James Hernandez McIntyre

External world skepticism is often thought to entail that we should suspend judgment about the existence of the external world. I challenge this orthodoxy by arguing that, when combined with a plausible principle of ontological parsimony, the skeptical challenge intensifies into an argument for outright disbelief in the external world. The principle in question instructs us to prima facie disbelieve in undetectable postulates. This bears on the skeptical challenge because, by the skeptic’s lights, the external world and its constituents qualify as undetectable. This result poses problems for concessive responses to skepticism, such as entitlement-based and pragmatic strategies.

This paper argues that if an artificial system realizes consciousness, then it’s likely to realize not just one, but many independent minds at once. This view is motivated by analogy to split-brain cases, in which the two hemispheres of a patient’s brain are functionally disconnected. While there are several ways to interpret this phenomenon, it’s highly plausible that split-brain patients have two distinct minds—one per hemisphere. If functional disconnection within a single biological brain can generate multiple minds, then we should say the same about artificial conscious systems, where such disconnections are likely to be even more extreme. Thus, each instance of an overarching artificial conscious system (e.g., a particular user’s distinct, functionally independent interaction with that system) will realize an independent mind, leading to a radical multiplicity of artificial minds. This carries several ethical implications for AI design and system updating. Moreover, it introduces the prospect that human interests may soon find themselves overwhelmed in the moral calculus by an enormous number of artificial minds.

Philosophers often defend appeals to parsimony by invoking its central role in science. I argue that this move fails once we distinguish between two uses of parsimony: non-ideal and ideal. Non-ideal parsimony enjoys strong inductive support in science, since complex models are prone to overfit to predictively irrelevant noise. But philosophical data aren’t significantly noisy in the relevant sense: when our intuitions are unreliable, their unreliability typically reflects systematic bias rather than noise, which parsimony doesn’t mitigate. Philosophers therefore need ideal parsimony, which finds only weak support from science. Thus, the scientific analogy cannot vindicate the philosopher’s use of parsimony.

This paper argues that two of the theoretical virtues—ontological and ideological parsimony—reduce to a more foundational and familiar theoretical imperative to avoid positing brute facts. Following recent developments in metaphysics, I characterize ontological and ideological parsimony as principles that impose theoretical costs on fundamental ontology and primitive ideology respectively. I subsequently argue that both sorts of theoretical commitment entail costly brute (i.e., ungrounded) commitments. We can thus understand ontological and ideological parsimony as reducing to brute fact parsimony. After defending this proposal against objections, I extend it to the non-fundamental by introducing the notion of domain-relative bruteness, thereby explaining parsimony’s role in the special sciences. If my argument is successful, then it is now possible to weigh previously incommensurable virtues against each other. I suggest that this may facilitate progress in the debate about the laws of nature.

Generative AI systems require vast amounts of training data, much of it scraped from the internet without creators’ consent. Critics often characterize this practice as “theft,” but such claims require showing that AI training violates creators’ property rights. However, this position risks implying that human learning and works of inspiration would count as theft, too. This paper develops two arguments for moral restrictions that apply uniquely to AI training, not to human learning. The first contends that even if creators possess extensive moral property rights over the use of their content, in the human case, these property rights are typically overridden by a countervailing right to freedom of thought. Since AI systems lack such a right, creators’ property rights remain intact, requiring AI companies to obtain permission to use their work. The second argument contends that even without such broad property rights, creators have the moral right to restrict which copying technologies may be used to capture their work, allowing them to block web crawlers for AI training while permitting ordinary human browsing. The paper concludes by exploring the policy implications of these moral arguments.

Misinformation is paradigmatically false. However, a great deal of contemporary political content misleads by invoking defective concepts, even when the content is true. This paper offers an account of conceptual misinformation, according to which the use of a concept constitutes misinformation when that concept both fails to adequately capture the world’s structure and receives uptake. Central to my account is the metaphysical notion of “joint-carving” concepts—those that capture the world’s structure and figure into good explanations. This notion can be extended to the social and normative domains, where socially and normatively joint-carving concepts (e.g., “farmer,” “legal free speech”) yield better social and normative explanations than gerrymandered concepts (e.g., “Pacific Islander,” “consequence-free speech”). The need to avoid conceptual misinformation carries practical consequences and imposes a constraint on ameliorative conceptual engineering projects.