PHILOSOPHICAL STUDIES最新文献

Large language models now possess human-level linguistic abilities in many contexts. This raises the concern that they can be used to deceive and manipulate on unprecedented scales, for instance spreading political misinformation on social media. In future, agentic AI systems might also deceive and manipulate humans for their own purposes. In this paper, first, I argue that AI-generated content should be subject to stricter standards against deception and manipulation than we ordinarily apply to humans. Second, I offer new characterizations of AI deception and manipulation meant to support such standards, according to which a statement is deceptive (resp. manipulative) if it leads human addressees away from the beliefs (resp. choices) they would endorse under “semi-ideal” conditions. Third, I propose two measures to guard against AI deception and manipulation, inspired by this characterization: “extreme transparency” requirements for AI-generated content and “defensive systems” that, among other things, annotate AI-generated statements with contextualizing information. Finally, I consider to what extent these measures can protect against deceptive behavior in future, agentic AI systems.

This paper examines the question as to whether propositional seeing is best thought of as a way of knowing a proposition to be true. After showing how Pritchard’s distinction between objective and subjective goodness motivates a negative answer to this question, I examine a challenge raised by Ghijsen for Pritchard’s construal of that distinction. I then turn to the connection between propositional seeing and belief. I argue that doxasticism about propositional seeing – the claim that propositional seeing involves belief – ultimately lacks independent motivation and I offer a model of propositional seeing that explains how propositional perception can provide one with a rational basis for forming a perceptual belief. Finally, I discuss in what way the proposed model of propositional seeing may remain compatible with the claim that propositional seeing is a way of knowing a proposition to be true.

Much of revisionist just war theory is individualistic in nature: morality in war is just an extension of morality in interpersonal circumstances, so that killing in war is subject to the same moral principles that govern personal self-defense and defense of others. Recent work in the ethics of self-defense suggests that this individualism leads to a puzzle, which I call the puzzle of aggregation, when many threateners contribute to a single threatened harm. In this paper, I investigate the moral problems posed by the puzzle of aggregation and develop a novel account recently proposed by Jeff McMahan, which he calls “proportionality in the aggregate”. I argue that accepting proportionality in the aggregate as a moral constraint on the use of force has significant implications for the ethics of war and self-defense.

This paper defends a novel version of moderate pessimism about moral deference, i.e., the view that we have pro tanto reason to try to avoid moral deference. The problem with moral deference is that it puts one in a bad position to form what I call morally worthy attitudes, i.e., non-cognitive attitudes that have moral worth in the same sense that certain actions have moral worth. Forming morally worthy attitudes requires a special sensitivity to the sufficiency of the moral reasons one has to form these attitudes. However, when one defers about a moral claim, one takes it fully on trust, which entails that one is strongly disposed to not intentionally try to take precautions against believing that moral claim epistemically defectively (e.g., falsely or without sufficient justification). Thus, moral deference makes one strongly disposed to not try to acquire the sensitivity required for forming morally worthy attitudes. However, morally worthy attitudes are more valuable than merely fitting attitudes and plausibly have both final and non-final value. Thus, moral deference interferes with gaining something with final and non-final value.

This paper argues that God plausibly has facts of self that function as modifiers of the normative reasons that apply to him. Facts of self are subjective facts like the fact that one has certain commitments, the fact that one has a certain character, the fact that one has a certain practical identity, the fact that one has certain projects. There is a widespread intuition (the normative significance of self) that facts of self influence what an agent’s sufficient reasons are. While this intuition is widespread in ethics, its implications for God’s practical life have received little scholarly attention. Facts of God’s self have, however, received some attention in the context of what I call the divine mechanism complaint, but their normative roles have been undertheorized. The divine mechanism complaint is that on certain conceptions of God’s relation to reasons, God is objectionably mechanical. I take this complaint as requiring that God have some influence on what his sufficient reasons are. An adequate account of the normative significance of God’s self, then, can answer the divine mechanism complaint, providing us with a plausible picture of God’s practical life. I provide such an account, arguing that God need not be objectionably mechanical if his facts of self function as modifiers of his normative reasons.

My interest is in an apparent tension between two epistemological theses. The first is veritism, which is roughly the claim that truth is the fundamental epistemic good. The second is the idea that understanding is the proper goal of inquiry. The two theses seem to be in tension because the former seems to imply that the proper goal of inquiry should be truth rather than understanding. And yet there is a strong prima facie case to be made for thinking that properly conducted inquiry aims at an elevated epistemic standing like understanding rather than merely true belief. I suggest that this putative tension is one of the reasons why veritism is these days not widely endorsed. As I show, however, there is in fact no tension between these two claims, at least once they are each properly understood. Indeed, I will be suggesting that there is a plausible conception of veritism which would explain why intellectual exemplars seek out understanding in inquiry.

How can a persisting object change whilst remaining the same object? Lewis, who frames this as the problem of temporary intrinsics, presents us with the perdurance solution: objects persist by having temporal parts which may have differing properties. And in doing so he characterises the opposing view as persisting but not by having temporal parts – a view he calls endurance. But this dichotomous picture of Lewis, although now widely embraced, misses out the orthodox historic view – a view I call temporal holism: objects persist by having temporal parts to which they are ontologically prior. (In the perduring solution, by contrast, the temporal parts are ontologically prior.) This paper sets out this temporal holist solution and makes clear its differences from perdurantist and endurantist solutions. Although temporal holism has a long and illustrious history, this history has not been explicitly recognised. I begin the task of recognising this history in this paper, in order to make clear the nature of temporal holism, and to show that it is a long-established, well supported and distinctive position. The paper sets out, too, how temporal holism solves other ontological problems so that, despite its current neglect, temporary holism has the potential to greatly enrich contemporary philosophical debates.

In the current debate on the relation between how-possibly explanations (HPEs) and understanding, two seemingly irreconcilable positions have emerged, which either deny or assert HPEs’ contribution to understanding. We argue, in contrast, that there is substantial room for reconciliation between these positions. First, we show that a shared assumption is unfounded: HPEs can be interpreted as being correct explanations. Second, we argue that what we call the standard account is actually compatible with the claim that HPEs may improve understanding. Our analysis not only indicates that there is room for reconciliation, but also specifies the potential remaining disagreements.

A jury needs “proof beyond a reasonable doubt” in order to convict a defendant of a crime. The standard is vexingly difficult to pin down, but some legal epistemologists have given this account: knowledge is the standard of legal proof. On this account, a jury should deliver a guilty verdict just in case they know that the defendant is guilty. In this paper, I’ll argue that legal proof requires more than just knowledge that a defendant is guilty. In cases of “merely predictive evidence,” a jury knows that the defendant is guilty but does not have legal proof. What are they missing? Evidence that is causally downstream from the crime. Legal proof requires a “smoking gun.” The point generalizes outside of the courtroom. A professor needs to read a term paper before assigning a grade, even if she knows the student will produce A + work. You may know that your roommate will forget to water the plants while you are away—she is scatterbrained and always forgets these things—but you can’t blame her until you get back home and see that the plants are wilting. In order to have appropriate reactions or reactive attitudes, we must respond causally to what other people have done.

How reliable are causal inferences in complex empirical scenarios? For example, a physician prescribes a drug to a patient, and then the patient undergoes various changes to their symptoms. They then increase their confidence that it is the drug that causes such changes. Are such inferences reliable guides to the causal relation in question, particularly when the physician can gain a large volume of such clinical experience by treating many patients? The evidence-based medicine movement says no, while some physicians and philosophers support such appeals to first-person experience. We develop a formal model and simulate causal inference based on clinical experience. We conclude that in very particular clinical scenarios such inferences can be reliable, while in many other routine clinical scenarios such inferences are not reliable.