Publications
"Using People to Serve Their Own Interests," Journal of Moral Philosophy, November 2024
"Can Relative Prioritarianism Accommodate the Shift?" Ethics, July 2024
"Can Relative Prioritarianism Accommodate the Shift?" Ethics, July 2024
Working Papers
"an ex ante theory of reparations"
Philosophers and legal scholars generally agree that in ordinary cases where a wrongful event occurs, the perpetrator owes their victim compensation. But it is less clear what grounds compensation for historical atrocities, such as the enslavement of Black people, or the seizing of Indigenous land, where both the direct victims and the perpetrators of the atrocity no longer exist. In a working paper, I develop a response to the claim that we often lack the information required to successfully implement a principle of reparations for historical injustices. I argue that we can resolve this epistemic problem for any theory of reparations by being sensitive to the probabilities of individuals possessing relevant characteristics, rather than by being sensitive to whether individuals actually possess those characteristics, in our assessment of claim rights to reparations borne by presently existing people.
"THE wrong of discrimination in LLMs"
I consider what implications the wrong of discrimination has for how AI companies ought – both legally and morally – to train LLMs (AI models trained on large amounts of data). Sophia Moreau recently argued that the wrong of discrimination can be partially explained by the burdens that it places on the target’s deliberative freedom. I argue that this idea can help us make progress on legal and ethical questions regarding LLM training. Anthropic recently published a paper showing that in certain contexts, their AI generates results that demonstrate positive discrimination toward certain historically marginalized social groups. But to what extent should we prioritize eliminating positive discrimination of this kind, especially when doing so may come at the cost of other desirable LLM features? I argue that in a world where any given individual will predictably have a negative bias toward Black candidates but not white candidates (for example), the white candidate will retain substantial deliberative freedom despite an LLM’s favorable bias toward Black candidates. I conclude that there is at least one sense in which positive discrimination toward socially disadvantaged groups in LLMs is not as pressing a moral concern as negative discrimination, and consider the legal implications of this result.
"Do Merely Statistical PeOPLE Have a Standing to Complain?"
Ex ante contractualists evaluate the strength of competing complaints against an action by considering the prospects that action imposes onto each particular person. Ex post contractualists evaluate the strength of competing complaints by considering the outcomes that are expected to result from the associated action. Ex ante contractualists face a serious objection according to which, when faced with a decision between saving a larger number or a smaller number of people, we ought to save the larger number. Ex ante contractualism sometimes fails to yield this result. In reply, some ex ante contractualists have adopted a pluralist view of moral rightness, according to which ex ante contractualism governs the so-called person-directed domain of moral rightness. Some critics believe this move is ad hoc. I argue, however, that there are strong independent grounds for favoring the pluralist view. In particular, only ex ante contractualism (and not other forms of contractualism, such as ex post contractualism) can be said to govern the domain of person-directed obligation as Scanlonian contractualism purports to. If we are sympathetic to Scanlonian contractualism, we must be pluralists.
"Can ex post contractualism accommodate separateness?
In this paper I argue that the ex post contractualist cannot adequately explain the separateness of persons, the moral significance of the boundaries between persons, because ex post justification is not sensitive to the distinctiveness of claim-bearers. This is a surprising result, because it is among contractualism’s greatest selling points that it upholds the separateness of persons whereas utilitarianism and related models do not. I argue that to accommodate separateness, it is not enough to deny that competing claims aggregate, as is commonly claimed. In addition, the contractualist must recognize the moral significance of a single person bearing two possible complaints (say) rather than distinct people issuing the complaints. To do so, they must care about the claims issued by particular people rather than the mere expectation that complaints will arise.
"a solution to the problem of ex ante identification"
I defend ex ante contractualism against a problem raised in the literature for ex ante models of evaluation which ask us to evaluate from the perspective of the agent: What is it to know the claimant ex ante? A person’s ex ante prospects sometimes depend on the guise under which she is regarded by the agent. The ex ante prospects of the person in trouble, for example, might differ from the ex ante prospects of Anne, the person in front of me, if I do not know whether Anne is the person in trouble. I argue that minimal information is required to know the person to whom justification is owed in the relevant way.
"The informational constraints of ex ante contractualism"
According to the ex ante model for evaluating decisions under risk, does the strength of competing claims depend on the agent’s ex ante information, the patient’s ex ante information, the ex ante information of the more knowledgeable party, or an information set otherwise specified? This ambiguity is best brought out by cases in which the patient knows the outcomes of each available action beforehand whereas the agent does not. Should the ex ante model treat such cases as decisions under risk and evaluate on the basis of the ex ante prospects as understood by the agent, or should it regard such cases as non-risky and evaluate on the basis of the prospects as understood by the patient? The agent-centered model is often assumed in the contractualist literature, but it has not to date been carefully defended and compared against alternative interpretations of the ex ante model. I call attention to this concern and offer a contractualist-based explanation for why we should prioritize the agent’s ex ante perspective.