I argue that even while accepting that there are moral and social norms governing belief, there are cases in which a subject is blameworthy for a belief but isn’t plausibly morally or socially blameworthy. If this latter view is correct, then we may need to develop a new account of blame which can be applied to beliefs which are not morally or socially blameworthy. (shrink)

Knowledge ascriptions are a central topic of research in both philosophy and science. In this collection of new essays on knowledge ascriptions, world class philosophers offer novel approaches to this long standing topic.

It has become recently popular to suggest that knowledge is the epistemic norm of practical reasoning and that this provides an important constraint on the correct account of knowledge, one which favours subject-sensitive invariantism over contextualism and classic invariantism. I argue that there are putative counterexamples to both directions of the knowledge norm. Even if the knowledge norm can be defended against these counterexamples, I argue that it is a delicate issue whether it is true, one which relies on fine (...) distinctions among a variety of relevant notions of propriety which our intuitions may reflect. These notions variously apply to the agent herself, her character traits, her beliefs, her reasoning and any resultant action. Given the delicacy of these issues, I argue that the knowledge norm is not a fixed point from which to defend substantive and controversial views in epistemology. Rather, these views need to be defended on other grounds. (shrink)

The World Health Organisation declared COVID-19 a global pandemic on 11th March 2020, recognising that the underlying SARS-CoV-2 has caused the greatest global crisis since World War II. In this chapter, we present a framework to evaluate whether and to what extent the use of digital systems that track and/or trace potentially infected individuals is not only legal but also ethical.

An approach to explaining the nature and source of logic and its laws with a rich historical tradition takes the laws of logic to be laws of thought. This view seems intuitively compelling, after all, logic seems to be intimately related with how we think. But how exactly should we understand it? And what arguments can we give in favour? I will propose one line of argument for the claim that the laws of logic are laws of thought. I will (...) motivate the claim that there is a certain phenomenon, namely, that there are logical principles which are immune to rational doubt. I will then give an argument to the best explanation; I will argue that the best explanation of this phenomenon is to take the laws of logic to be constitutive-normative laws of thought. The proposal, and some responses to potential objections, will have a notably Kantian flavour. (shrink)

Stanley and Williamson have defended the intellectualist thesis that knowing-how is a subspecies of knowing-that by appeal to the syntax and semantics of ascriptions of knowing-how. Critics have objected that this way of defending intellectualism places undue weight on linguistic considerations and fails to give sufficient attention to empirical considerations from the scientific study of the mind. In this paper, I examine and reject Stanley's recent attempt to answer the critics.

In this paper, I defend McKinsey's argument (Analysis 1991) that Burge's antiindividualist position is incompatible with privileged access, viz. the claim that each subject can know his own thought contents just by reflection and without having undertaken an empirical investigation. I argue that Burge thinks that there are certain necessary conditions for a subject to have thoughts involving certain sorts of concepts; these conditions are appropriately different for thoughts involving natural kind concepts and thoughts involving non-natural kind concepts. I use (...) Burge's commitment to these entailments to show that his antiindividualist position is incompatible with privileged access. (shrink)

Research groups around the world are currently busy trying to invent new life in the laboratory, looking for extraterrestrial life, or making machines increasingly more life-like. In the case of astrobiology, any newly discovered life would likely be very old, but when discovered it would be new to us. In the case of synthetic organic life or life-like machines, humans will have invented life that did not exist before. Together, these endeavors amount to what we call the emerging plurality of (...) lives because we see a future where we are surrounded by, and interact with, life of many different origins, as well as entities that we have not traditionally seen as being alive. In the future these entities will likely possess properties that we so far have only associated with living beings, thereby transcending and perhaps blurring the borders between life and non-life. (shrink)

According to Ian Hacking, some human kinds are subject to a peculiar type of classificatory instability: individuals change in reaction to being classified, which in turn leads to a revision of our understanding of the kind. Hacking’s claim that these ‘human interactive kinds’ cannot be natural kinds has been vehemently criticized on the grounds that similar patterns of instability occur in paradigmatic examples of natural kinds. I argue that the dialectic of the extant debate misses the core conceptual problem of (...) human interactive kinds. The problem is not that these kinds are particularly unstable but ‘capricious’—their members behave in wayward, unexpected manners that defeats existing theoretical understanding. The reason for that, I argue, is that human interactive kinds are often ‘hybrid kinds’ consisting of a base kind and an associated status, which makes mechanisms that support patterns of change and stability systematically difficult to understand and predict. 1Introduction2The Extant Discussion 2.1Hacking’s account of interactive kinds2.2Classificatory feedback in non-human kinds3Natural kinds and Ontological Instability 3.1Understanding instability3.2The problem of stabilizing feedback3.3Summary4Capricious Kinds 4.1Biased conceptualization4.2Studying social status5Conclusion. (shrink)

I consider but reject one broad strategy for answering the threshold problem for fallibilist accounts of knowledge, namely what fixes the degree of probability required for one to know? According to the impurist strategy to be considered, the required degree of probability is fixed by one's practical reasoning situation. I distinguish two different ways to implement the suggested impurist strategy. According to the Relevance Approach, the threshold for a subject to know a proposition at a time is determined by the (...) practical reasoning situations she is then in to which that particular proposition is relevant. According to the Unity Approach, the threshold for a subject to know any proposition whatsoever at a time is determined by a privileged practical reasoning situation she then faces, most plausibly the highest stakes practical reasoning situation she is then in. I argue that neither way of implementing the impurist strategy succeeds and so impurism does not offer a satisfactory response to the threshold problem. (shrink)

We introduce the concept of assertion, survey existing views about it, and detail the contents of the remainder of the book.

Plato links pleasure with illusion, and this link explains his rejection of the view that all desires are rational desires for the good. The Protagoras and Gorgias show connections between pleasure and illusion; the Republic develops these into a psychological theory. One part of the soul is not only prone to illusions, but also incapable of the kind of reasoning that can dispel them. Pleasure appears good; therefore this part of the soul (the appetitive part) desires pleasures qua good but (...) ignores reasoning about what is really good. Hence the new moral psychology of the Republic: not all desires are rational, and thus virtue depends on bringing one's non-rational desires under the control of reason. (shrink)

If there is any consensus about knowledge in contemporary epistemology, it is that there is one primary kind: knowledge-that. I put forth a view, one I find in the works of Aristotle, on which knowledge-of – construed in a fairly demanding sense, as being well-acquainted with things – is the primary, fundamental kind of knowledge. As to knowledge-that, it is not distinct from knowledge-of, let alone more fundamental, but instead a species of it. To know that such-and-such, just like to (...) know a person or place, is to be well-acquainted with a portion of reality – in this case a fact. In part by comparing classic Gettier cases to cases in which one has true impressions of but fails to know a person, I argue that this account not only respects our intuitions about knowledge-that – in particular that it is or entails non-accidentally true justified belief – but also explains them, providing a compelling analysis. (shrink)

Vetter's Potentiality is an exposition and development of a new account of possibility and necessity, given in terms of potentialities. In this critical notice, I give an outline of some of the key claims of the book. I then raise some issues for the extent to which Vetter's view can accommodate genuine de re modalities, especially those of possible existence and non-existence.

Recently, a debate has developed between those who claim that essence can be explained in terms of de re modality (modalists), and those who claim that de re modality can be explained in terms of essence (essentialists). The aim of this paper is to suggest that we should reassess. It is assumed that either necessity is to be accounted for in terms of essence, or that essence is to be accounted for in terms of necessity. I will argue that we (...) should assume neither. I discuss what role these key notions – essence and necessity – can reasonably be thought to contribute to our understanding of the world, and argue that, given these roles, there is no good reason to think that we should give an account of one in terms of the other. I conclude: if we can adequately explain de re modality and essence at all, we should aim to do so separately. (shrink)

I consider but reject one broad strategy for answering the threshold problem for fallibilist accounts of knowledge, namely what fixes the degree of probability required for one to know? According to the impurist strategy to be considered, the required degree of probability is fixed by one's practical reasoning situation. I distinguish two different ways to implement the suggested impurist strategy. According to the Relevance Approach, the threshold for a subject to know a proposition at a time is determined by the (...) practical reasoning situations she is then in to which that particular proposition is relevant. According to the Unity Approach, the threshold for a subject to know any proposition whatsoever at a time is determined by a privileged practical reasoning situation she then faces, most plausibly the highest stakes practical reasoning situation she is then in. I argue that neither way of implementing the impurist strategy succeeds and so impurism does not offer a satisfactory response to the threshold problem. (shrink)

It has been suggested that to overcome the challenges facing the UK’s National Health Service (NHS) of an ageing population and reduced available funding, the NHS should be transformed into a more informationally mature and heterogeneous organisation, reliant on data-based and algorithmically-driven interactions between human, artificial, and hybrid (semi-artificial) agents. This transformation process would offer significant benefit to patients, clinicians, and the overall system, but it would also rely on a fundamental transformation of the healthcare system in a way that (...) poses significant governance challenges. In this article, we argue that a fruitful way to overcome these challenges is by adopting a pro-ethical approach to design that analyses the system as a whole, keeps society-in-the-loop throughout the process, and distributes responsibility evenly across all nodes in the system. (shrink)

de Anima III.10 characterizes akrasia as a conflict between phantasia (“imagination”) on one side and rational cognition on the other: the akratic agent is torn between an appetite for what appears good to her phantasia and a rational desire for what her intellect believes good. This entails that akrasia is parallel to certain cases of perceptual illusion. Drawing on Aristotle's discussion of such cases in the de Anima and de Insomniis , I use this parallel to illuminate the difficult discussion (...) of akrasia in Nicomachean Ethics VII.3, arguing that its account of akrasia as involving ignorance is compatible with, and in fact crucially supplements, the more straightforward account we find elsewhere in the corpus of akrasia as a struggle between desires. (shrink)

Healthcare systems across the globe are struggling with increasing costs and worsening outcomes. This presents those responsible for overseeing healthcare with a challenge. Increasingly, policymakers, politicians, clinical entrepreneurs and computer and data scientists argue that a key part of the solution will be ‘Artificial Intelligence’ (AI) – particularly Machine Learning (ML). This argument stems not from the belief that all healthcare needs will soon be taken care of by “robot doctors.” Instead, it is an argument that rests on the classic (...) counterfactual definition of AI as an umbrella term for a range of techniques that can be used to make machines complete tasks in a way that would be considered intelligent were they to be completed by a human. Automation of this nature could offer great opportunities for the improvement of healthcare services and ultimately patients’ health by significantly improving human clinical capabilities in diagnosis, drug discovery, epidemiology, personalised medicine, and operational efficiency. However, if these AI solutions are to be embedded in clinical practice, then at least three issues need to be considered: the technical possibilities and limitations; the ethical, regulatory and legal framework; and the governance framework. In this article, we report on the results of a systematic analysis designed to provide a clear overview of the second of these elements: the ethical, regulatory and legal framework. We find that ethical issues arise at six levels of abstraction (individual, interpersonal, group, institutional, sectoral, and societal) and can be categorised as epistemic, normative, or overarching. We conclude by stressing how important it is that the ethical challenges raised by implementing AI in healthcare settings are tackled proactively rather than reactively and map the key considerations for policymakers to each of the ethical concerns highlighted. (shrink)

In this paper I investigate what it is for a group to believe something for a reason. I defend a non-summative account on which a group can believe that p for a reason even though none of its members believe that p for that reason. By contrast, a summative account would hold that the reason for which a group believes that p is a function of the reason for which its members believe that p. I argue that the proposed non-summative (...) account deals better with cases in which members of a group believe that p for different reasons. I also defend it against a range of objections, including that it conflicts with epistemic norms for assertion and action. (shrink)

The main contribution of this paper is a new account of how a community may introduce a term for a natural kind in advance of knowing the correct scientific account of that kind. The account is motivated by the inadequacy of the currently dominant accounts of how a community may do this, namely those proposed by Kripke and by Putman. Their accounts fail to deal satisfactorily with the facts that (1) typically, an item that instantiates one natural kind instantiates several (...) - 'the higher-level natural kinds problem', and (2) natural kinds often occur in nature in impure form - 'the composition problem' .On the account I propose, a term for a natural kind gains its reference by being associated with a recognitional capacity for that kind. I show how members of a scientifically ignorant community could have a recognitional capacity for a natural kind, say gold, as opposed to a certain kind of appearance, for instance the appearance that gold actually has. I argue that members of such a community can have recognitional capacities for particular natural kinds despite the actual or possible existence of duplicate kinds, e.g. water. After developing the account in detail, I show how it can deal with the two problems faced by Kripke's and Putnam's problem. The case of natural kind terms is crucial to the central debate in the philosophy of language and the philosophy of mind about whether we can refer non-descriptively to objects and kinds in the world. I take the account I propose to be a non-descriptive account of linguistic reference to natural kinds that can be used to support externalism in the philosophy of mind. (shrink)

I will ask the conditional question: if folk attributions of "know" are not sensitive to the stakes and/or the salience of error, does this cast doubt on contextualism or subject-sensitive invariantism (SSI)? I argue that if it should turn out that folk attributions of knowledge are insensitive to such factors, then this undermines contextualism, but not SSI. That is not to say that SSI is invulnerable to empirical work of any kind. Rather, I defend the more modest claim that leading (...) versions of SSI are not undermined by one particular kind of experimental result, namely the recent suggestion that knowledge attributions are insensitive to the stakes. (shrink)

The idea that one can blamelessly violate a norm is central to ethics and epistemology. The paper examines the prospects for an account of blameless norm violation applicable both to norms governing action and norms governing belief. In doing so, I remain neutral on just what are the norms governing action and belief. I examine three leading suggestions for understanding blameless violation of a norm which is not overridden by another norm: doxastic accounts; epistemic accounts; and appeal to expected value. (...) We see that all of these accounts face problems when understood as accounts of blameless norm violation applicable to both belief and action. This leaves a variety of options including seeking an alternative account of blameless norm violation common to belief and action; concluding that we cannot determine the correct account of blameless norm violation independently of what are the norms of belief; and abandoning the project of finding a common account of blameless norm violation common to ethics and epistemology. (shrink)

The aim of this chapter is to explore to some extent the relationship between identity and necessity in logic and metaphysics. First, I provide a historically-based summary of proofs of the necessity of identity, highlighting the importance of the role that self-identity plays. Second, I introduce two examples of metaphysical topics where the necessity of identity has played a pivotal role: the necessary a posteriori, and the coincidence of material objects. I argue that important aspects of these debates rest on (...) how we represent identity. Third, I consider some recent work on generalized identity. This opens up new prospects for explaining why identity is necessary. (shrink)

In her 1938 paper ‘Logical and Metaphysical Necessity’, Martha Kneale introduces the necessary a posteriori. I present a critical summary of Kneale's argument that so-called ‘metaphysical propositions’ are necessary but not a priori. I argue that Kneale is well placed to offer a template for reconciling conceivability approaches to modal epistemology with the post-Kripkean trend for taking metaphysical necessities to have their source in mind-independent reality.

According to one contemporary formulation of infallibilism, probability 1 infallibilism, if a subject knows that p, then the probability of p on her evidence is 1. To avoid an implausible scepticism about knowledge, probability 1 infallibilism needs to allow that, in a wide range of cases, a proposition can be evidence for itself. However, such infallibilism needs to explain why it is typically infelicitous to cite p as evidence for p itself. I argue that probability 1 infallibilism has no explanation (...) of this infelicity and should be rejected. (shrink)