Foundation models such as GPT-4 are fine-tuned to avoid unsafe or otherwise problematic behavior, such as helping to commit crimes or 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)

A predominância dos temas literários e fenomenológicos nos textos sartrianos dos anos 1930 já chamou atenção de vários especialistas e comentadores da obra do autor. É notável, entretanto, como os exames dos textos dessa época contentam-se, no mais das vezes, em apontar o paralelismo ou a contemporaneidade entre fenomenologia e literatura ali presente, sem realçar a fundamentação que uma parece conferir à outra e que constitui uma das questões mais interessantes que a filosofia sartriana levanta. Há que se incluir, na (...) verdade, a estética literária de Sartre na compreensão de suas teses fenomenológicas; teses estas que, por outro lado, fornecem o solo filosófico da literatura sartriana, em termos tanto de autoria quanto de crítica. O propósito deste artigo é demonstrar, assim, uma dependência ou, se quisermos, implicação mútua entre literatura e fenomenologia não apenas central em relação àquele período da produção de Sartre como, também, determinante para seus projetos teóricos e ficcionais ulteriores. (shrink)

Initial responses to questionnaires used to assess participants' understanding of informed consent for malaria vaccine trials conducted in the United States and Mali were tallied. Total scores were analyzed by age, sex, literacy (if known), and location. Ninety-two percent (92%) of answers by United States participants and 85% of answers by Malian participants were correct. Questions more likely to be answered incorrectly in Mali related to risk, and to the type of vaccine. For adult participants, independent predictors of higher scores (...) were younger age and female sex in the United States, and male sex in Mali. Scores in the United States were higher than in Mali (P = 0.005). Despite this difference participants at both sites were well informed overall. Although interpretation must be qualified because questionnaires were not intended as research tools and were not standardized among sites, these results do not support concerns about systematic low understanding among research participants in developing versus developed countries. (shrink)

It is widely held that all lies are assertions: the traditional definition of lying entails that, in order to lie, speakers have to assert something they believe to be false. It is also widely held that assertion contrasts with presupposition and, in particular, that one cannot assert something by presupposing it. Together, these views imply that speakers cannot lie with presuppositions—a view that Andreas Stokke has recently explicitly defended. The aim of this paper is to argue that speakers can lie (...) with presuppositions, and to discuss some of the implications this outcome has for current research on lying, assertion and presupposition. (shrink)

The distinction between lying and mere misleading is commonly tied to the distinction between saying and conversationally implicating. Many definitions of lying are based on the idea that liars say something they believe to be false, while misleaders put forward a believed-false conversational implicature. The aim of this paper is to motivate, spell out, and defend an alternative approach, on which lying and misleading differ in terms of commitment: liars, but not misleaders, commit themselves to something they believe to be (...) false. This approach entails that lying and misleading involve speech-acts of different force. While lying requires the committal speech-act of asserting, misleading involves the non-committal speech-act of suggesting. The approach leads to a broader definition of lying that can account for lies that are told while speaking non-literally or with the help of presuppositions, and it allows for a parallel definition of misleading, which so far is lacking in the debate. (shrink)

This paper defends stage theory against the argument from diachronic counting. It argues that stage theorists can appeal to quantifier domain restriction in order to accommodate intuitions about diachronic counting sentences. Two approaches involving domain restriction are discussed. According to the first, domains of counting are usually restricted to stages at the time of utterance. This approach explains intuitions in many cases, but is theoretically costly and delivers wrong counts if diachronic counting is combined with fission or fusion. On the (...) second approach, domains of counting are usually restricted in an indeterminate way, so as to include at most one member of any maximal class of counterpart-interrelated stages (with respect to a certain utterance). This view can accommodate all the relevant intuitions about counting sentences, and it fits well with a new stage-theoretic view of reference that allows speakers to refer to both present and past stages. (shrink)

Defeasible argumentation has been applied successfully in several real-world domains in which it is necessary to handle incomplete and contradictory information. In recent years, there have been interesting attempts to carry out argumentation processes supported by massive repositories developing argumentative reasoning applications. One of such efforts builds arguments by retrieving information from relational databases using the DBI-DeLP framework; this article presents eDBI-DeLP, which extends the original DBI-DeLP framework by providing two novel aspects which refine the interaction between DeLP programs and (...) relational databases. First, we expand the expressiveness of dbi-delp programs by providing ways of controlling how the information in databases is recovered; this is done by introducing filters that enable an improved fine-grained control on the argumentation processes which become useful in applications, providing the semantics and the implementation of such filters. Second, we introduce an argument comparison criterion which can be adjusted at the level of literals to model particular features such as credibility and topic expertise, among others. These new tools can be particularly useful in environments such as medical diagnosis expert systems, decision support systems, or recommender systems based on argumentation, where datasets are often provided in the form of relational databases. (shrink)

Many recent definitions of lying are based on the notion of what is said. This paper argues that says-based definitions of lying cannot account for lies involving non-literal speech, such as metaphor, hyperbole, loose use or irony. It proposes that lies should instead be defined in terms of assertion, where what is asserted need not coincide with what is said. And it points to possible implications this outcome might have for the ethics of lying.

In arguing against a supposed ambiguity, philosophers often rely on the zeugma test. In an application of the zeugma test, a supposedly ambiguous expression is placed in a sentence in which several of its supposed meanings are forced together. If the resulting sentence sounds zeugmatic, that is taken as evidence for ambiguity; if it does not sound zeugmatic, that is taken as evidence against ambiguity. The aim of this article is to show that arguments based on the second direction of (...) the test are misguided: ambiguous expressions, and in particular philosophically contested ones, do not reliably lead to zeugmaticity, so an absence of zeugmaticity provides no meaningful evidence for an absence of ambiguity. (shrink)

Pictures are notably absent from the current debate about how to define lying. Theorists in this debate tend to focus on linguistic means of communication and do not consider the possibility of lying with photographs, drawings and other kinds of pictures. The aim of this paper is to show that such a narrow focus is misguided: there is a strong case to be made for the possibility of lying with pictures and this possibility allows for insights concerning the question of (...) how lying should be defined. (shrink)

We study how changes in unemployment risk affect firms’ productivity and whether firm-initiated policies can mitigate the moral hazard problem created by increases in unemployment insurance benefits that might decrease workers’ incentives to work hard. We focus on state-specific changes in UIB levels as a quasi-natural experiment. While a large body of research has examined UIBs, including their effect on unemployed workers, few studies investigate whether UIBs have any impact on a firm’s overall productivity. Using data on firm-level total factor (...) productivity and state-level UIBs, we find a negative association between productivity and UIBs. We also find that the negative association is weaker for firms with higher employee-welfare indices than for firms with lower indices, suggesting that the adverse effect of higher UIBs on productivity is mitigated by policies that benefit workers’ welfare. More specifically, we find that among policies that are under the umbrella of corporate social responsibility, a subset of employee-welfare policies are more effective in managing moral hazard problems than other policies. (shrink)

A RESPONSABILIDADE CIVIL PELO ABUSO DO DIREITO À LIBERDADE DE EXPRESSÃO EM CONFLITO COM OS DIREITOS DA PERSONALIDADE: UMA ANÁLISE DO CASO DAS BIOGRAFIAS NÃO AUTORIZADAS.

Intentionalism is the view that demonstratives, gradable adjectives, quantifiers, modals and other context‐sensitive expressions are intention‐sensitive: their semantic value on a given use is fixed by speaker intentions. The first aim of this paper is to defend Intentionalism against three recent objections, according to which speakers at least sometimes do not have suitable intentions when using supposedly intention‐sensitive expressions. Its second aim is to thereby shed light on the so far little‐explored question of which kinds of intentions can be semantically (...) relevant. (shrink)

Book reviewed in this article: Principles of Biomedical Ethics. By Tom L. Beauchamp and James F. Childress.

To understand the endogenous components of the event-related brain potential (ERP), we must use data about the components' antecedent conditions to form hypotheses about the information-processing function of the underlying brain activity. These hypotheses, in turn, generate testable predictions about the consequences of the component. We review the application of this approach to the analysis of the P300 component. The amplitude of the P300 is controlled multiplicatively by the subjective probability and the task relevance of the eliciting events, whereas its (...) latency depends on the duration of stimulus evaluation. These and other factors suggest that the P300 is a manifestation of activity occurring whenever one's model of the environment must be revised. Tests of three predictions based on this “context updating” model are reviewed. Verleger's critique is based on a misconstrual of the model as well as a partial and misleading reading of the relevant literature. (shrink)

ABSTRACT Copredication, as exhibited by sentences such as ‘That book is heavy but informative,’ is commonly seen as a phenomenon that is tied to sentences featuring polysemous expressions. David Liebesman and Ofra Magidor have recently attacked this view by arguing that ‘book’ has a single context-sensitive sense. The first aim of the present paper is to show that Liebesman and Magidor are wrong to claim that ‘book’ is univocal, but that they may nonetheless be right to question that copredication requires (...) polysemy. Its second aim is to consider implications of this result for the debates on copredication and on semantic variability. (shrink)

This work has been selected by scholars as being culturally important, and is part of the knowledge base of civilization as we know it. This work is in the "public domain in the United States of America, and possibly other nations. Within the United States, you may freely copy and distribute this work, as no entity (individual or corporate) has a copyright on the body of the work. Scholars believe, and we concur, that this work is important enough to be (...) preserved, reproduced, and made generally available to the public. We appreciate your support of the preservation process, and thank you for being an important part of keeping this knowledge alive and relevant. (shrink)

In the most recent debate of phenomenological as well as analytical provenance, Husserl’s reflection, although pioneer regarding ethics, seems to suffer of a general depreciation enrooted in his supposed excess rationalism. The husserlian ethics would remain subjected to the dominance of a logico-objectifying reason, which appears to be the fundamental dimension of intentionality. The wide range of life-experiences would then be crushed by the hegemony of objectifying acts. Are we really sure? Our contribution will rather consist in raising the originality (...) of Husserl’s method which lastly to reconcile two seemingly contradictory requirements: the autonomy of ethical and its dependence towards logical reason. We shall notice thus a progressive dissociatoon from Brentano’s teaching leading to an overall revision of intentionality. If anything such as rationalism exists in Husserl’s thinking, it can only be asserted on the condition of a radical criticism of the concept of “representation” – from where, as we shall see, emerges a newly articulated relationship between “object” and “value”. (shrink)

By its very name, phenomenology seems to invoke a priority claim on phenomena. And yet it has not been necessary to wait for phenomenology in oder to have a proper account of phenomena. One need only to take a look at the history of philosophy, from Plato to Kant, as well as at the history of sciences, from physics to psychology, so as to register a wide range of uses concerning the concept of phenomenon. The understanding of what a phenomenon (...) is, in a phenomenological sense, should therefore at least not be completely compatible with any other, if phenomenology has to exist as a discipline and as a method. That is one of the issues at stake in the debate between phenomenology and neokantianism, between Husserl, to a certain extent Brentano, and Natorp. It is on this debate that we shall draw attention, if we wish to grasp the reasons behind the phenomenological conception of the phenomenon, especially at the time of the Logische Untersuchungen. (shrink)

Within the Seinsfrage, the question of being, the «who» question, or Werfrage, arises at once. Who is this «who»? In other words, who am I? Questioning the being, inevitably means questioning ourselves. Would it thus be possible to conceive of a subject without being? This question, paradoxical as it may appear, nevertheless has a historiographical legitimacy, if we consider that «phenomenological anthropology», as the young Heidegger conceived it, precedes in effect the formulation of the Seinsfrage. From his very first lectures (...) in Fribourg, an access into subjectivity, apprehended as a «wordly ipseity », is thereby opened. While criticizing Natorp who unreservedly denied such a possibility, Heidegger yet ends up conceding the point almost entirely. This, at least, is our hypothesis: the «I» exists only at the cost of a tension between facticity and existence, a tension which, in the eyes of Natorp, leads to a confirmation that the «I», strictly speaking, cannot be – and this allows us therefore a glimpse into the legitimate possibility of a Werfrage without Seinsfrage. (shrink)

Where should we meet? This may seem like a trivial decision, mostly focused on convenience. However, location can represent integral values, attitudes, and priorities. Especially when the meeting i...

In 1877 Erwin Rohde suggested the conjecture Argino instead of the manuscript reading argiro in Ampelius 8, 16. His conjecture did not find approval among editors of the late antique author and is virtually forgotten. A fresh look at Rohde’s suggestion shows, however, that it is convincing with regards to palaeography and consents with literary and epigraphical evidence. With Argino instead of a conjecture like Bargyliis, the passage is no more a testimony for the Carian city of Bargylia, but for (...) the Ionian Erythrae and its religious history (Aphrodite, Heracles, and the Sibyl). (shrink)

Lying requires asserting a disbelieved proposition, that much is widely accepted in the debate on how to define lying. But what else is required? Does lying require a particular linguistic manner of expression, such as saying? Does the proposition asserted have to be false (and not merely disbelieved)? And does lying require an intention to deceive? The aim of this chapter is to provide an opinionated introduction to the debates on these questions that takes into account both theoretical considerations and (...) empirical data. (shrink)

Sam Berstler defends a general moral advantage for misleading over lying by arguing that liars, but not misleaders, act unfairly toward the other members of their linguistic community. This article spells out three difficulties for Berstler’s account. First, though Berstler aims to avoid an error theory, it is dubitable that her account fits with intuitions on the matter. Second, there are some lies that do not exhibit the unfairness Berstler identifies. Third, fairness is not the only morally relevant difference between (...) lying and misleading. (shrink)

We study first-order logic with two variables FO² and establish a small substructure property. Similar to the small model property for FO² we obtain an exponential size bound on embedded substructures, relative to a fixed surrounding structure that may be infinite. We apply this technique to analyse the satisfiability problem for FO² under constraints that require several binary relations to be interpreted as equivalence relations. With a single equivalence relation, FO² has the finite model property and is complete for non-deterministic (...) exponential time, just as for plain FO² With two equivalence relations, FO² does not have the finite model property, but is shown to be decidable via a construction of regular models that admit finite descriptions even though they may necessarily be infinite. For three or more equivalence relations, FO² is undecidable. (shrink)

In the past few years, machine learning (ML) tools have been implemented with success in the medical context. However, several practitioners have raised concerns about the lack of transparency—at the algorithmic level—of many of these tools; and solutions from the field of explainable AI (XAI) have been seen as a way to open the ‘black box’ and make the tools more trustworthy. Recently, Alex London has argued that in the medical context we do not need machine learning tools to be (...) interpretable at the algorithmic level to make them trustworthy, as long as they meet some strict empirical desiderata. In this paper, we analyse and develop London’s position. In particular, we make two claims. First, we claim that London’s solution to the problem of trust can potentially address another problem, which is how to evaluate the reliability of ML tools in medicine for regulatory purposes. Second, we claim that to deal with this problem, we need to develop London’s views by shifting the focus from the opacity of algorithmic details to the opacity of the way in which ML tools are trained and built. We claim that to regulate AI tools and evaluate their reliability, agencies need an explanation of how ML tools have been built, which requires documenting and justifying the technical choices that practitioners have made in designing such tools. This is because different algorithmic designs may lead to different outcomes, and to the realization of different purposes. However, given that technical choices underlying algorithmic design are shaped by value-laden considerations, opening the black box of the design process means also making transparent and motivating (technical and ethical) values and preferences behind such choices. Using tools from philosophy of technology and philosophy of science, we elaborate a framework showing how an explanation of the training processes of ML tools in medicine should look like. (shrink)

In several recent papers and a monograph, Andreas Stokke argues that questions can be misleading, but that they cannot be lies. The aim of this paper is to show that ordinary speakers disagree. We show that ordinary speakers judge certain kinds of insincere questions to be lies, namely questions carrying a believed-false presupposition the speaker intends to convey. These judgements are robust and remain so when the participants are given the possibility of classifying the utterances as misleading or as deceiving. (...) The judgements contrast with judgements participants give about cases of misleading or deceptive behaviour, and they pattern with judgements participants make about declarative lies. Finally, the possibility of lying with non-declaratives is not confined to questions: ordinary speakers also judge utterances of imperative, exclamative and optative sentences carrying believed-false presuppositions to be lies. (shrink)

Four ethical values — maximizing benefits, treating equally, promoting and rewarding instrumental value, and giving priority to the worst off — yield six specific recommendations for allocating medical resources in the Covid-19 pandemic: maximize benefits; prioritize health workers; do not allocate on a first-come, first-served basis; be responsive to evidence; recognize research participation; and apply the same principles to all Covid-19 and non–Covid-19 patients.

In the past few years, the ethical ramifications of AI technologies have been at the center of intense debates. Considerable attention has been devoted to understanding how a morally responsible practice of data science can be promoted and which values have to shape it. In this context, ethics and moral responsibility have been mainly conceptualized as compliance to widely shared principles. However, several scholars have highlighted the limitations of such a principled approach. Drawing from microethics and the virtue theory tradition, (...) in this paper, we formulate a different approach to ethics in data science which is based on a different conception of “being ethical” and, ultimately, of what it means to promote a morally responsible data science. First, we develop the idea that, rather than only compliance, ethical decision-making consists in using certain moral abilities, which are cultivated by practicing and exercising them in the data science process. An aspect of virtue development that we discuss here is moral attention, which is the ability of data scientists to identify the ethical relevance of their own technical decisions in data science activities. Next, by elaborating on the capability approach, we define a technical act as ethically relevant when it impacts one or more of the basic human capabilities of data subjects. Therefore, rather than “applying ethics”, data scientists should cultivate ethics as a form of reflection on how technical choices and ethical impacts shape one another. Finally, we show how this microethical framework concretely works, by dissecting the ethical dimension of the technical procedures involved in data understanding and preparation of electronic health records. (shrink)

In the science and values literature, scholars have shown how science is influenced and shaped by values, often in opposition to the ‘value free’ ideal of science. In this paper, we aim to contribute to the science and values literature by showing that the relation between science and values flows not only from values into scientific practice, but also from (allegedly neutral) science to values themselves. The extant literature in the ‘science and values’ field focuses by and large on reconstructing, (...) post hoc, how values have influenced science; our reconstruction of the case studies, instead, aims to show that scientific concepts and methods _too, because of specific identifiable characteristics,_ can promote some values rather than (or at the expense of) others. We explain this bidirectional relation in analogy to debates on the normativity of technical artifacts and on feminist approaches in science, and we illustrate our claims with cases from the health sciences and machine learning. While our arguments in this paper also draw on post hoc reconstructions, we intend to show where, in the science in the making, we should engage not only with the question whether a practice is value-laden, but also how specific conceptual and methodological choices can influence values down the road. All in all, these considerations expand the ways in which philosophers can contribute to more value-aware scientific practices. (shrink)

In Emanuel Adler's distinctive constructivist approach to international relations theory, international practices evolve in tandem with collective knowledge of the material and social worlds. This book - comprising a selection of his journal publications, a new introduction and three previously unpublished articles - points IR constructivism in a novel direction, characterized as 'communitarian'. Adler's synthesis does not herald the end of the nation-state; nor does it suggest that agency is unimportant in international life. Rather, it argues that what mediates (...) between individual and state agency and social structures are communities of practice, which are the wellspring and repositories of collective meanings and social practices. The concept of communities of practice casts new light on epistemic communities and security communities, helping to explain why certain ideas congeal into human practices and others do not, and which social mechanisms can facilitate the emergence of normatively better communities. (shrink)

In the last five years Europe experienced a dramatic increase in anti-Semitic incidents. Discourse once reputed unacceptable is now routinely voiced in mainstream circles, the press, and the corridors of power. This article sets out to explain the nature of this phenomenon, the reasons for its recurrence, and its severity.