This work consists of two parts. Part I will be a contribution to a philo- sophical discussion of the nature of causal explanation. It will present my contrastive counterfactual theory of causal explanation and show how it can be used to deal with a number of problems facing theories of causal explanation. Part II is a contribution to a discussion of the na- ture of interest explanation in social studies of science. The aim is to help to resolve some controversies (...) concerning interest explanation by explicating the concept of interest and its explanatory uses by using the account of explanation developed in Part I. (shrink)

Pluralism with respect to the structure of explanations of facts is not uncommon. Wesley Salmon, for instance, distinguished two types of explanation: causal explanations (which provide insight in the causes of the fact we want to explain) and unification explanations (which fit the explanandum into a unified world view). The pluralism which Salmon and others have defended is compatible with several positions about the exact relation between these two types of explanations. We distinguish four such positions, and argue in favour (...) of one of them. We also compare our results with the views of some authors who have recently written on this subject. (shrink)

When is a cause of a cause of an effect also a cause of that effect? The right answer is either Sometimes or Always . In favour of Always , transitivity is considered by some to be necessary for distinguishing causes from redundant non-causal events. Moreover transitivity may be motivated by an interest in an unselective notion of causation, untroubled by principles of invidious discrimination. And causal relations appear to add up like transitive relations, so that the obtaining of the (...) overarching relation is not independent of the obtaining of the intermediaries. On the other hand, in favour of Sometimes , often we seem not to treat events that are very spatiotemporally remote from an effect as its causes, even when connected to the effect in question by a chain of counterfactual or chance-raising dependence. Moreover cases of double prevention provide counterexamples to causal transitivity even over short chains. According to the argument of this paper, causation is non-transitive. Transitizing causation provides no viable account of causal redundancy. An unselective approach to causation may motivate resisting the distance counterexamples to transitivity, but it does not help with double prevention, and even makes it more intractable. The strongest point in favour of transitivity is the adding up of causal relations, and this is the point that extant non-transitizing analyses have not adequately addressed. I propose a necessary condition on causation that explains the adding up phenomenon. In doing so it also provides a unifying explanation of distance and double prevention counterexamples to transitivity. (shrink)

In the literature on scientific explanation, there is a classical distinction between explanations of facts and explanations of laws. This paper is about explanations of facts. Our aim is to analyse the role of unification in explanations of this kind. We discuss five positions with respect to this role, argue for two of them and refute the three others.

What conception of causation is at work in the law? I argue that the law implicitly relies on a contrastive conception. In a liability case where the defendant's breach of duty must be shown to have caused the plaintiff's damages, it is not enough to consider what would have happened if the cause had not occurredthe law requires us to look to a specific replacement for the effect, which in this case is the hypothetical outcome in which the plaintiff came (...) off better. In place of I suggest the more explicit An explicitly contrastive approach can thus potentially help the lawyer phrase her causal question in a more explicit way, while shedding light on our conception of causation. (shrink)

Philosophers of science have often favoured reductive approaches to how-possibly explanation. This article identifies three alternative conceptions making how-possibly explanation an interesting phenomenon in its own right. The first variety approaches “how possibly X?” by showing that X is not epistemically impossible. This can sometimes be achieved by removing misunderstandings concerning the implications of one’s current belief system but involves characteristically a modification of this belief system so that acceptance of X does not result in contradiction. The second variety offers (...) a potential how-explanation of X. It is usually followed by a range of further potential how-explanations of the same phenomenon. In recent literature the factual claims implied by the second variety have been downplayed whereas the heuristic role of mapping the space of conceptual possibilities has been emphasized. I will focus especially on this truth-bracketing sense of potentiality when looking closer at the second variety in the paper. The third variety has attracted less interest. It presents a partial how-explanation of X. Typically it aims to establish the existence of a mechanism by which X could be and was generated. The third conception stands out as the natural alternative for the advocate of ontic how-possibly explanations. This article transfers Salmon’s view that explanation-concepts can be broadly divided into epistemic, modal, and ontic to the context of how-possibly explanations. Moreover, it is argued that each of the three above-mentioned varieties of how-possibly explanation occurs in science. To recognize this may be especially relevant for philosophers. We are often misled by the promises of various why-explanation accounts, and seem to have forgotten nearly everything about the diversity of how-possibly explanations. (shrink)

To operationalize the methodological assessment of evolutionary psychology, three requirements are proposed that, if satisfied, would show that a hypothesis is not a just-so story: (1) theoretical entrenchment (i.e., that the hypothesis under consideration is a consequence of a more fundamental theory that is empirically well-confirmed across a very wide range of phenomena), (2) predictive success (i.e., that the hypothesis generates concrete predictions that make it testable and eventually to a certain extent corroborated), and (3) failure of rival explanations (i.e., (...) that crucial and successful predictions attributed to the hypothesis in question cannot be derived from alternative hypotheses). The author argues that the hypothesis about evolutionary sex differences in human jealousy satisfies all three requirements. Key Words: evolutionary psychology • adaptationism • philosophy of science • testability. (shrink)

It it tempting to think that if an outcome had some probability of not occurring, then we cannot explain why that outcome in fact occurred. Despite this intuition, most philosophers of science have come to admit the possibility of indeterministic explanation. Yet some of them continue to hold that if an outcome was not determined, it cannot be explained why that outcome rather than some other occurred. I argue that this is an untenable compromise: if indeterministic explanation is possible, then (...) indeterministic contrastive explanation is possible too. In order to defend this conclusion, I develop an account of contrastive explanation. (shrink)

In a brief passage, David Lewis derives from quantum-theory a dilemma regarding the explanation of chance events which he tries to solve by first distinguishing plain from contrastive why-questions have answers. His brevity warrants elaboration and critique. I endorse his derivation, but I make a structural objection to his solution. Once a further distinction is drawn between different kinds of contrastive why-question, his solution can be modified and refined so as to go some way to meeting this objection. However, it (...) cannot resist a deeper objection to the effect that chance events can be given contrastive explanations after all. The alternative solution to the dilemma that I recommend holds Lewis's conception of the 'essence of chance' to be illusory. (shrink)

I argue that the contrast between models and theories is important for public policy issues. I focus especially on the way a mathematical model explains just one aspect of the data.

How did I raise my arm? The simple answer is that I raised it as a consequence of intending to raise it. A slightly more complicated response would mention the absence of any factors which would inhibit the execution of the intention- and a more complicated one still would specify the intention in terms of a goal (say, drinking a beer) which requires arm-raising as a means towards that end. Whatever the complications, the simple answer appears to be on the (...) right track. (shrink)

Explores the science and creative process behind Poe’s cosmological treatise. Silver Winner for Philosophy, 2017 Foreword INDIES Book of the Year Awards In 1848, almost a year and a half before Edgar Allan Poe died at the age of forty, his book Eureka was published. In it, he weaved together his scientific speculations about the universe with his own literary theory, theology, and philosophy of science. Although Poe himself considered it to be his magnum opus, Eureka has mostly been overlooked (...) or underappreciated, sometimes even to the point of being thought an elaborate hoax. Remarkably, however, in Eureka Poe anticipated at least nine major theories and developments in twentieth-century science, including the Big Bang theory, multiverse theory, and the solution to Olbers’ paradox. In this book—the first devoted specifically to Poe’s science side—David N. Stamos, a philosopher of science, combines scientific background with analysis of Poe’s life and work to highlight the creative and scientific achievements of this text. He examines Poe’s literary theory, theology, and intellectual development, and then compares Poe’s understanding of science with that of scientists and philosophers from his own time to the present. Next, Stamos pieces together and clarifies Poe’s theory of scientific imagination, which he then attempts to update and defend by providing numerous case studies of eureka moments in modern science and by seeking insights from comparative biography and psychology, cognitive science, neuroscience, and evolution. David N. Stamos teaches philosophy at York University in Toronto. He is the author of several books, including Darwin and the Nature of Species, also published by SUNY Press. (shrink)

Most psychological research aims to uncover generalizations about the mind that hold across subjects. Philosophical discussions of scientific explanation have focused on such generalizations, but in doing so, have often overlooked an important phenomenon: variation. Variation is ubiquitous in psychology and many other domains, and an important target of explanation in its own right. Here I characterize explananda that concern individual differences and formulate an account of what it takes to explain them. I argue that the notion of actual difference (...) making, the only causal concept in the literature that explicitly addresses variation, cannot be used to ground such an account. Instead, I propose a view on which explaining individual differences involves identifying causes that could be intervened on to reduce the variability in the population. This account provides criteria of success for explaining variation and deepens our understanding of causal explanation. (shrink)

Armstrong famously argued in favour of introducing totality facts in our ontology. Contrary to fully negative (absence) facts, totality facts yield a theory of “moderate” or “partial” negativity, which allegedly provides an elegant solution to the truthmaking problem of negative claims and, at the same time, avoids postulating (many) first-order absences. Friends of totality facts argue that partial negativity is (i) tolerable vis-à-vis the Eleatic principle qua mark of the real, and (ii) achieves a significant advantage in terms of ontological (...) parsimony. But are totality facts, which are partially negative, really more ontologically acceptable than fully negative facts? In this paper, we argue that, comparatively, the case for totality facts is weaker than commonly assumed and that, ultimately, the answer is negative. (shrink)

Historians have claimed that the two closely related concepts of animism and natural teleology were both decisively rejected in the Scientific Revolution. They tout Robert Boyle as an early modern warden against pre-modern animism. Discussing Avicenna, Aquinas, and Buridan, as well as Renaissance psychology, I instead suggest that teleology went through a slow and uneven process of rationalization. As Neoplatonic theology gained influence over Aristotelian natural philosophy, the meaning of animism likewise grew obscure. Boyle, as some historians have shown, exemplifies (...) this uneven process. There is an unresolved tension between his religious convictions and the implicit animism of his empirical practice. (shrink)

Abstract‘Behavioural modernity’ isn’t what it used to be. Once conceived as an integrated package of traits demarcated by a clear archaeological signal in a specific time and place, it is now disparate, archaeologically equivocal, and temporally and spatially spread. In this paper we trace behavioural modernity’s empirical and theoretical developments over the last three decades, as surprising discoveries in the material record, as well the reappraisal of old evidence, drove increasingly sophisticated demographic, social and cultural models of behavioural modernity. We (...) argue, however, that some approaches to identifying and categorizing modernity have not kept up with this new picture. This is due to what we term ‘Rubicon expectations’: classificatory and interpretive practices which look for or assume clear demarcations in behavioural and cultural processes. We develop a philosophical account of ‘investigative disintegration’ to capture how our understanding of behavioural modernity has changed, and how Rubicon-based practices have become inadequate. Disintegration, in the form we analyse, occurs when scientists’ conception of a phenomenon shifts sufficiently to reshape an investigation’s epistemic structure. For behavioural modernity, the explanatory weight which once lay on identifying ‘switch-points’ in the innate suite of hominin cognitive capacities, lies now in understanding the social and demographic environments that were capable of sustaining and nourishing more complex material cultures. Finally, we argue that the phenomenon itself has not disintegrated to the point that we are left with no interesting explanandum: for all its mosaic, disparate nature, there are still good reasons for behavioural modernity to retain its central place in investigation of our species' origins. (shrink)

Everyday life reasoning and argumentation is defeasible and uncertain. I present a probability logic framework to rationally reconstruct everyday life reasoning and argumentation. Coherence in the sense of de Finetti is used as the basic rationality norm. I discuss two basic classes of approaches to construct measures of argument strength. The first class imposes a probabilistic relation between the premises and the conclusion. The second class imposes a deductive relation. I argue for the second class, as the first class is (...) problematic if the arguments involve conditionals. I present a measure of argument strength that allows for dealing explicitly with uncertain conditionals in the premise set. (shrink)

In this BA dissertation, I deploy examples of non-causal explanations of physical phenomena as evidence against the view that causal models of explanation can fully account for explanatory practices in science. I begin by discussing the problems faced by Hempel’s models and the causal models built to replace them. I then offer three everyday examples of non-causal explanation, citing sticks, pilots and apples. I suggest a general form for such explanations, under which they can be phrased as inductive-statistical arguments incorporating (...) plausible assumptions. I then show the applicability of this form to explanatory practices in thermal physics. I explore the possibility that population genetics provides a similar form of explanation, and offer a novel defence of the statistical interpretation of population genetics proposed by Matthen and Ariew (2002). I close with remarks concerning how, when faced with competing causal and non-causal explanations of the same phenomenon, we can perform an Inference to the Best Explanation. (shrink)

Sometimes, learning about the origins of a belief can make it irrational to continue to hold that belief—a phenomenon we call ‘genealogical defeat’. According to explanationist accounts, genealogical defeat occurs when one learns that there is no appropriate explanatory connection between one’s belief and the truth. Flatfooted versions of explanationism have been widely and rightly rejected on the grounds that they would disallow beliefs about the future and other inductively-formed beliefs. After motivating the need for some explanationist account, we raise (...) some problems for recent versions of explanationism. Learning from their failures, we then produce and defend a more resilient explanationism. (shrink)

This essay presents an argument, which it calls the Bias Argument, with the dismaying conclusion that (almost) everyone should significantly reduce her confidence in (too many) philosophical beliefs. More precisely, the argument attempts to show that the most precious philosophical beliefs are biased, as the pervasive and permanent disagreement among the leading experts in philosophy cannot be explained by the differences between their evidence bases and competences. After a short introduction, the premises of the Bias Argument are spelled out in (...) the first part. The second part explains why the objections to the Bias Argument are not compelling. Even though the essay does not adopt the conclusion of the Bias Argument, partly because it seems to be self-defeating, the authors know no plausible way to refute its premises. Thus, the primary aim of the essay is to clarify why the aporetic situation of the Bias Argument arises. (shrink)

For millennia self has been conjectured to be necessary for consciousness. But scant empirical evidence has been adduced to support this hypothesis. Inconsistent explications of “self” and failure to design apt experiments have impeded progress. Advocates of phenomenological psychiatry, however, have helped explicate “self,” and employed it to explain some psychopathological symptoms. In those studies, “self” is understood in a minimalist sense, sheer “for-me-ness.” Unfortunately, explication of the “minimal self” (MS) has relied on conceptual analysis, and applications to psychopathology have (...) been hermeneutic, allowing for many degrees of interpretive latitude. The result is that MS’s current scientific status is analogous to that of the “atom,” at the time when “atom” was just beginning to undergo transformation from a philosophical to a scientific concept. Fortunately, there is now an opportunity to promote a similar transformation for “MS.” Discovery of the brain’s Default Mode Network (DMN) opened the door to neuroimaging investigations of self. Taking the DMN and other forms of intrinsic activity as a starting point, an empirical foothold can be established, one that spurs experimental research and that enables extension of research into multiple phenomena. New experimental protocols that posit “MS” can help explain phenomena hitherto not thought to be related to self, thereby hastening development of a mature science of self. In particular, targeting phenomena wherein consciousness is lost and recovered, as in some cases of Unresponsive Wakefulness Syndrome (UWS), allow for design of neuroimaging probes that enable detection of MS during non-conscious states. These probes, as well as other experimental protocols applied to NREM Sleep, General Anesthesia (GA), and the waking state, provide some evidence to suggest that not only can self and consciousness dissociate, MS might be a necessary precondition for conscious experience. Finally, these findings have implications for the science of consciousness: it has been suggested that “levels of consciousness” (LoC) is not a legitimate concept for the science of consciousness. But because we have the conceptual and methodological tools with which to refine investigations of MS, we have the means to identify a possible foundation—a bifurcation point—for consciousness, as well as the means by which to measure degrees of distance from that foundation. These neuroimaging investigations of MS position us to better assess whether LoC has a role to play in a mature science of consciousness. (shrink)

Can an identity be the proper subject of an explanation? A popular stance, albeit not one often argued for, gives a negative answer to this question. Building from a contentious passage from Jaegwon Kim in this direction, we reconstruct an argument to the conclusion that identities, to the extent in which they are necessary, cannot be explained. The notion of contrastive explanation, characterized as difference-seeking, will be crucial for this argument; however, we will eventually find the argument to be unsatisfactory. (...) On the contrary, the discussion provides enough resource to sketch a very simple framework for a non-causal contrastive explanation of identities. Many instances will be provided, with different varieties of explanans, ultimately suggesting that certain entailment or biconditional principles involving identities (first and foremost, so-called two-level identity criteria) may indeed be taken to have an inherent explanatory value. (shrink)

A well-known objection to Humean accounts of laws charges them with circularity. While the view has it that particular facts explain the natural laws, natural laws are often relied upon in order to explain particular facts. Thus, the Humean is committed to circular explanations—or so goes the argument. In this paper, I review two ways of dealing with the circularity objection against Humean views of laws. Then, by introducing a contrastive treatment of explanations, I put forward a new one, which, (...) if it does not end up dethroning its competitors, I am hoping is still worth exploring. (shrink)

Intellectual progress involves forming a more accurate picture of the world. But it also figuring out which concepts to use for theorizing about the world. Bayesian epistemology has had much to say about the former aspect of our cognitive lives, but little if at all about the latter. I outline a framework for formulating questions about conceptual change in a broadly Bayesian framework. By enriching the resources of Epistemic Utility Theory with a more expansive conception of epistemic value, I offer (...) a picture of our cognitive economy on which adopting new conceptual tools can sometimes be epistemically rational. (shrink)

In this book, I defend the present-centered approach in historiography of science (i.e. study of the history of science), build an account for causal explanations in historiography of science, and show the fruitfulness of the approach and account in when we attempt to understand science. -/- The present-centered approach defines historiography of science as a field that studies the developments that led to the present science. I argue that the choice of the targets of studies in historiography of science should (...) be directly connected to our values and preferences in an intersubjective process. The main advantage of this approach is that it gives a clear motivation for historiography of science and avoids or solves stubborn conceptual and practical problems within the field. -/- The account of causal explanations is built on the notions of counterfactual scenarios and contrastive question-answer pairs. I argue that if and only if we track down patterns of counterfactual dependencies, can we understand history. Moreover, I define the notions of historical explanation, explanatory competition, explanatory depth, and explanatory resources. -/- Finally, I analyze the existing historiography of science with the framework built in the previous chapter, and I show that this framework clarifies many first-order (i.e. concerning the history of science) and meta-level issues (i.e. concerning the nature of science in general) that historians and philosophers tackle. As an illustration of the philosophical power of the framework, I explicate the notion of local explanation and analyze the question of whether the developments of science were necessary or contingent. (shrink)

Recent work on interpretability in machine learning and AI has focused on the building of simplified models that approximate the true criteria used to make decisions. These models are a useful pedagogical device for teaching trained professionals how to predict what decisions will be made by the complex system, and most importantly how the system might break. However, when considering any such model it’s important to remember Box’s maxim that "All models are wrong but some are useful." We focus on (...) the distinction between these models and explanations in philosophy and sociology. These models can be understood as a "do it yourself kit" for explanations, allowing a practitioner to directly answer "what if questions" or generate contrastive explanations without external assistance. Although a valuable ability, giving these models as explanations appears more difficult than necessary, and other forms of explanation may not have the same trade-offs. We contrast the different schools of thought on what makes an explanation, and suggest that machine learning might benefit from viewing the problem more broadly. (shrink)

Corroborative evidence may be understood as having two epistemic effects: a primary effect by which it offers direct evidence for some claim, and a secondary effect by which it bolsters the appraised probative, or evidential, value of some other piece of evidence for that claim. This paper argues that the bolstering effect of corroborative evidence is epistemically legitimate because corroboration provides a reason to count the belief based on the initial evidence as sensitive to, and safe from, defeat in a (...) way that it was not previously recognized to be. Discovering that our initial evidence tracks the truth in a way we previously did not recognize provides a reason to positively reappraise the probative value of that evidence. The final section of the paper relates the proposed sensitivity- and safety-based account of corroboration to an explanation-based account. (shrink)

Robert Kane’s libertarian theory is often attacked on the grounds that undetermined self-forming actions are not amenable to contrastive explanation. I propose that we should understand contrastive explanations in terms of an appeal to structuring causes. Doing so reveals that Kane’s claim that there can be no contrastive explanation for self-forming actions is not an unwanted implication of his appeal to indeterminism, but is actually an implication of the fact that the agent’s will is not yet appropriately structured. I then (...) explain how this can assist Kane with the luck objection. (shrink)

Many recent decisions in tort law attempt to combine two conceptually incommensurable features: a traditional 'but for' test of factual causation, and the scientific or medical evidence that is required to explain how some injury occurred. Even when applied to macroscopic objects, the 'but for' test fails to identify causes, because it merely rephrases in the language of possible worlds what may be inferred from what is inductively known about the actual world. Since scientific theories explain the occurrence of events (...) without reference to causes, this test of causation is not only ineffectual but also inapplicable to scientific or medical evidence. Tort law, therefore, should abandon the 'but for' test; when scientific or medical evidence is required in a tort case, the courts should rely on the explanatory strategies that function within the theories from which they borrow such evidence. (shrink)

We often explain by citing an absence or an omission. Apart from the problem of assigning a causal role to such apparently negative factors as absences and omissions, there is a puzzle as to why only some absences and omissions, out of indefinitely many, should figure in explanations. In this paper we solve this ’many absences problem’ by using the contrastive model of explanation. The contrastive model of explanation is developed by adapting Peter Lipton’s account. What initially appears to be (...) only a trivial amendment to Lipton’s Difference Condition enables us both to offer a much more satisfactory solution to the ’many absences problem’ than David Lewis did, and also to explain why explanation in terms of absences and omissions should be so common. (shrink)

Bayesianism provides a rich theoretical framework, which lends itself rather naturally to the explication of various “contrastive” and “non-contrastive” concepts. In this (brief) discussion, I will focus on issues involving “contrastivism”, as they arise in some of the recent philosophy of science, epistemology, and cognitive science literature surrounding Bayesian confirmation theory.

Traditional accounts of explanation fail to illuminate the explanatory relevance of “models that are descriptively false” in the sense that the regularities they entail fail to obtain. In this paper, I propose an account of explanation, which I call ‘explanation by concretization’, that serves to explicate the explanatory relevance of such models. Starting from a highly abstract and idealized model, causal explanations of the absence of regularities are sought by adding complexity to the model or by concretizing it. Whether this (...) process is successful depends on whether the abstractions and idealizations in the basic model succeed in isolating a mechanism, i.e. in representing how it operates when interfering factors are absent. This account is developed in the context of economics and contrasted to those of Daniel Hausman and Nancy Cartwright. I go on to provide an account of how unrealistic models can be used for providing understanding of the way mechanisms work. (shrink)

Scientists and laypeople alike use the sense of understanding that an explanation conveys as a cue to good or correct explanation. Although the occurrence of this sense or feeling of understanding is neither necessary nor sufficient for good explanation, it does drive judgments of the plausibility and, ultimately, the acceptability, of an explanation. This paper presents evidence that the sense of understanding is in part the routine consequence of two well-documented biases in cognitive psychology: overconfidence and hindsight. In light of (...) the prevalence of counterfeit understanding in the history of science, I argue that many forms of cognitive achievement do not involve a sense of understanding, and that only the truth or accuracy of an explanation make the sense of understanding a valid cue to genuine understanding. (shrink)

Nancy Cartwright argues that so-called capacities, not universal laws of nature, best explain the often complex way events actually unfold. On this view, science would represent a world that is fundamentally "dappled", or disunified, and not, as orthodoxy would perhaps have it, a world unified by universal laws of nature. I argue, first, that the problem Cartwright raises for laws of nature seems to arise for capacities too, so why reject laws of nature? Second, that in so far as there (...) is a problem, it concerns the role of counterfactuals in explanation; I then briefly propose a simple model of counterfactual explanation. Finally, I investigate how a sophisticated version of the regularity theory of laws of nature (that of Ramsey-Lewis) can be neutral between the empirical hypotheses that the world is unified, and that the world is disunified. (shrink)

In this paper, we argue that one way to approach what is known in the literature as the “Trust Gap” in Medical AI is to focus on explanations from an Explainable AI (xAI) perspective. Against the current framework on xAI – which does not offer a real solution – we argue for a pragmatist turn, one that focuses on understanding how we provide explanations in Traditional Medicine (TM), composed by human agents only. Following this, explanations have two specific relevant components: (...) they are usually (i) social and (ii) abductive. Explanations, in this sense, ought to provide understanding by answering contrastive _why-questions_: “Why had P happened instead of Q?” (Miller in AI 267:1–38, 2019) (Sect. 1 ). In order to test the relevancy of this concept of explanation in medical xAI, we offer several reasons to argue that abductions are crucial for medical reasoning and provide a crucial tool to deal with trust gaps between human agents (Sect. 2 ). If abductions are relevant in TM, we can test the capability of Artificial Intelligence systems on this merit. Therefore, we provide an analysis of the capacity for social and abductive reasoning of different AI technologies. Accordingly, we posit that Large Language Models (LLMs) and transformer architectures exhibit a noteworthy potential for effective engagement in abductive reasoning. By leveraging the potential abductive capabilities of LLMs and transformers, we anticipate a paradigm shift in the integration of explanations within AI systems. This, in turn, has the potential to enhance the trustworthiness of AI-driven medical decisions, bridging the Trust Gap that has been a prominent challenge in the field of Medical AI (Sect. 3 ). This development holds the potential to not only improve the interpretability of AI-generated medical insights but also to guarantee that trust among practitioners, patients, and stakeholders in the healthcare domain is still present. (shrink)

Artificial intelligence tools used in employment decision-making cut across the multiple stages of job advertisements, shortlisting, interviews and hiring, and actual and potential bias can arise in each of these stages. One major challenge is to mitigate AI bias and promote fairness in opaque AI systems. This paper argues that the equal opportunity merit principle is an ethical approach for fair AI employment decision-making. Further, explainable AI can mitigate the opacity problem by placing greater emphasis on enhancing the understanding of (...) reasonable users and affected persons as to the AI output. Both the equal opportunity merit principle and explainable AI should be integrated in the design and implementation of AI employment decision-making systems so as to ensure, as far as possible, that the AI output is arrived at through a fair process. (shrink)

Necessity and sufficiency are the building blocks of all successful explanations. Yet despite their importance, these notions have been conceptually underdeveloped and inconsistently applied in explainable artificial intelligence, a fast-growing research area that is so far lacking in firm theoretical foundations. In this article, an expanded version of a paper originally presented at the 37th Conference on Uncertainty in Artificial Intelligence, we attempt to fill this gap. Building on work in logic, probability, and causality, we establish the central role of (...) necessity and sufficiency in XAI, unifying seemingly disparate methods in a single formal framework. We propose a novel formulation of these concepts, and demonstrate its advantages over leading alternatives. We present a sound and complete algorithm for computing explanatory factors with respect to a given context and set of agentive preferences, allowing users to identify necessary and sufficient conditions for desired outcomes at minimal cost. Experiments on real and simulated data confirm our method’s competitive performance against state of the art XAI tools on a diverse array of tasks. (shrink)

Craig casts doubt upon the project of trying to give the traditional sort of necessary and sufficient conditions for A knows that p. He interprets the inadequacy of existing analyses of knowledge as evidence that our concept of knowledge is complex and diffuse, and concludes that we should aim to understand it by thinking about the rôle the concept plays in our lives, rather than by trying to find necessary and sufficient conditions for the truth of knowledge ascriptions.There is surely (...) something right about Craig's view: we are unlikely to succeed in any attempt to analyse away the intricacies in our concept of knowledge. We cannot realistically hope to uncover a set of necessary and sufficient conditions for A knows that p which are in all cases either clearly satisfied or clearly not satisfied. Nor, I suspect, is it possible to offer necessary and sufficient conditions for knowledge which are widely accepted as being more securely understood than knowledge itself. (shrink)

Our epistemic access to the past is infamously patchy: historical information degrades and disappears and bygone eras are often beyond the reach of repeatable experiments. However, historical scientists have been remarkably successful at uncovering and explaining the past. I argue that part of this success is explained by the exploitation of dependencies between historical events, entities, and processes. For instance, if sauropod dinosaurs were hot blooded, they must have been gluttons; the high-energy demands of endothermy restrict sauropod grazing strategies. Understanding (...) such dependencies extends our reach into the past in spite of incomplete data. In addition, this serves as a counterexample to two accounts of method in the historical sciences. By one, historical science proceeds by identifying ‘smoking guns’: traces that discriminate between live hypotheses. By the other, historical hypotheses are supported by consilience: the convergence of independent lines of evidence. However, testing for ‘coherency’ between past hypotheses also plays a critical role in historical confirmation. Just as historical scientists exploit dependencies between past entities and present entities to infer what the past was like, they also exploit dependencies between past entities themselves. I do not suggest that archetypical historical science proceeds in this manner. Rather, the lesson I draw is that historical methodology cannot be characterized as archetypically relying on one method or another. Historical science is at base opportunistic, and is resistant to unitary analyses. 1 Introduction2 Snowballs and Explosions: The Basic Idea3 Were Sauropods Endothermic?4 Dependent Entities and Interdependent Explanations5 Smoking Guns and Consilience. (shrink)

The proposition that Tweety is a bird coheres better with the proposition that Tweety has wings than with the proposition that Tweety cannot fly. This relationship of contrastive coherence is the focus of the present paper. Based on recent work in formal epistemology we consider various possibilities to model this relationship by means of probability theory. In a second step we consider different applications of these models. Among others, we offer a coherentist interpretation of the conjunction fallacy.

In the context of the free will debate, both compatibilists and event-causal libertarians consider that the agent’s mental states and events are what directly causes her decision to act. However, according to the ‘disappearing agent’ objection, if the agent is nothing over and above her physical and mental components, which ultimately bring about her decision, and that decision remains undetermined up to the moment when it is made, then it is a chancy and uncontrolled event. According to agent-causalism, this sort (...) of problem can be overcome if one realizes that the agent herself, as an irreducible substance, is the true originator of her actions. I’ll present arguments that favor this view. Event-causalists have countered that if the agent identifies with some of the inner states that play the self-determining causal role in bringing about the action, then it is as though the action was directly caused by herself. I’ll object that this is not a distinctive aspect of free agency. Agent-causalism has been criticized from most naturalistically inclined fronts, and it must address risks of implausibility, contradiction and unintelligibility. Even though I’ll acknowledge these challenges, I’ll still argue that libertarian free will cannot be defended by any reductionist alternative, and that agent-causalism does not conflict with contemporary science but only with some of its unproven assumptions. (shrink)

Many philosophers think that the Substitution Objection decisively refutes Russellianism. This objection claims that sentences (1) and (2) can differ in truth value. Therefore, it says, the sentences express different propositions, and so Russellianism is false.

Selection explanations explain some non-accidental generalizations in virtue of a selection process. Such explanations are not particulaizable - they do not transfer as explanations of the instances of such generalizations. This is unlike many explanations in the physical sciences, where the explanation of the general fact also provides an explanation of its instances (i.e. standard D-N explanations). Are selection explanations (e.g. in biology) therefore a different kind of explanation? I argue that to understand this issue, we need to see that (...) a standard D-N explanation of some non-accidental generalization (al Fs are Gs) may also ipso facto explain its contrapositive (all non-Gs are non-Fs), but the explanation is particularizable with respect to the former but not to the latter. This can be seen by noting that the Raven Paradox counterexample to the H-D model of confirmation also generates a counterexample to the D-N model of explanation (all ravens are black does not explain why the non-black shoe is a non-raven). In such cases it is natural to take the generalization with the positive predicates to have a particularizable explanation. However, this need not be the case, and in selection explanations it is the generalization with the positive predicates whose explanation is no particularizable. Thus there is no need to suppose that selection explanations are fundamentally different. (shrink)

In this paper I develop a theory of contrastive why questions that establishes under what conditions it is sensible to ask "why p rather than q?". p and q must be outcomes of a single type of causal process.