Advancements in machine learning have fuelled the popularity of using AI decision algorithms in procedures such as bail hearings, medical diagnoses and recruitment. Academic articles, policy texts, and popularizing books alike warn that such algorithms tend to be opaque: they do not provide explanations for their outcomes. Building on a causal account of transparency and opacity as well as recent work on the value of causal explanation, I formulate a moral concern for opaque algorithms that is yet to receive (...) a systematic treatment in the literature: when such algorithms are used in life-changing decisions, they can obstruct us from effectively shaping our lives according to our goals and preferences, thus undermining our autonomy. I argue that this concern deserves closer attention as it furnishes the call for transparency in algorithmic decision-making with both new tools and new challenges. (shrink)

It is known that big data analytics and AI pose a threat to privacy, and that some of this is due to some kind of “black box problem” in AI. I explain how this becomes a problem in the context of justification for judgments and actions. Furthermore, I suggest distinguishing three kinds of opacity: 1) the subjects do not know what the system does (“shallow opacity”), 2) the analysts do not know what the system does (“standard black box (...) opacity”), or 3) the analysts cannot possibly know what the system might do (“deep opacity”). If the agents, data subjects as well as analytics experts, operate under opacity, then these agents cannot provide justifications for judgments that are necessary to protect privacy, e.g., they cannot give “informed consent”, or guarantee “anonymity.” It follows from these points that agents in big data analytics and AI ofen cannot make the judgments needed to protect privacy. So I conclude that big data analytics makes the privacy problems worse and the remedies less effective. As a positive note, I provide a brief outlook on technical ways to handle this situation. (shrink)

This paper develops a methodological conception of opacity grounded in the structural incompletion of reflective systems. In comparative and decolonial contexts, the demand for mutual legibility often reproduces the same totalizing logic it seeks to resist. Against this tendency, opacity is reframed not as a barrier to understanding but as its generative limit-condition: the recursive point at which a system confronts the non-coincidence that makes coherence possible. Drawing on Bernard Williams’s account of ethical integrity and Enrique Dussel’s notion (...) of exteriority—and distinguishing this structural use of opacity from Édouard Glissant’s political right to opacity—the paper identifies a shared recursive pattern underlying ethical self-limitation and philosophical responsibility. Recursion is developed as a formal operation through which linguistic, cognitive, ethical, and comparative systems sustain identity by incorporating their own remainder. On this view, the decolonial condition reflects philosophy’s exposure to its own internal exteriority: the recognition that universality and plurality are structurally intertwined. The result is a set of methodological principles—non-coincidence, integrity, and recursion—offered as a framework for comparative and world philosophy that preserves relation without transparency and universality without totalization. (shrink)

The Axiomatic Opacity Constraint (AOC) states that no agent embedded in a causally closed system operating under linear temporal ordering can achieve infallible epistemic access to its own future states while retaining the capacity for causal intervention. Previously established as a corollary of the Fixed-Point Paradox Theorem, the AOC merits independent examination. We prove that the AOC is axiomatic in the strict sense: derivable from temporal linearity, informational closure, and non-contradiction alone, prior to any empirical commitments. We establish formal (...) connections with Gödel incompleteness, Turing undecidability, and Lawvere's categorical fixed-point theorem, situating the AOC as a fourth member of the self-reference family. We derive the Probability Corollary: no contingent future event can receive credence one for any embedded agent retaining causal agency. We prove the Optimal Epistemic Position Theorem: deliberative efficiency DE = ρ·Ω (reason-responsiveness times residual epistemic uncertainty) is maximized at an interior noise level σ* = γσ_R, with maximum efficiency DE* = (1−φ²)/4. Neither certainty nor randomness is optimal; the wisest agent operates at the equipartition point. We extend the AOC to multi-agent systems, derive structural consequences for communication and moral assessment, and establish that the first-person perspectival asymmetry is a formally necessary feature of any agent retaining causal agency rather than a contingent limitation. The epistemology of perspective is not incidental to agency but constitutive of it. (shrink)

Federated learning (FL) is a machine learning (ML) approach that allows multiple devices or institutions to collaboratively train an ML model without sharing their local data with a third-party. It has recently received significant attention as a promising way to overcome longstanding ethical obstacles to training medical ML models with patient health data. This paper examines the promise of FL in healthcare from an ethical perspective. It argues that medical FL generates a new variety of opacity – federation (...) class='Hi'>opacity, wherein stakeholders cannot access, analyze, or curate the data on which a model has been trained – which (a) presents distinctive ethical challenges concerning institutional fairness and accountability in medical ML; and (b) makes FL models especially vulnerable to data poisoning attacks. It then identifies several key claims about the expected benefits of FL in healthcare and argues that they may be either exaggerated, misleading, or incomplete – often due to the problem of federation opacity. (shrink)

Referential opacity is the failure of substitutivity of identity (SI, for short) and in Quine’s view of existential generalization (EG, for short) as well. Quine thinks that its “solution” in epistemic and doxastic contexts, which relies on the notion of exportation, leads to undesirable results. But epistemic logicians such as Jaakko Hintikka and Wolfgang Lenzen provide another solution based on a different diagnosis: opacity is not, as in Quine’s view, due to the absence of reference, it is rather (...) due to the plurality of references; therefore, one must stabilize the reference to restore SI and EG. However, Hintikka’s semantics remains classical in its treatment of existence, which makes EG non-applicable to possible objects, while in Lenzen’s system, EG is applicable by means of a particular quantifier. But this requires adding the predicate of existence to account for real objects. In this paper, I will show the advantages and disadvantages of both solutions and will end by providing an alternative approach to the problem of non-existent objects, which stays into the frame of possible worlds semantics, but introduces some more restrictions, eliminates the problematic predicate of existence, and applies a neutral quantifier to possible non-existent objects. (shrink)

Computer simulations are often claimed to be opaque and thus to lack transparency. But what exactly is the opacity of simulations? This paper aims to answer that question by proposing an explication of opacity. Such an explication is needed, I argue, because the pioneering definition of opacity by P. Humphreys and a recent elaboration by Durán and Formanek are too narrow. While it is true that simulations are opaque in that they include too many computations and thus (...) cannot be checked by hand, this doesn’t exhaust what we might want to call the opacity of simulations. I thus make a fresh start with the natural idea that the opacity of a method is its disposition to resist knowledge and understanding. I draw on recent work on understanding and elaborate the idea by a systematic investigation into what type of knowledge and what type of understanding are required if opacity is to be avoided and why the required sort of understanding, in particular, is difficult to achieve. My proposal is that a method is opaque to the degree that it’s difficult for humans to know and to understand why its outcomes arise. This proposal allows for a comparison between different methods regarding opacity. It further refers to a kind of epistemic access that is important in scientific work with simulations. (shrink)

ABSTRACT This paper offers an account of institutional testimonial justice and describes one way that it breaks down, which we call institutional opacity. An institution is opaque when it becomes resistant to epistemic evaluation and understanding by its agents and users. When one cannot understand the inner workings of an institution, it becomes difficult to know how to comport oneself testimonially. We offer an account of an institutional ethos to explain what it means for an institution to be testimonially (...) just; we then describe how an ethos of institutional testimonial justice can break down when the institution becomes opaque. An opaque institution is especially problematic for individuals and groups already rendered epistemically vulnerable during their interactions with that institution, which we call epistemically vulnerabilised individuals. We articulate the features of an encounter between an epistemically vulnerabilised individual and an opaque institution. We end by tracing ameliorative strategies that could help repair a deteriorated institutional ethos of testimonial justice. (shrink)

Philosophers have recently focused on critical, epistemological challenges that arise from the opacity of deep neural networks. One might conclude from this literature that doing good science with opaque models is exceptionally challenging, if not impossible. Yet, this is hard to square with the recent boom in optimism for AI in science alongside a flood of recent scientific breakthroughs driven by AI methods. In this paper, I argue that the disconnect between philosophical pessimism and scientific optimism is driven by (...) a failure to examine how AI is actually used in science. I show that, in order to understand the epistemic justification for AI-powered breakthroughs, philosophers must examine the role played by deep learning as part of a wider process of discovery. The philosophical distinction between the 'context of discovery' and the 'context of justification' is helpful in this regard. I demonstrate the importance of attending to this distinction with two cases drawn from the scientific literature, and show that epistemic opacity need not diminish AI's capacity to lead scientists to significant and justifiable breakthroughs. (shrink)

This article explores the epistemic dynamics of intimacy through a close reading of Clarice Lispector’s Água Viva. Beginning with an illustration of intimacy as it emerges in the book, I then build on the hypothesis that intimacy – a relational mode often associated with sharing from the “innermost” – involves a desire by the self to know, or understand, both the other and itself. This dual yearning for knowledge makes intimacy a fertile ground for rethinking the epistemic relation between self (...) and other. Reading the contradictions and tensions of the desire for knowledge in Água Viva, I demonstrate how the book can help us reimagine some of the hegemonic imaginaries regarding ways of knowing in intimacy. Building on feminist scholars such as Luce Irigaray and Vrinda Dalmiya, I align my intervention with efforts to resist acts of grasping or reducing the other when it comes to knowing and relating intimately. Drawing on Édouard Glissant’s theorization of opacity, I argue that Água Viva reimagines the epistemic relation in intimacy in ways that are less discovery- and transparency-oriented. Through a kaleidoscopic interplay of contradictions, the text offers a way of understanding intimacy that resists appropriation and embraces the shifting, unfixed nature of relational knowledge. While Lispector’s book proposes the idea of opacity as a condition for intimacy, it also makes us wonder: if we are trying to deal with the potential of epistemic violence in intimacy, what can come after the paradoxical and troubling yearning for both clarity and opacity? Água Viva, I suggest, invites us to address this question and speculate about alternative ways of knowing in intimacy. (shrink)

This paper critically examines a family of arguments by analogy which suggest that the trust granted to an AI system should mirror the one usually granted by a layperson to a human expert. I particularly dispute the idea that, on the grounds that both share some degree of opacity, human experts and AI systems can be considered epistemic ‘black boxes’ and both be subject to the same blind trust on the part of non-experts. To uncover the problematic nature of (...) this rather widespread analogy, I proceed by identifying a form of ambivalence in the notion of opacity mobilised, as well as a number of highly charged presuppositions concerning expertise (notably relating to a kind of obsession with what is sometimes dubbed its ‘veritistic’ character). I suggest that such a reductionist, monomaniacal conception of expertise is flawed, negligent or inadequate. The other forgotten facets of expertise are not merely cosmetic, but constitutive of it. I show that artificial systems cannot instantiate them, and that we cannot expect them to ever do so. (shrink)

This paper explores the view that the vocabulary of metaphysical fundamentality is opaque, using Sider’s theory of structure as a motivating case study throughout. Two conceptions of fundamentality are distinguished, only one of which can explain why the vocabulary of fundamentality is opaque.

Do we have introspective access to our own thoughts? Peter Carruthers challenges the consensus that we do: he argues that access to our own thoughts is always interpretive, grounded in perceptual awareness and sensory imagery. He proposes a bold new theory of self-knowledge, with radical implications for understanding of consciousness and agency.

Abstract (English) — This paper examines the role of opacity in the thought of Bernard Williams and Enrique Dussel, and develops a philosophical method oriented toward non-totalization. In Williams’s ethics, opacity marks the internal limits of moral justification: the individual cannot and should not make all aspects of their ethical life transparent to others or to themselves. Integrity, for Williams, resides not in public coherence but in the lived coherence of one’s commitments, which remain partially inarticulable. In contrast, (...) Dussel identifies opacity at the structural level—as the condition of exteriority that totalizing systems must exclude in order to sustain their coherence. Through his concept of analectics, Dussel maintains that this exclusion is not accidental but constitutive: the Other is not simply marginalized, but rendered epistemically invisible. By placing these two accounts into dialogue, the paper argues that opacity should be treated not as a failure of knowledge or clarity, but as a methodological principle. This principle finds further support in recursive systems, where self-reference generates non-coincidence from within. The resulting framework affirms opacity as a condition of ethical, structural, and conceptual integrity. Rather than seeking philosophical totality, the method outlined here sustains the limits of systems as an active site of reflection. Opacity, in this sense, becomes not what philosophy must overcome, but what it must think with. -/- Resumen (Español) — Este artículo examina el papel de la opacidad en el pensamiento de Bernard Williams y Enrique Dussel, y desarrolla un método filosófico orientado hacia la no-totalización. En la ética de Williams, la opacidad marca los límites internos de la justificación moral: el individuo no puede ni debe hacer completamente transparentes todos los aspectos de su vida ética, ni para los demás ni para sí mismo. La integridad, según Williams, no reside en la coherencia pública, sino en la coherencia vivida de los propios compromisos, que permanecen en parte inarticulables. En contraste, Dussel identifica la opacidad en un nivel estructural, como la condición de exterioridad que los sistemas totalizantes deben excluir para sostener su coherencia. A través de su concepto de analéctica, Dussel sostiene que esta exclusión no es accidental sino constitutiva: el Otro no es simplemente marginado, sino hecho epistemológicamente invisible. Al poner en diálogo estas dos concepciones, el artículo argumenta que la opacidad debe entenderse no como un fallo del conocimiento o de la claridad, sino como un principio metodológico. Este principio recibe apoyo adicional en sistemas recursivos, donde la autorreferencia genera no coincidencia desde dentro. El marco resultante afirma la opacidad como una condición de integridad ética, estructural y conceptual. En lugar de buscar la totalidad filosófica, el método aquí propuesto sostiene los límites de los sistemas como un sitio activo de reflexión. La opacidad, en este sentido, no es lo que la filosofía debe superar, sino aquello con lo que debe pensar. (shrink)

Nietzsche believes that we do not know our own actions, nor their real motives. This belief, however, is but a consequence of his assuming a quite general skepticism about introspection. The main aim of this paper is to offer a reading of this last view, which I shall call the Inner Opacity (IO) view. In the first part of the paper I show that a strong motivation behind IO lies in Nietzsche’s claim that self-knowledge exploits the same set of (...) cognitive capacities as well as the same folk-psychological framework involved in outward-directed mind-reading. In the second part I turn to Nietzsche’s view of agency and argue that he sees a fundamental discrepancy between the conscious attitudes we have introspective access to, on the one hand, and the subpersonal processes and states occurring at the unconscious level of the drives, on the other hand. (shrink)

This is a reply to commentators in the Journal of the Philosophy of Sport's special issue symposium on GAMES: AGENCY AS ART. I respond to criticisms concerning the value of achievement play and striving play, the transparency and opacity of play, the artistic status of games, and many more.

This article argues for a redefinition of difficulty in relation to the inextricable violence of modernity and examines the consecutive challenge to notions of understanding and interpretation — of a text, of language or of the other — that this repositioning brings. To this end, the article offers a nuanced rereading of Steiner’s canonical fourfold categorization of difficulty, in dialogue with, first, Édouard Glissant’s opacity and, second, Jonathan Pugh and David Chandler’s theorizations of ‘abyssal thought’, an approach emerging from (...) Caribbean and critical Black studies, exploring the key challenge it poses to forging an unmasterful, after Julietta Singh, comparative literary practice. With an equal attention to the theoretical and pedagogical dimensions of reading difficulty, the article emphasizes the importance of the accretive and embodied nature of the reading process and offers a repositioning of textual difficulty, as a meeting site of opacities, and a reformulation of a difficult, unmasterful reading practice, after catastrophe and with others. (shrink)

Phonological opacity and paradigmatic effects (“synchronic analogy”) have long been of interest in relation to change, naturalness, and the phonology/morphology interface. Their investigation has now acquired a new urgency, because they call into question OT’s postulate that constraints are evaluated in parallel. Conceptually, parallelism is one of the basic and most interesting tenets of OT, and so there are good methodological reasons to try hard to save it in the face of such recalcitrant data. The price to be paid (...) for it is the introduction of otherwise unneeded powerful new types of Faithfulness constraints, such as Output/Output (O/O) con- straints, Paradigm Uniformity constraints, and Sympathy constraints, which have turned out to compromise the OT program very severely. (shrink)

Modern computer systems are ubiquitous in contemporary life yet many of them remain opaque. This poses significant challenges in domains where desiderata such as fairness or accountability are crucial. We suggest that the best strategy for achieving system transparency varies depending on the specific source of opacity prevalent in a given context. Synthesizing and extending existing discussions, we propose a taxonomy consisting of eight sources of opacity that fall into three main categories: architectural, analytical, and socio-technical. For each (...) source, we provide initial suggestions as to how to address the resulting opacity in practice. The taxonomy provides a starting point for requirements engineers and other practitioners to understand contextually prevalent sources of opacity, and to select or develop appropriate strategies for overcoming them. (shrink)

In ‘Bureaucratic respectful equality’, Christopher Nathan puts forward two challenges for the author’s claim that basic equality can be grounded in a form of ‘opacity respect’ appropriately shown b...

The paper offers a critical examination of a prominent, “quasi-deflationist” argument advanced in the contemporary debate on the semantic paradoxes against non-naive and non-transparent theories of truth. The argument claims that truth unrestrictedly fulfils certain expressive functions, and that its so doing requires the unrestricted validity of naivety and transparency principles. The paper criticises the quasi-deflationist argument by considering some kinds of cases in which transparency and naivety arguably fail. In some such cases truth still fulfils the relevant expressive functions (...) without being transparent or naive; in some other such cases, truth does not fulfil the relevant expressive functions and other conceptual resources must be called upon. Thus, in different ways, all such cases belie the quasi-deflationist argument’s insistence that naivety and transparency should be unrestrictedly valid for truth unrestrictedly to fulfil the relevant expressive functions. There might however be other reasons for solving the semantic paradoxes by revising classical logic, and the paper in effect closes by offering versions of the liar paradox that rely on compelling but opacity-friendly truth-theoretic principles. (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)

It is well known that experts’ opinion and testimony take on a decisive weight in judicial fact-finding, raising issues and perplexities that have long been under scholarly scrutiny. In this paper I argue that expert’s opinions have a much wider impact on legal decision-making. In particular, they may generate a problem that I will call ‘the opacity of law’. A legal text, such as a statute or regulation, becomes opaque if a legal authority is not able to grasp its (...) full linguistic content but is nevertheless in a position to use it, thanks to an expert’ opinion, in legal decision-making. When this occurs, not only do experts contribute to fact-finding but also to determining the content of the law. In the paper I analyse the linguistic and cognitive sources of this phenomenon, its characteristics and troublesome consequences, and the different kinds of opacity that may affect legal decision-making. (shrink)

This chapter offers a new framework for approaching the notions of linguistic density and compression so central to discussions of the nature of poetic language. For this chapter, semantic density and like features of poetic language turn out to be forms of poetic opacity. Indeed, the particular forms of opacity that poetry is apt to produce are the ‘the mark of the poetic’, its distinguishing feature and that upon which a definition, such as poetry will admit of, can (...) be constructed. The chapter identifies various forms of poetic opacity and explains their interrelation and aesthetic significance. The chapter concludes by drawing a striking analogy between our experience of poetic opacity and Richard Wollheim’s notion of the ‘twofoldness’ of representational painting. (shrink)

Kant’s moral philosophy both enjoins the acquisition of self-knowledge as a duty, and precludes certain forms of its acquisition via what has become known as the Opacity Thesis. This article looks at several recent attempts to solve this difficulty and argues that they are inadequate. I argue instead that the Opacity Thesis rules out only the knowledge that one has acted from genuine moral principles, but does not apply in cases of moral failure. The duty of moral self-knowledge (...) applies therefore only to one’s awareness of one’s status as a moral being and to the knowledge of one’s moral failings, both in particular actions and one’s overall character failings, one’s vices. This kind of knowledge is morally salutary as an aid to discovering one’s individual moral weakness as well as the subjective ends for which one acts, and in this way for taking up the morally required end of treating human beings as human beings. In this way, moral self-knowledge can be understood as a necessary element of moral improvement, and I conclude by suggesting several ways to understand it thereby as genuinely primary among the duties to oneself. (shrink)

It has been widely accepted that Kant holds the “Opacity Thesis,” the claim that we cannot know the ultimate grounds of our actions. Understood in this way, I shall argue, the Opacity Thesis is at odds with Kant's account of practical self-consciousness, according to which I act from the (always potentially conscious) representation of principles of action and that, in particular, in acting from duty I act in consciousness of the moral law's determination of my will. The (...) class='Hi'>Opacity Thesis thus threatens to render acting from duty unintelligible. To diffuse the threat, I argue, first, that we need not attribute the Opacity Thesis to Kant. Kant's concern with the ubiquity of moral self-opacity does not imply the strong skeptical conclusion that knowledge of the grounds of one's action is impossible. Second, I show how moral self-opacity in cases of morally bad action emerges from the intrinsic inability of representing oneself, insofar as one is pursuing the indeterminate end of “happiness.” Thus, moral self-opacity does not undermine the will's self-consciousness but is born of it. (shrink)

Human opacity is the intrinsic quality of unknowabil- ity of human beings with respect to machines. The descriptive rela- tionship between humans and machines, which captures how much information one can gather about the other, can be explicated using an opacity-transparency relationship. This relationship allows us to describe and normatively evaluate a spectrum of opacity where humans and machines may be either opaque or transparent. In this paper, we argue that the advent of Affective Computing (AC) has (...) begun to shift the ideal position of humans on this spectrum towards greater transparency, while much of this technology is shifting towards opacity. We explore the implications of this shift with regard to the affective information of humans and how the threat to human opacity by AC systems has various adverse repercussions, such as infringement of one’s autonomy, deception, manipulation, and increased anxiety. There are also distributive consequences that expose vulnerable groups to unjustified burdens and reduce them to mere profiles. We further provide an assessment of current AC technology, which follows the descriptive relationship between humans and machines from the lens of opacity and transparency. Finally, we foresee and address three possible objections to our claims. These are the beneficence of AC systems, their relation to privacy, and their restrictive capacity to capture human affects. Through these arguments, the paper aims to bring attention to the ontological relationship between humans and machines from the perspective of opacity and transparency while emphasizing on the gravity of the ethical concerns raised by their threat to human opacity. (shrink)

“It is as if I have a brick on my stomach”—our stomach aches and we need to use an ‘as if’ to explain others what our experience feels like. This metaphoric language is necessary for us to describe the quality of the pain we experience. The bioethicist Engelhardt reminds us how our body is often opaque to us (1973). The opacity and recalcitrance of the body is comparable to the opacity and autonomous life of the algorithms that we (...) use to organize our society. As they are becoming more and more the body of our collective lives, committing to make an algorithm as translucent as possible might be as challenging as making human bodies as readable as possible. In the same way as we need to know why our stomach hurts if we want to cure it, so we need to read through the opacity of an algorithm if we want to fix the malfunctioning of certain technology in our lives. In this paper I will examine how phenomenology and its metaphorical language can help the digital language of decision-making algorithms, especially in medical humanities, to comply to the ethical commitment to transparency. Namely, I will use phenomenology to shed light on the co-constitutive triad of user-designer-system in decision-making algorithms. (shrink)

Philosophy and Phenomenological Research, EarlyView.

What is narrative? What is distinctive about the great literary narratives? In virtue of what is a narrative fictional or non-fictional? In this important new book Peter Lamarque, one of the leading philosophers of literature at work today, explores these and related questions to bring new clarity and insight to debates about narrative in philosophy, critical theory, and narratology.

This paper looks at the debates between internalism and externalism in mind and epistemology. In each realm, internalists face what we call 'The Connection Problem', while externalists face what we call 'The Problem of Opacity'. We offer an integrated account of thought content and epistemic warrant that overcomes the problems. We then apply the framework to debates between internalists and externalists in metaethics.

Deep neural networks have become increasingly successful in applications from biology to cosmology to social science. Trained DNNs, moreover, correspond to models that ideally allow the prediction of new phenomena. Building in part on the literature on ‘eXplainable AI’, I here argue that these models are instrumental in a sense that makes them non-explanatory, and that their automated generation is opaque in a unique way. This combination implies the possibility of an unprecedented gap between discovery and explanation: When unsupervised models (...) are successfully used in exploratory contexts, scientists face a whole new challenge in forming the concepts required for understanding underlying mechanisms. (shrink)

While Deep Neural Networks (DNNs) can perform a wide range of tasks at human or greater-than-human level of competence, they are also notoriously opaque. This paper aims to shed light on both the specific nature of this opacity and what it would take to fully or partially remove it. We begin by drawing a clarificatory distinction between two basic dimensions of opacity of complex systems – internal and relational – and explain how various kinds of opacity invoked (...) in recent discussions of DNNs can be understood in terms of these basic dimensions. We then discuss the prospects of removing opacity from DNNs following the methods of two subfields of research in AI – mechanistic interpretability and explainable artificial intelligence (XAI). We finish with a critical discussion of the relevance of what Sullivan (2022a) calls ‘link uncertainty’. (shrink)

We explore the view that Frege's puzzle is a source of straightforward counterexamples to Leibniz's law. Taking this seriously requires us to revise the classical logic of quantifiers and identity; we work out the options, in the context of higher-order logic. The logics we arrive at provide the resources for a straightforward semantics of attitude reports that is consistent with the Millian thesis that the meaning of a name is just the thing it stands for. We provide models to show (...) that some of these logics are non-degenerate. (shrink)

This essay explicates Édouard Glissant’s aesthetics of opacity in terms of its formation and significance. This theory comes into form in the historical condition of colonial alterity. In The Poetics of Relation, Glissant extrapolates opacity as the fundamental of aesthetics from such linguistic activities as creole languages and improvised stories found in the Caribbean islands. More than a postcolonial defense of identity alterity, opacity denotes the linguistic expression of material alterity. It means an involuntary flourishing of linguistically (...) enhanced dynamic of exchange, connection, and making in the landscapes of compelling affordances. Such languages cannot be reduced to texts because they are derived from the inevitably alien ground called "the other of Thought,” or a recognition and practice of radical difference. The significance of the aesthetics of opacity lies in that, Glissant asserts, humans can linguistically express the engagement with material ecologies while avoiding the authoritative domination of reason. (shrink)

Key issues in epistemology for the most part have to do with epistemic values such as justification, truth, and knowledge—that is, values related to the epistemic status of our propositional attitudes, mental events, and states. However, another important issue that is worth examining is the extent to which a subject is in a position to evaluate the strength of her epistemic position. In this paper, I wish to emphasize two properties of our mental states that play a decisive part in (...) that respect: their opacity and transparency. In the following, I will assume that a mental state is opaque whenever it presents itself with no underlying reason, whereas states that are supported by apparent reasons are transparent. One main argument that I will defend is that even when the opacity and transparency of our mental states are not reliable cues, still they remain highly informative. Notably, I draw some implications relative to the externalist/internalist debate in epistemology. I examine the claim that only mental states that are justified are well grounded or justification conferring and the idea that often stands behind, that is, that only states whose reasons are accessible are justified—namely, the Accessibility Requirement. I also examine the source of a recent debate in epistemology: the epistemic status of our intuitive states, as these are perfect instances of opaque mental states. I conclude that intuitive states in some respect are less misleading than states that are supported by apparent reasons. (shrink)

There seems to be a lot of opacity in our language. Quotation is opaque. The modal idioms are apparently opaque. Propositional attitude ascriptions seem opaque, as do the environments created by verbs such as ‘seeks’ and ‘fears’. Opacity raises a number of issues — first and foremost, whether there is such a thing. This article concentrates on the question of whether there is any opacity to be found in natural language, examining various reasons one might have for (...) denying that apparent opacity is genuine. (shrink)

The main goal of this paper is to defend the thesis that the content of know-how states is an accuracy assessable type of nonconceptual content. My argument proceeds in two stages. I argue, first, that the intellectualist distinction between types of ways of grasping the same kind of content is uninformative unless it is tied in with a distinction between kinds of contents. Second, I consider and reject the objection that, if the content of know-how states is non-conceptual, it will (...) be mysterious why attributions of knowing how create opaque contexts. I show that the objection conflates two distinct issues: the nature of the content of know-how states and the semantic evaluability of know-how ascriptions. (shrink)

The aim of this literature review was to compose a narrative review supported by a systematic approach to critically identify and examine concerns about accountability and the allocation of responsibility and legal liability as applied to the clinician and the technologist as applied the use of opaque AI-powered systems in clinical decision making. This review questions if it is permissible for a clinician to use an opaque AI system in clinical decision making and if a patient was harmed as a (...) result of using a clinician using an AIS’s suggestion, how would responsibility and legal liability be allocated? Literature was systematically searched, retrieved, and reviewed from nine databases, which also included items from three clinical professional regulators, as well as relevant grey literature from governmental and non-governmental organisations. This literature was subjected to inclusion/exclusion criteria; those items found relevant to this review underwent data extraction. This review found that there are multiple concerns about opacity, accountability, responsibility and liability when considering the stakeholders of technologists and clinicians in the creation and use of AIS in clinical decision making. Accountability is challenged when the AIS used is opaque, and allocation of responsibility is somewhat unclear. Legal analysis would help stakeholders to understand their obligations and prepare should an undesirable scenario of patient harm eventuate when AIS were used. (shrink)

The problem discussed here is to find a basis for a uniform treatment of the relation between pronouns and their antecedents, taking into account both linguists' and philosophers' approaches. The two main candidates would appear to be the linguists' notion of coreference and the philosophers' notion of pronouns as variables. The notion of coreference can be extended to many but not all cases where the antecedent is non-referential. The pronouns-as-variables approach appears to come closer to full generality, but there are (...) some examples of pronouns of laziness which appear to resist either of the two approaches. (shrink)

Artificial intelligent (AI) systems used in medicine are often very reliable and accurate, but at the price of their being increasingly opaque. This raises the question whether a system’s opacity undermines the ability of medical doctors to acquire knowledge on the basis of its outputs. We investigate this question by focusing on a case in which a patient’s risk of recurring breast cancer is predicted by an opaque AI system. We argue that, given the system’s opacity, as well (...) as the possibility of malfunctioning AI systems, practitioners’ inability to check the correctness of their outputs, and the high stakes of such cases, the knowledge of medical practitioners is indeed undermined. They are lucky to form true beliefs based on the AI systems’ outputs, and knowledge is incompatible with luck. We supplement this claim with a specific version of the safety condition on knowledge, Safety*. We argue that, relative to the perspective of the medical doctor in our example case, his relevant beliefs could easily be false, and this despite his evidence that the AI system functions reliably. Assuming that Safety* is necessary for knowledge, the practitioner therefore doesn’t know. We address three objections to our proposal before turning to practical suggestions for improving the epistemic situation of medical doctors. (shrink)