Head-driven phrase structure grammar (HPSG) is one of the most prominent theories employed in deep parsing of natural language. Many linguistic theories are arguably best formalized in extensions of modal or dynamic logic (Keller, Feature logics, infinitary descriptions and grammar, 1993; Kracht, Linguistics Philos 18:401–458, 1995; Moss and Tiede, In: Blackburn, van Benthem, and Wolther (eds.) Handbook of modal logic, 2006), and HPSG seems to be no exception. Adequate extensions of dynamic logic have not been studied in detail, however; (...) the most important aspect is the reference to sets of substructures. In this paper, an adequate extension is identified, and some important results are established: Satisfiability is highly undecidable, and model checking is shown to be in EXPTIME and PSPACE-hard. A fragment with polynomial time model checking procedures is identified; it is shown to cover considerable fragments of HPSG. (shrink)

Landgrebe and Smith :2061–2081, 2021) present an unflattering diagnosis of recent advances in what they call language-centric artificial intelligence—perhaps more widely known as natural language processing: The models that are currently employed do not have sufficient expressivity, will not generalize, and are fundamentally unable to induce linguistic semantics, they say. The diagnosis is mainly derived from an analysis of the widely used Transformer architecture. Here I address a number of misunderstandings in their analysis, and present what I take to be (...) a more adequate analysis of the ability of Transformer models to learn natural language semantics. To avoid confusion, I distinguish between inferential and referential semantics. Landgrebe and Smith ’s analysis of the Transformer architecture’s expressivity and generalization concerns inferential semantics. This part of their diagnosis is shown to rely on misunderstandings of technical properties of Transformers. Landgrebe and Smith also claim that referential semantics is unobtainable for Transformer models. In response, I present a non-technical discussion of techniques for grounding Transformer models, giving them referential semantics, even in the absence of supervision. I also present a simple thought experiment to highlight the mechanisms that would lead to referential semantics, and discuss in what sense models that are grounded in this way, can be said to understand language. Finally, I discuss the approach Landgrebe and Smith advocate for, namely manual specification of formal grammars that associate linguistic expressions with logical form. (shrink)

1. Fairness is about the just distribution of society's resources, and in ML, the main resource being distributed is model performance, e.g. the translation quality produced by machine translation...

I identify a class of arguments against multiple realization (MR): BookofSand arguments. The arguments are in their general form successful under reasonably uncontroversial assumptions, but this, on the other hand, turns the table on identity theory: If arguments from MR can always be refuted by BookofSand arguments, is identity theory falsifiable? In the absence of operational demarcation criteria, it is not. I suggest a parameterized formal demarcation principle for brain state/process types and show how it can be used to identify (...) previously unconsidered contenders for evidence for MR, e.g., binary classification, division, and sorting. For these to be actual instances of MR, the corresponding psychological kinds must be verifiably, relevantly similar. I also briefly discuss possible linguistic, behavioral, and experimental demarcation criteria for psychological kinds. (shrink)

Taking up the controversial issue of the value of the laws of non-ideal cities in Plato’s Statesman, the paper argues for a modified version of the traditional interpretation, as defended against Christopher Rowe’s influential criticism. The paper agrees with the traditional view that the established laws of non-ideal cities are assumed to be good laws and that the Eleatic Stranger’s justification for this assumption can be found in 300b. But it also argues that this defence of the traditional interpretation must (...) be based on a revised understanding of the role of legislation in the Stranger’s conception of ideal rule. The unique contribution of the expert to good government consists in his distinctive ability to judge what is appropriate in each case, not in his work as legislator, which, on closer inspection, turns out to be based on the very same ancestral traditions that the second best city must resort to. (shrink)

Deep neural networks are said to be opaque, impeding the development of safe and trustworthy artificial intelligence, but where this opacity stems from is less clear. What are the sufficient properties for neural network opacity? Here, I discuss five common properties of deep neural networks and two different kinds of opacity. Which of these properties are sufficient for what type of opacity? I show how each kind of opacity stems from only one of these five properties, and then discuss to (...) what extent the two kinds of opacity can be mitigated by explainability methods. (shrink)

Scientists increasingly approach the world through machine learning techniques, but philosophers of science often question their epistemic status. Some philosophers have argued that the use of unsupervised clustering algorithms is more justified than the use of supervised classification, because supervised classification is more biased, and because (parametric) simplicity plays a different and more interesting role in unsupervised clustering. I call these arguments the No-Bias Argument and the Simplicity-Truth Argument. I show how both arguments are fallacious and how, on the contrary, (...) the use of supervised classification is at least as justified as the use of unsupervised clustering. (shrink)

Most, if not all, philosophers agree that computers cannot learn what words refers to from raw text alone. While many attacked Searle’s Chinese Room thought experiment, no one seemed to question this most basic assumption. For how can computers learn something that is not in the data? Emily Bender and Alexander Koller ( 2020 ) recently presented a related thought experiment—the so-called Octopus thought experiment, which replaces the rule-based interlocutor of Searle’s thought experiment with a neural language model. The Octopus (...) thought experiment was awarded a best paper prize and was widely debated in the AI community. Again, however, even its fiercest opponents accepted the premise that what a word refers to cannot be induced in the absence of direct supervision. I will argue that what a word refers to _is_ probably learnable from raw text alone. Here’s why: higher-order concept co-occurrence statistics are stable across languages and across modalities, because language use (universally) reflects the world we live in (which is relatively stable). Such statistics are sufficient to establish what words refer to. My conjecture is supported by a literature survey, a thought experiment, and an actual experiment. (shrink)

The first part of the article demonstrates how the concept of mental resilience has developed from the 1970s to the 2010s, spreading from the field of developmental psychopathology to a wide range of psychological and psychotherapeutic disciplines. Today, there are many varied definitions of mental resilience, ranging from a relation to the concept of mental vulnerability to a related concept of mental strength. The second part of the article demonstrates how the current popularisation of the idea of individual resilience is (...) associated with a social and cultural problematisation of the vulnerable life as well as with a psychopolitical and governmental focus on enhancing the mental capacity of citizens. (shrink)

The first part of the article demonstrates how the concept of mental resilience has developed from the 1970s to the 2010s, spreading from the field of developmental psychopathology to a wide range of psychological and psychotherapeutic disciplines. Today, there are many varied definitions of mental resilience, ranging from a relation to the concept of mental vulnerability to a related concept of mental strength. The second part of the article demonstrates how the current popularisation of the idea of individual resilience is (...) associated with a social and cultural problematisation of the vulnerable life as well as with a psychopolitical and governmental focus on enhancing the mental capacity of citizens. (shrink)

This collection breaks new ground in four key areas of feminist social thought: the sex/gender debates; challenges to liberalism/equality; feminist ethics; and feminist perspectives on global ethics and politics in the 21st century. Altogether, the essays provide an innovative look at feminist philosophy while making substantive contributions to current debates in gender theory, ethics, and political thought.

Developers of artificial intelligence (AI) often cannot explain the inferences their neural networks make, at least not in ways that satisfy user needs. XAI—explainable artificial intelligence—aims to develop techniques for providing such explanations. XAI researchers have adopted techniques that, to philosophers, seem representationalist–internalist, leading some philosophers to call for more externalist alternatives. But is explaining AI models through causally related external factors feasible? I suggest we compare the idea of externalist XAI to so‐called functionalist XAI. Two common arguments against functionalist (...) XAI are the Epistemic Argument and the Trust Argument, respectively. I discuss whether these arguments can be met by externalists, as well as a new, previously unnoticed challenge from random initialization. (shrink)

The article discusses the relation between political office and the rule of law in Plato’s dialogue Statesman. Taking its starting-point from an observation about the Statesman’s peculiar approach to constitutional analysis, the article argues that what Plato is concerned to show is how the reconceptualisation of the role of law in government proposed in that dialogue has important implications for what we take the role of the institution of office-holding to be. While Greek political tradition held the main aim of (...) archē to be the formal circumscription and control of official power within a constitutional order, Plato insists that it should primarily be understood as ensuring that the exercise of political power approximates, by means of law, the ideal rule by a political expert. The article ends by pointing out how this reading complements another recent discussion of office-holding in the Statesman. (shrink)

The political outlook of the so-called ‘Anonymus Iamblichi’ has been a subject of controversy in the scholarly literature, with some commentators judging him to be a committed democrat, while others see in him a partisan of aristocracy or even oligarchy. This disagreement is not surprising, for the text contains passages that seem to pull in opposite directions. The article suggests that we move beyond the one-dimensional oligarch-or-democrat model traditionally employed and instead approach the issue from a fresh angle, applying the (...) more nuanced typology for understanding ancient social criticism developed by Josiah Ober.The article begins by situating the author within the social landscape of classical Athens. The resulting characterization presents the author as a representative of a distinctive social type: the Athenian ‘rich quietist’, who prefers private economic endeavours to engagement in public affairs, and who is ideologically committed to democratic government but also highly critical of how democratic society treats its wealthy citizens. This characterization helps make better sense of the seemingly contradictory political indications in the text. In particular, fragments 6 and 7 should be read as a single line of argument, which takes the form of an ‘immanent critique’ of contemporary Athenian democracy. According to ‘Anonymus Iamblichi’, democratic society rightly praises the rule of law as a distinctive democratic value, but it falls disappointingly short of that ideal in its treatment of its own wealthy elite citizens. (shrink)

The paper challenges the traditional assumption that the fragments of ‘Anonymus Iamblichi’ are best understood and interpreted against the intellectual and cultural background of the so-called ‘sophistic movement’. I begin by suggesting that we can distinguish, in the fragments, between two separate ‘discourses’ concerning nomos and its role in human life: an abstract ‘sophistic’ discourse, centered around the defense of nomos against the antinomian champions of natural pleonexia, and another, less abstract and more polemical discourse on nomos, which is aimed (...) at the author’s contemporary Athens. I argue that the author’s engagement with well-known sophistic ideas is best understood as instrumental to his polemical agenda: it provides him with a powerful intellectual framework in which to articulate his criticism of democratic society, especially with regard to traditional notions of social ‘benefaction’ and the relation between rich and poor. (shrink)

The National Science Foundation's Second Merit Criterion, or Broader Impacts Criterion , was introduced in 1997 as the result of an earlier Congressional movement to enhance the accountability and responsibility as well as the effectiveness of federally funded projects. We demonstrate that a robust understanding and appreciation of NSF BIC argues for a broader conception of research ethics in the sciences than is currently offered in Responsible Conduct of Research training. This essay advocates augmenting RCR education with training regarding broader (...) impacts. We demonstrate that enhancing research ethics training in this way provides a more comprehensive understanding of the ethics relevant to scientific research and prepares scientists to think not only in terms of responsibly conducted science, but also of the role of science in responding to identified social needs and in adhering to principles of social justice. As universities respond to the mandate from America COMPETES to “provide training and oversight in the responsible and ethical conduct of research”, we urge institutions to embrace a more adequate conception of research ethics, what we call the Ethical Dimensions of Scientific Research, that addresses the full range of ethical issues relevant to scientific inquiry, including ethical issues related to the broader impacts of scientific research and practice. (shrink)

There are many things that can be done to educate young people about controversial topics - including historical monuments - in schools. At the same time, however, we argue that there is little warrant for optimism concerning the educational potential of classroom instruction given the interpretative frame of the state-approved history curriculum; the onerous institutional constraints under which school teachers must labour; the unusual constellation of talents history teachers must possess; the frequent absence of marginalized voices in these conversations; and (...) finally, the not unlikely indifference - if not outright hostility - expressed by far too many members of the dominant group. For these reasons, we think it best to expand the scope of educational possibilities one is willing to consider. (shrink)

Knowledge claims are abundant in the literature on large language models (LLMs); but can we say that GPT-4 truly "knows" the Earth is round? To address this question, we review standard definitions of knowledge in epistemology and we formalize interpretations applicable to LLMs. In doing so, we identify inconsistencies and gaps in how current NLP research conceptualizes knowledge with respect to epistemological frameworks. Additionally, we conduct a survey of 100 professional philosophers and computer scientists to compare their preferences in knowledge (...) definitions and their views on whether LLMs can really be said to know. Finally, we suggest evaluation protocols for testing knowledge in accordance to the most relevant definitions. (shrink)

The prospect of using memory modifying technologies raises interesting and important normative concerns. We first point out that those developing desirable memory modifying technologies should keep in mind certain technical and user-limitation issues. We next discuss certain normative issues that the use of these technologies can raise such as truthfulness, appropriate moral reaction, self-knowledge, agency, and moral obligations. Finally, we propose that as long as individuals using these technologies do not harm others and themselves in certain ways, and as long (...) as there is no prima facie duty to retain particular memories, it is up to individuals to determine the permissibility of particular uses of these technologies. (shrink)

This study is part of a comprehensive investigation of ethical thinking among male and female physicians and nurses. Nine women physicians with different levels of expertise, working in various wards in paediatric clinics at two of the university hospitals in Norway, narrated 37 stories about their experience of being in ethically difficult care situations. All of the interviewees’ narrations were concerned with problems relating to both action ethics and relation ethics. The main focus was on problems in a relation ethics (...) perspective. The most common themes in an action ethics perspective were overtreatment and withholding treatment. The more experienced physicians reasoned differently from the group of less experienced physicians and they coped with pressure in different ways. The less experienced physicians disclosed their professional experience yet seemed uncertain, while putting on an air of certainty, but the more experienced physicians disclosed both their professional and personal experience of caregiving and they seemed to allow themselves to feel uncertain in their certainty. Both groups emphasized a need for deep discussion between colleagues about their being in ethically difficult care situations. (shrink)

Anthropogenic climate change is arguably one of the biggest problems that confront us today. There is ample evidence that climate change is likely to affect adversely many aspects of life for all people around the world, and that existing solutions such as geoengineering might be too risky and ordinary behavioural and market solutions might not be sufficient to mitigate climate change. In this paper, we consider a new kind of solution to climate change, what we call human engineering, which involves (...) biomedical modifications of humans so that they can mitigate and/or adapt to climate change. We argue that human engineering is potentially less risky than geoengineering and that it could help behavioural and market solutions succeed in mitigating climate change. We also consider some possible ethical concerns regarding human engineering such as its safety, the implications of human engineering for our children and society, and we argue that these concerns can be addressed. Our upshot is that human engineering deserves further consideration in the debate about climate change. (shrink)