Foundation models such as GPT-4 are fine-tuned to avoid unsafe or otherwise problematic behavior, so that, for example, they refuse to comply with requests for help with committing crimes or with producing racist text. One approach to fine-tuning, called reinforcement learning from human feedback, learns from humans' expressed preferences over multiple outputs. Another approach is constitutional AI, in which the input from humans is a list of high-level principles. But how do we deal with potentially diverging input from humans? How (...) can we aggregate the input into consistent data about "collective" preferences or otherwise use it to make collective choices about model behavior? In this paper, we argue that the field of social choice is well positioned to address these questions, and we discuss ways forward for this agenda, drawing on discussions in a recent workshop on Social Choice for AI Ethics and Safety held in Berkeley, CA, USA in December 2023. (shrink)

Several claims made by Wilkins & Wakefield require qualification. First, the proposed delineation of the parietal-occipital-temporal junction (POT) is overly restrictive. Second, focusing exclusively on the evolutionary importance of manual manipulation oversimplifies interacting evolutionary preconditions for language. Finally, Wilkins and Wakefield's perspective adheres to a homocentric, formal, linguistic definition of language instead of viewing language as a multimodal sensory enhancement system unique to each species.

Is self-disclosure of negative information a viable strategy for a company to lessen the damage done to consumer responses? Three experiments assessed whether self-disclosing negative information in itself lessened the damaging impact of this information compared to third-party disclosure of the same information. Results indicated that mere self-disclosure of a negative event positively affected consumers’ choice behavior, perceived company trustworthiness, and company evaluations compared to third-party disclosure. The effectiveness of the self-disclosure strategy was moderated by the initial reputation of a (...) company, such that its impact was only observed for companies that had a poor reputation at the outset. For them, self-disclosure considerably lessened the impact of negative information compared to third-party disclosure. For companies that enjoyed a positive reputation, type of disclosure did not affect consumer responses. Mediation analysis showed that perceptions of company trustworthiness underlie the effects of the self-disclosure strategy on consumer judgment. (shrink)

Research on the asymmetric effect of negative versus positive affective states on scope of attention, both at a perceptual and a conceptual level, is abundant. However,...

I present a new argument for the repugnant conclusion. The core of the argument is a risky, intrapersonal analogue of the mere addition paradox. The argument is important for three reasons. First, some solutions to Parfit’s original puzzle do not obviously generalize to the intrapersonal puzzle in a plausible way. Second, it raises independently important questions about how to make decisions under uncertainty for the sake of people whose existence might depend on what we do. And, third, it suggests various (...) difficulties for leading views about the value of a person’s life compared to her nonexistence. (shrink)

The standard view of "believes" and other propositional attitude verbs is that such verbs express relations between agents and propositions. A sentence of the form “S believes that p” is true just in case S stands in the belief-relation to the proposition that p; this proposition is the referent of the complement clause "that p." On this view, we would expect the clausal complements of propositional attitude verbs to be freely intersubstitutable with their corresponding proposition descriptions—e.g., "the proposition that p"—as (...) they are in the case of "believes." In many cases, however, intersubstitution of that-clauses and proposition descriptions fails to preserve truth value or even grammaticality. These substitution failures lead some philosophers to reject the standard view of propositional attitude reports. Others conclude that propositional attitude verbs are systematically ambiguous. I reject both these views. On my view, the that-clause complements of propositional attitude verbs denote propositions, but proposition descriptions do not. (shrink)

I defend the view that a reason for someone to do something is just a reason why she ought to do it. This simple view has been thought incompatible with the existence of reasons to do things that we may refrain from doing or even ought not to do. For it is widely assumed that there are reasons why we ought to do something only if we ought to do it. I present several counterexamples to this principle and reject some (...) ways of understanding "ought" so that the principle is compatible with my examples. I conclude with a hypothesis for when and why the principle should be expected to fail. (shrink)

Negative acknowledgement is an impression management technique that uses the admission of an unfavourable quality to mitigate a negative response. Although the technique has been clearly demonstrated, the underlying process is not well understood. The current research identifies a key mediator and moderator while also demonstrating that the effect extends beyond the specific acknowledged domain to the overall evaluation of a target object. The results of study 1 indicate that negative acknowledgement works through mitigating negatively valenced cognitive responses. People who (...) are presented with a negative acknowledgement are less likely to counterargue when forming an evaluation. The results of study 2 reveal that individual differences in need for structure impact the effectiveness of the technique. People who are high in need for structure are more susceptible to the effect presumably because of their desire for easy-to-use information that aids the formation and maintenance of simple knowledge structures. (shrink)

According to asymmetric comparativism, it is worse for a person to exist with a miserable life than not to exist, but it is not better for a person to exist with a happy life than not to exist. My aim in this paper is to explain how asymmetric comparativism could possibly be true. My account of asymmetric comparativism begins with a different asymmetry, regarding the (dis)value of early death. I offer an account of this early death asymmetry, appealing to the (...) idea of conditional goods, and generalize it to explain how asymmetric comparativism could possibly be true. I also address the objection that asymmetric comparativism has unacceptably antinatalist implications. (shrink)

According to the person-affecting restriction, one distribution of welfare can be better than another only if there is someone for whom it is better. Extant problems for the person-affecting restriction involve variable-population cases, such as the nonidentity problem, which are notoriously controversial and difficult to resolve. This paper develops a fixed-population problem for the person-affecting restriction. The problem reveals that, in the presence of incommensurable welfare levels, the person-affecting restriction is incompatible with minimal requirements of impartial beneficence even in fixed-population (...) contexts. (shrink)

Tsui and Weymark (Economic Theory, 1997) have shown that the only continuous social welfare orderings on the whole Euclidean space which satisfy the weak Pareto principle and are invariant to individual-specific similarity transformations of utilities are strongly dictatorial. Their proof relies on functional equation arguments which are quite complex. This note provides a simpler proof of their theorem.

Although considerable attention has been devoted to the concepts of “visible” and “invisible” victims in general medical practice, especially in relation to resource allocation, far less consideration has been devoted to these concepts in behavioral health. Distinctive features of mental health care in the United States help explain this gap. This essay explores three specific ways in which the American mental health care system protects potentially “visible” individuals at the expense of “invisible victims” and otherwise fails to meet the needs (...) of great numbers of people with serious psychiatric conditions: prioritization of the wrong patients, incentivization of excessive caution among providers, and a narrow definition of psychiatry's purview. While each of these practices has been discussed elsewhere in the literature, they are rarely considered as part of an interrelated and systemic problem. Reconceptualizing these three issues as aspects of the larger conflict between the interests of “visible” and “invisible” victims may prove a path toward reform. (shrink)

Matthew Adler's Measuring Social Welfare is an introduction to the social welfare function (SWF) methodology. This essay questions some ideas at the core of the SWF methodology having to do with the relation between the SWF and the measure of well-being. The facts about individual well-being do not single out a particular scale on which well-being must be measured. As with physical quantities, there are multiple scales that can be used to represent the same information about well-being; no one scale (...) is special. Like physical laws, the SWF and its ranking of distributions cannot depend on exactly which of these scales we use. Adler and other theorists in the SWF tradition have used this idea to derive highly restrictive constraints on the shape of the SWF. These constraints rule out seemingly plausible views about distributive justice and population ethics. I argue, however, that these constraints stem from a simple but instructive mistake. The SWF should not be applied to vectors of numbers such as 1 and 2, but rather to vectors of dimensioned quantities such as 1 util and 2 utils. This seemingly pedantic suggestion turns out to have far-reaching consequences. Unlike the orthodox SWF approach, treating welfare levels as dimensioned quantities lets us distinguish between real changes in well-being and mere changes in the unit of measurement. It does this without making the SWF depend on the scale on which welfare is measured, and in a way that avoids the restrictive constraints on the shape of the SWF. (shrink)

Substituted judgment has increasingly become the accepted standard for rendering decisions for incapacitated adults in the USA. A broad exception exists with regard to patients with diminished capacity secondary to depressive disorders, as such patients’ previous wishes are generally not honoured when seeking to turn down life-preserving care or pursue aid-in-dying. The result is that physicians often force involuntary treatment on patients with poor medical prognoses and/or low quality of life as a result of their depressive symptoms when similarly situated (...) incapacitated patients without such depressive symptoms would have their previous wishes honoured via substituted judgment. This commentary argues for reconsidering this approach and for using a substituted judgment standard for a subset of EMP/LQL patients seeking death. (shrink)

The public discourse surrounding sex and severe disability over the past 40 years has largely focused on protecting vulnerable populations from abuse. However, health professionals and activists are increasingly recognising the inherent sexuality of disabled persons and attempting to find ways to accommodate their intimacy needs. This essay explores several ethical issues arising from such efforts.

Lara Buchak argues for a version of rank-weighted utilitarianism that assigns greater weight to the interests of the worse off. She argues that our distributive principles should be derived from the preferences of rational individuals behind a veil of ignorance, who ought to be risk averse. I argue that Buchak’s appeal to the veil of ignorance leads to a particular way of extending rank-weighted utilitarianism to the evaluation of uncertain prospects. This method recommends choices that violate the unanimous preferences of (...) rational individuals and choices that guarantee worse distributions. These results, I suggest, undermine Buchak’s argument for rank-weighted utilitarianism. (shrink)

Many economists and philosophers assume that status quo bias is necessarily irrational. I argue that, in some cases, status quo bias is fully rational. I discuss the rationality of status quo bias on both subjective and objective theories of the rationality of preferences. I argue that subjective theories cannot plausibly condemn this bias as irrational. I then discuss one kind of objective theory, which holds that a conservative bias toward existing things of value is rational. This account can fruitfully explain (...) some compelling aspects of common sense morality, and it may justify status quo bias. (shrink)

Physicians expressing opinions on medical matters that run contrary to the consensus of experts pose a challenge to licensing bodies and regulatory authorities. While the right to express contrarian views feeds a robust marketplace of ideas that is essential for scientific progress, physicians advocating ineffective or dangerous cures, or actively opposing public health measures, pose a grave threat to societal welfare. Increasingly, a distinction has been made between professional speech that occurs during the physician-patient encounter and public speech that transpires (...) beyond the clinical setting, with physicians being afforded wide latitude to voice empirically false claims outside the context of patient care. This paper argues that such a bifurcated model does not sufficiently address the challenges of an age when mass communications and social media allow dissenting physicians to offer misleading medical advice to the general public on a mass scale. Instead, a three-tiered model that distinguishes between citizen speech, physician speech and clinical speech would best serve authorities when regulating physician expression. (shrink)

The Rachels–Temkin spectrum arguments against the transitivity of better than involve good or bad experiences, lives, or outcomes that vary along multiple dimensions—e.g., duration and intensity of pleasure or pain. This paper presents variations on these arguments involving combinations of good and bad experiences, which have even more radical implications than the violation of transitivity. These variations force opponents of transitivity to conclude that something good is worse than something that isn’t good, on pain of rejecting the good altogether. That (...) is impossible, so we must reject the spectrum arguments. (shrink)

Totalism is the view that one distribution of well-being is better than another just in case the one contains a greater sum of well-being than the other. Many philosophers, following Parfit, reject totalism on the grounds that it entails the repugnant conclusion: that, for any number of excellent lives, there is some number of lives that are barely worth living whose existence would be better. This paper develops a theory of welfare aggregation—the lexical-threshold view—that allows totalism to avoid the repugnant (...) conclusion, as well as its analogues involving suffering populations and the lengths of individual lives. The theory is grounded in some independently plausible views about the structure of well-being, identifies a new source of incommensurability in population ethics, and avoids some of the implausibly extreme consequences of other lexical views, without violating the intuitive separability of lives. (shrink)

This paper is about the role of interpersonal comparisons in Harsanyi's aggregation theorem. Harsanyi interpreted his theorem to show that a broadly utilitarian theory of distribution must be true even if there are no interpersonal comparisons of well-being. How is this possible? The orthodox view is that it is not. Some argue that the interpersonal comparability of well-being is hidden in Harsanyi's premises. Others argue that it is a surprising conclusion of Harsanyi's theorem, which is not presupposed by any one (...) of the premises. I argue instead that Harsanyi was right: his theorem and its weighted-utilitarian conclusion do not require interpersonal comparisons of well-being. The key to making sense of this possibility is to treat Harsanyi's weights as dimensional constants rather than dimensionless numbers. (shrink)

Priority for solid organ transplant generally does not consider the underlying cause of the need for transplantation. This paper argues that a distinctive set of factors justify assigning lower priority to willfully unvaccinated individuals who require transplant as a result of suffering from COVID‐19. These factors include the personal responsibility of the patients for their own condition and the public outrage likely to ensue if willfully unvaccinated patients receive organs at the expense of vaccinated ones. The paper then proposes a (...) three‐prong test for similar deviations from the current allocation standard that incorporates patient responsibility, foreseeability and avoidability, and the frequency of the occurrence. (shrink)

Is well-being the kind of thing that can be summed across individuals? This paper takes a measurement-theoretic approach to answering this question. To make sense of adding well-being, we would need to identify some natural "concatenation" operation on the bearers of well-being that satisfies the axioms of extensive measurement and can therefore be represented by the arithmetic operation of addition. I explore various proposals along these lines, involving the concatenation of segments within lives over time, of entire lives led alongside (...) one another or in sequence, and of evaluatively basic propositions via conjunction. All of these proposals turn out to carry highly controversial commitments about the good. I do not claim that these commitments are unacceptable. But they suggest that we cannot simply take for granted, as many philosophers do, that there is any such thing as the sum of well-being. (shrink)

The integrity of the patient–physician relationship depends on maintaining professional boundaries. While ethicists and professional organisations have devoted significant consideration to the subject of sexual boundary transgressions, the subject of non-sexual boundaries, especially outside the mental health setting, has been largely neglected. While professional organisations may offer guidance on specific subjects, such as accepting gifts or treating relatives, as well as general guidance on transparency and conflict of interest, what is missing is a principle-based method that providers can use to (...) assess non-sexual interactions with patients that transcend norms of practice. This paper attempts to offer an operational model for such assessment that considers not only the traditional emphasis on beneficence, but also incorporates concerns over entanglement and concordance. (shrink)

This paper presents a new kind of problem in the ethics of distribution. The problem takes the form of several “calibration dilemmas,” in which intuitively reasonable aversion to small-stakes inequalities requires leading theories of distribution to recommend intuitively unreasonable aversion to large-stakes inequalities. We first lay out a series of such dilemmas for prioritarian theories. We then consider a widely endorsed family of egalitarian views and show that they are subject to even more forceful calibration dilemmas than prioritarian theories. Finally, (...) we show that our results challenge common utilitarian accounts of the badness of inequalities in resources. (shrink)

The “Goldwater rule,” a policy adopted by the American Psychiatry Association in 1973, prohibits organization members from diagnosing or offering professional opinions regarding the mental health of public figures without both first-hand evaluation and authorization. Initially developed in response to a controversial survey of APA members during the 1964 Presidential election campaign, the ethics rule faced few large scale challenges until the election of Donald Trump in 2016. Since that time, a significant number of psychiatrists have either violated or criticized (...) the rule openly. This paper argues that whatever the initial merits of the rule, the prohibition has since been rendered obsolete by the combined lack of professional consensus supporting the policy, absence of a meaningful enforcement mechanism, and the credible statements of non-APA members in the mental health professions regarding public figures. (shrink)

Since several landmark legal decisions in the 1970s and 1980s, substituted judgment has become widely accepted as an approach to decision‐making for incapacitated patients that incorporates their autonomy and interests. Two notable exceptions have been cases involving minors and those involving cognitively or psychiatrically impaired individuals who never previously possessed the ability to contemplate the medical decisions involved in their care. While a best interest standard may have universal merit in pediatric cases, this paper argues that substituted judgement has been (...) rejected too broadly for “never‐capacitated” patients. It then lays out a three‐condition test for using substituted judgment in these cases. For a subset of “never‐capacitated” patients who belong to distinctive religious or cultural communities with clear medical values and have an appropriate surrogate, these values may prove helpful in vindicating the interests of the never‐capacitated and should not be blanketly dismissed. (shrink)

The American Journal of Bioethics, Volume 12, Issue 9, Page 9-11, September 2012.

This essay assesses the degree to which the deaf were regarded as a disabled population in medical, religious, and legal thought during the Renaissance, chronologically identified with the period between approximately 1500 and 1650. The primary geographic focus rests on the German-speaking lands of central Europe. Analysis shows that the idea of deafness as a disability here was composite one, making connections between inability to hear and intellectual impairment, moral deficiency, and disease. This contrasts with recent findings elsewhere in Europe, (...) where intellectuals were more focused on sign languages as a means of integrating the deaf into hearing society. (shrink)

This paper develops and explores a new framework for theorizing about the measurement and aggregation of well-being. It is a qualitative variation on the framework of social welfare functionals developed by Amartya Sen. In Sen’s framework, a social or overall betterness ordering is assigned to each profile of real-valued utility functions. In the qualitative framework developed here, numerical utilities are replaced by the properties they are supposed to represent. This makes it possible to characterize the measurability and interpersonal comparability of (...) well-being directly, without the use of invariance conditions, and to distinguish between real changes in well-being and merely representational changes in the unit of measurement. The qualitative framework is shown to have important implications for a range of issues in axiology and social choice theory, including the characterization of welfarism, axiomatic derivations of utilitarianism, the meaningfulness of prioritarianism, the informational requirements of variable-population ethics, the impossibility theorems of Arrow and others, and the metaphysics of value. (shrink)

“An original, compelling, and provocative exploration of ethical issues in our society, with thoughtful and balanced commentary. I have not seen anything like it.” —Alan Lightman, author of Einstein’s Dreams Drawing upon the author’s two decades teaching medical ethics, as well as his work as a practicing psychiatrist, this profound and addictive little book offers up challenging ethical dilemmas and asks readers, What would you do? A daughter gets tested to see if she’s a match to donate a kidney to (...) her father. The test reveals that she is not the man’s biological daughter. Should the doctor tell the father? Or the daughter? A deaf couple prefers a deaf baby. Should they be allowed to use medical technology to ensure they have a child who can’t hear? Who should get custody of an embryo created through IVF when a couple divorces? Or, when you or a loved one is on life support, Who says you’re dead? In short, engaging scenarios, Dr. Appel takes on hot-button issues that many of us will confront: genetic screening, sexuality, privacy, doctor-patient confidentiality. He unpacks each hypothetical with a brief reflection drawing from science, philosophy, and history, explaining how others have approached these controversies in real-world cases. Who Says You’re Dead? is designed to defy easy answers and to stimulate thought and even debate among professionals and armchair ethicists alike. (shrink)

How should we choose between uncertain prospects in which different possible people might exist at different levels of wellbeing? Alex Voorhoeve and Marc Fleurbaey offer an egalitarian answer to this question. I give some reasons to reject their answer and then sketch an alternative, which I call person-affecting prioritarianism.

Recent findings indicate that complex cognitive functions are organized at a global level in the brain and rely on large-scale information processing requiring functional integration of multiple disparate neural assemblies. The critical question of the integration of distributed brain activities is whether the essential integrative role can be attributed to a specific structure in the brain or whether this ability is inherent to the cognitive network as a whole. The results of the present study show that mean values of the (...) running correlation function in frontal-temporal EEG pairs with one electrode in the anterior cingulate cortex are significantly higher than the same values in other frontal-temporal pairs. These findings indicate a particular role of the ACC in large-scale communication, which could reflect its unique integrative functions in cognitive processing. (shrink)

Minor-attracted persons raise multiple ethical and legal challenges. Sexual contact between adults and children is justly prohibited on child welfare grounds. Advances in technology raise the prospect of interventions for minor-attracted persons that have the potential to reduce harm to children by diverting would-be offenders to other endeavors that nonetheless may generate moral disgust This essay examines three of these potential harm reduction technologies (sex robots, haptic devices and synthetic child pornography) and raises the possibility that their use can be (...) justified and their acceptance morally obligatory if doing so reduces harm to real children. (shrink)

Müller provides a valuable synthesis of neurobiological evidence on the epigenetic development of neural structures involved in language acquisition. The pluripotentiality of developing neural tissue crucially constrains linguistic/cognitive theorizing about supposedly innate neural mechanisms and contributes significantly to our understanding of experience–dependent processes involved in language acquisition. Without this understanding, any proposed explanation of language acquisition is suspect.