Abstract It is often argued that maximizing normative theories are inappropriate for bounded agents. I argue that leading motivations for theories of bounded rationality are substantially compatible with some traditional maximizing normative theories. This finding expands the audience for theories of bounded rationality by making room for advocates of maximizing normative theories, and has implications for knowledge-action principles, the relationship between descriptive and normative theorizing, and the definitions of central tenets of bounded rationality.
Abstract Standard accounts of higher-order defeat run into well-known Bayesian hurdles. In this paper, I explore the prospects of overcoming those hurdles by applying a thoroughgoing instrumentalist approach: treating our reasoning capacities as we’d treat any other instrument. I argue that instrumentalism cannot coherently accommodate the defeatist’s position when that position involves genuine higher-order uncertainty—that is, rational uncertainty about what rationality requires. At the end of the paper I gesture at an approach to modeling our responses to defeaters that is both instrumentalist and Bayesian, but that abandons the thought that the cases typically described as cases of higher-order defeat involve genuinely higher-order uncertainty.
Abstract Two central questions regarding future bias are (a) why we are future-biased, and (b) whether future bias is justified. With respect to (a), I propose a Humean explanation of future bias in terms of the workings of the imagination, which makes our imaginings of future events more vivid than our imaginings of past events. However, in order to defend this explanation, a potential tension within Humean psychology needs to be resolved: since, per Hume, memory is more vivid than imagination, we ought to be past-biased. My proposed resolution makes use of Hume’s mechanism of sympathy (when applied to one’s own future self). With respect to (b), I argue that, rather than focusing on the question of rationality, we should focus on whether future bias is a virtue, by considering, in a Humean spirit, how agents’ temporal preferences affect others, when assessed from ‘the general point of view’.
Abstract People sometimes experience shame in response to being looked down upon, despite rejecting the negative attitudes taken towards them by others. I aim to explain why—while also holding onto the idea that shame is a response to an apparent flaw in oneself. At the heart of my proposal is the idea that many people treat being well-regarded by specific communities of others as a personal ideal. I defend this account against an objection from Calhoun, namely that it conflicts with the claim that shame in response to oppression is generally rational. On the contrary, I argue that we can only fully make sense of what is bad about such shame by abandoning this claim. Finally, I contrast my view with an alternative based on the ideal of ‘self-presentation’. I object that this fails to emphasize the importance of the negative valence of being looked down upon.
Epistemic charity says that we should (or may) interpret others in the most favourable way consistent with the evidence. Though this could be refined in several ways, this paper develops an argument against all of them, insofar as they endorse a certain 'signature' of charity, where the favourable interpretation is combined with an acknowledgement of, and uncertainty over, countervailing evidence. Though initially plausible, the signature is susceptible to a version of triviality arguments prominent in the literature on conditionals. The second half of the paper evaluates responses to the argument and generalizations of it to other default epistemic rules. Ultimately, a resolution is suggested, requiring that the friend of epistemic charity combine their optimism with realism in a different way, via fragmented states of mind.
Debates about animal sentience often appeal to precaution: where the evidence of sentience is strong enough, taxa should be treated as sentient in law and policy. I argue that what is called 'evidence of sentience' in fact has two dimensions-robustness of indicators and credibility of the indicator-sentience link-and that standard, threshold-based formulations of precaution ignore this structure. Using a simple model, I show that conjunctive high-threshold rules are over-conservative and excessively fragile, and I sketch an inverse-threshold-with-floors principle better suited to sentience policy.
This paper investigates the concept of epistemic community. Despite its frequent usage, the term lacks a clear and unified characterization. I argue that epistemic communities are essential to understanding how individuals form justified beliefs in contexts of epistemic dependence. Through an extensive review of prominent theorists, I demonstrate that epistemic communities are often used to ground trust, justify testimony, and structure the distribution of epistemic resources and norms. I propose a functional definition: an epistemic community is a group that shares epistemic norms and resources through networks of epistemic dependence. These communities range from local, tight-knit groups to diffuse, societal structures, and can confer either epistemic value or distortion. I show that epistemic communities are not just descriptive units but normative forces in belief formation.
What kind of mind is necessary and sufficient for moral status-for having interests that matter morally in and of themselves? This paper defends Affective Sentientism, the view that moral status requires the capacity for affective experiences such as pleasure, pain, and emotion. I argue that affective consciousness is what makes an entity a welfare subject, and that all and only welfare subjects have moral status. The connection between affect and welfare is systematic: I show that every major candidate welfare good has an affective component, often because such goods require the capacity to care. I also argue that even if the properties of being phenomenally conscious, being autonomous, or having a neutral level of welfare could come apart from affective consciousness, none of them could ground moral status in its absence. I conclude by exploring the implications for the moral status of animals and artificial intelligence systems.
In the face of rapidly improving AI capabilities, AI theorists have warned about the risks of ceding power to autonomous artificial agents. In particular, going sufficiently far down this path might mean 'losing control of our future': an outcome some have ranked as a worst-case catastrophe nearly on par with human extinction. A natural way to motivate this judgment is via the thought that losing human control might lead to our domination by future AI systems. Republican political philosophers such as Pettit have held that domination characteristically involves important kinds of harms, even when the dominating party is benevolent and non-interfering. Is this thought correct? After developing an account of what it means to control humanity's future, I show that AI systems possessing such control would indeed be likely to dominate humans on our best theories of arbitrary power. However, contrary to republican orthodoxy, I argue that this need not lead to two of the three primary harms traditionally associated with domination: the psychological and social costs of subordination, and the burden of obligatory strategic behavior. Meanwhile, the third harm-pervasive uncertainty about interference-remains a serious concern even in otherwise human-friendly AI power scenarios. These results should be of interest to theoretical political philosophy, as they suggest that artificial power has unique features which are as yet underrepresented in discussions of domination.
Synergy effects are taken to occur if equally well-informed and competent people agree, or nearly agree, that a certain proposition is true (or that it is false); these people should then take their agreement as confirmation that their reasoning has been correct and adopt an even greater (or lower) credence towards the issue under debate. Thus understood, synergy effects appear to occur quite frequently. The paper argues that a necessary condition for synergy effects to occur is that we cannot take our evidence at face value. This might happen either because we should be epistemically cautious in cases of presumed asymmetrical proneness to error (cases in which the relevant evidence supports either credence 1 or credence 0 towards the proposition at issue are the prime examples hereof); or because background information rules out the option supported by our remaining evidence. This result preserves the intuitive appeal of considerations on synergy but significantly restricts their scope. It thereby also reconciles considerations on synergy with a natural interpretation of conciliatory views on peer disagreement.
Frankfurt's well-known analysis is useful for understanding bullshit assertions in spoken and written language. But the term "bullshit" is routinely applied to a wider range of phenomena. Paper work is bullshit. Textbook prices are bullshit. Airport security lines are bullshit. Even within the realm of language, Frankfurt's analysis applies straightforwardly only to utterances that purport to aim at the truth. Yet, bullshit is possible not only in speech acts other than assertion, but also in conversational contexts where truth-even for assertions-is not the goal. I argue that Frankfurt's essay contains the resources for a generalized theory of bullshit that encompasses the full range of phenomena, and of which the classic Frankfurtian analysis is a special instance. I contrast the account with other adaptations of Frankfurt's approach. I conclude by exploring the important role that the concept of bullshit can play in debating and evaluating our practices more generally.
Social philosophers have long been interested in the connections between the social world and our ways of thinking or talking about it. Some suggest a tight connection between our language and the sorts of social kinds we encounter. The thought is that labels, that is, linguistically expressed concepts, sometimes confirm their own application: that labelling someone or something as a K, for some kind concept K, makes that person or thing a K. I argue that generally, being labelled as a K is neither necessary nor sufficient for being a K. Then, I explain that familiar social critiques don't require the thought that label-application causes or constitutes kind membership.
Power theorists are divided among those who think that a dispositionalist metaphysics imbues the physical world with a dynamic and active aspect, and those who deny that. So far there has been little success in clarifying the exact nature of the two positions and their disagreement beyond trading metaphors. In this paper we suggest that one way to elucidate the idea that powers are dynamic and active is to think that a robust conception of change must play a central role in power-based explanations. We use the grounding of modality as an example. We claim that friends of dynamic powers should embrace a specific version of Dispositionalism, one in which the modal facts change over time. We then argue that this view entails an especially strong version of Necessitism and Permanentism. In turn, according to an influential view in the philosophy of time, this amounts to adopting a Moving Spotlight Theory of time. We provide some details of this power-based moving spotlight theory, its features, and potential flaws.
I argue that agents can be praiseworthy for doing the right thing despite their being unaware of their doing that thing (and not only that it is right). Differently put, I propose that negligence has a positive counterpart, which I dub "posigence." To motivate this proposal, I introduce cases in which an agent has sufficient reason to act in a way that promotes her own well-being but sets back (or at least, fails to promote) the interests of another, but inadvertently omits to do so, manifesting therein good quality of will.
I argue that our intrinsic epistemic motivation extends beyond curiosity to include interest-a distinct affect through which we evaluate objects as worthy of exploration through sustained attention. While an agent's curiosity is directed at certain questions or specific unknowns, and aims at acquiring the relevant knowledge through goal-directed inquiry, an agent's interest is an attitude about objects and facts which she aims to get acquainted with through sustained attention. Drawing on valuationist theories of cognition, I argue that interest serves as the phenomenological manifestation of our appraisal of stimuli as worthy of attention, playing a crucial role in proper epistemic attunement, and explaining some forms of improper attunement. While fitting curiosity might support virtuous inquiry, fitting interest might support virtuous attending.
I here critically discuss "the Selection Problem", a vexing problem for metaphysicians who are realists about fictional entities and who further claim that such entities are identified with nonactualia. The Selection Problem arises due to the fact that the nonactualist is forced to defend the view that ficta are somehow discovered, or selected, by authors who then present them to us so we may entertain their exploits. But how exactly does an author manage this when there are a vast number of eligible candidates to be found in the relevant nonactual realms? I argue that the realist who is also a nonactualist would do well to adopt an ontology on which ficta are metaphysically vague in the sense that they are not wholly found at any particular nonactual realm, but are at once spread out over them all.
