This paper responds to recent criticisms of the idea that true causal claims, satisfying a minimal “interventionist” criterion for causation, can differ in the extent to which they satisfy other conditions—called stability and proportionality—that are relevant to their use in explanatory theorizing. It reformulates the notion of proportionality so as to avoid problems with previous formulations. It also introduces the notion of conditional independence or irrelevance, which I claim is central to understanding the respects and the extent to which upper (...) level explanations can be “autonomous”. (shrink)

Causal exclusion arguments are taken to threaten the autonomy of the special sciences, and the causal efficacy of mental properties. A recent line of response to these arguments has appealed to “independently plausible” and “well grounded” theories of causation to rebut key premises. In this paper I consider two papers which proceed in this vein and show that they share a common feature: they both require causes to be proportional to their effects. I argue that this feature is a bug, (...) and one that generalises: any attempt to rescue the autonomy of the special sciences, or the efficacy of the mental, from exclusion worries had better not look to proportionality for help. (shrink)

The intuitive distinction between natural and unnatural properties (e.g., green vs. grue) informs our theorizing not only in fundamental physics, but also in non-fundamental domains. This paper develops a reductive account of this broad notion of naturalness that covers non-fundamental properties: for a property to be natural, I propose, is for it to figure in a law of nature. After motivating the account, I defend it from a potential circularity charge. I argue that a suitably broad notion of lawhood can (...) be defined independently of naturalness, if we help ourselves to the notion of a fundamental physical property. I end by showing how the notions of naturalness and lawhood that my account delivers help illuminate a number of other important philosophical notions, like causation, reference/meaning, and rational induction. (shrink)

We propose a formal framework for interpretable machine learning. Combining elements from statistical learning, causal interventionism, and decision theory, we design an idealised explanation game in which players collaborate to find the best explanation for a given algorithmic prediction. Through an iterative procedure of questions and answers, the players establish a three-dimensional Pareto frontier that describes the optimal trade-offs between explanatory accuracy, simplicity, and relevance. Multiple rounds are played at different levels of abstraction, allowing the players to explore overlapping causal (...) patterns of variable granularity and scope. We characterise the conditions under which such a game is almost surely guaranteed to converge on a optimal explanation surface in polynomial time, and highlight obstacles that will tend to prevent the players from advancing beyond certain explanatory thresholds. The game serves a descriptive and a normative function, establishing a conceptual space in which to analyse and compare existing proposals, as well as design new and improved solutions. (shrink)

According to the so-called 'proportionality principle', causes should be proportional to their effects: they should be both enough and not too much for the occurrence of their effects. This principle is the subject of an ongoing debate. On the one hand, many maintain that it is required to address the problem of causal exclusion and take it to capture a crucial aspect of causation. On the other hand, many object that it renders accounts of causation implausibly restrictive and often reject (...) the principle wholesale. I argue that there is exaggeration on both sides. While one half of the principle is overly demanding, the other half is unobjectionable. And while the unobjectionable half does not block exclusion arguments on its own, it provides a nuanced picture of higher-level causation, fits with recent developments in philosophy of causation, and motivates adjustments to standard difference-making accounts of causation. I conclude that at least half of the proportionality principle is worth taking seriously. (shrink)

Proportionality—the idea that causes are neither too general nor too specific for their effects—seems to recommend implausibly disjunctive causes (McGrath, 1998 ; Shapiro & Sober, 2012 ; Franklin-Hall, 2016 ). I argue that this problem should be avoided by appeal to the notion of cohesion. I propose an account of cohesion in terms of the similarity structure of property-spaces, argue that it is not objectionably mysterious, and that alternative approaches—based on naturalness, interventionism, and contrastivism—are inadequate without appeal to it. In (...) an appendix, I show how my proposal can be perspicuously formalized by adapting structural equation models. (shrink)

Supporters of the autonomy of higher-level causation (or explanation) often appeal to proportionality, arguing that higher-level causes are more proportional than their lower-level realizers. Recently, measures based on information theory and causal modeling have been proposed that allow one to shed new light on proportionality and the related notion of specificity. In this paper we apply ideas from this literature to the issue of higher vs. lower-level causation (and explanation). Surprisingly, proportionality turns out to be irrelevant for the question of (...) whether higher-level causes (or explanations) can be autonomous; specificity is a much more informative notion for this purpose. (shrink)

We propose a formal framework for interpretable machine learning. Combining elements from statistical learning, causal interventionism, and decision theory, we design an idealisedexplanation gamein which players collaborate to find the best explanation(s) for a given algorithmic prediction. Through an iterative procedure of questions and answers, the players establish a three-dimensional Pareto frontier that describes the optimal trade-offs between explanatory accuracy, simplicity, and relevance. Multiple rounds are played at different levels of abstraction, allowing the players to explore overlapping causal patterns of (...) variable granularity and scope. We characterise the conditions under which such a game is almost surely guaranteed to converge on a (conditionally) optimal explanation surface in polynomial time, and highlight obstacles that will tend to prevent the players from advancing beyond certain explanatory thresholds. The game serves a descriptive and a normative function, establishing a conceptual space in which to analyse and compare existing proposals, as well as design new and improved solutions. (shrink)

It is often natural to compare two events by describing one as ‘more of a cause’ of some effect than the other. But what do such comparisons amount to, exactly? This paper aims to provide a guided tour of the recent literature on ‘degrees of causation’. Section 2 looks at what I call ‘dependence measures’, which arise from thinking of causes as difference‐makers. Section 3 looks at what I call ‘production measures’, which arise from thinking of causes as jointly sufficient (...) for their effects. Finally, section 4 examines the important question of whether there is any sense in which an agent is more responsible for an outcome in virtue of her action being more of a cause of it. I describe a puzzle that emerges from this question, and explore various strategies for resolving it. (shrink)

The experimental interventions that provide evidence of causal relations are notably similar to those that provide evidence of constitutive relevance relations. In the first two sections, I show that this similarity creates a tension: there is an inconsistent triad between Woodward’s popular interventionist theory of causation, Craver’s mutual manipulability account of constitutive relevance in mechanisms, and a variety of arguments for the incoherence of inter-level causation. I argue for an interpretation of the views in which the tension is merely apparent. (...) I propose to explain inter-level relations without inter-level causation by appealing to the notion of fat-handed interventions, and an argument against inter-level causation which dissolves the problem. (shrink)

Explainable artificial intelligence (XAI) aims to help people understand black box algorithms, particularly of their outputs. But what are these explanations and when is one explanation better than another? The manipulationist definition of explanation from the philosophy of science offers good answers to these questions, holding that an explanation consists of a generalization that shows what happens in counterfactual cases. Furthermore, when it comes to explanatory depth this account holds that a generalization that has more abstract variables, is broader in (...) scope and/or more accurate is better. By applying these definitions and contrasting them with alternative definitions in the XAI literature I hope to help clarify what a good explanation is for AI. (shrink)

Theories of explanation need to account for a puzzling feature of our explanatory practices: the fact that we prefer explanations that are relatively abstract but only moderately so. Contra Franklin-Hall ([2016]), I argue that the interventionist account of explanation provides a natural and elegant explanation of this fact. By striking the right balance between specificity and generality, moderately abstract explanations optimally subserve what interventionists regard as the goal of explanation, namely identifying possible interventions that would have changed the explanandum.

Unsupervised learning algorithms are widely used for many important statistical tasks with numerous applications in science and industry. Yet despite their prevalence, they have attracted remarkably little philosophical scrutiny to date. This stands in stark contrast to supervised and reinforcement learning algorithms, which have been widely studied and critically evaluated, often with an emphasis on ethical concerns. In this article, I analyze three canonical unsupervised learning problems: clustering, abstraction, and generative modeling. I argue that these methods raise unique epistemological and (...) ontological questions, providing data-driven tools for discovering natural kinds and distinguishing essence from contingency. This analysis goes some way toward filling the lacuna in contemporary philosophical discourse on unsupervised learning, as well as bringing conceptual unity to a heterogeneous field more often described by what it isnot(i.e., supervised or reinforcement learning) than by what itis. I submit that unsupervised learning is not just a legitimate subject of philosophical inquiry but perhaps the most fundamental branch of all AI. However, an uncritical overreliance on unsupervised methods poses major epistemic and ethical risks. I conclude by advocating for a pragmatic, error-statistical approach that embraces the opportunities and mitigates the challenges posed by this powerful class of algorithms. (shrink)

Let ‘semantic relevance’ be the thesis that the wide semantic properties of representational mental states (like beliefs and desires) are causally relevant to behavior. A popular way of arguing for semantic relevance runs as follows: start with a sufficient counterfactual condition for causal or explanatory relevance, and show that wide semantic properties meet it with respect to behavior (e.g., Loewer & Lepore (1987,1989), Rescorla (2014), Yablo (2003)).This paper discusses an in‐principle limitation of this strategy: even the most sophisticated counterfactual criteria (...) systematically misclassify irrelevant properties as relevant when they stand in certain kinds of modal co‐variation or ‘entanglement’ relations to genuinely relevant properties. This entanglement problem, I argue, is more general and more serious than proponents of the counterfactual strategy have recognized: it threatens recent interventionist arguments for semantic relevance, and is not easily solved by appeal to proportionality or naturalness. I end by suggesting that proponents of semantic relevance may need to shift their attention from patterns of counterfactuals to the lawful psychological generalizations that explain them. (shrink)

A cause is proportional to its effect when, roughly speaking, it is at the right level of detail. There is a lively debate about whether proportionality is a necessary condition for causation. One of the main arguments against a proportionality constraint on causation is that many ordinary and seemingly perfectly acceptable causal claims cite causes that are not proportional to their effects. In this paper, I suggest that proponents of a proportionality constraint can respond to this objection by developing an (...) idea that is present in Yablo’s early work on proportionality, but which has strangely been ignored by both Yablo and others in the subsequent debate. My suggestion is that proportionality—and, indeed, causation itself—is relative to a domain of events. At the metaphysical level, this means that the causal relation has an extra relatum—namely, a domain of events. At the level of language, it introduces a new way in which causal claims are context-sensitive: what is expressed by a causal claim depends on the contextually relevant domain of events. As I argue, this suggestion allows us to accommodate the truth of ordinary causal claims while extending the explanatory benefits of a proportionality constraint. (shrink)

Binding specificity is a centrally important concept in molecular biology, yet it has received little philosophical attention. Here I aim to remedy this by analyzing binding specificity as a causal property. I focus on the concept’s role in drug design, where it is highly prized and hence directly studied. From a causal perspective, understanding why binding specificity is a valuable property of drugs contributes to an understanding of causal selection—of how and why scientists distinguish between causes, not just causes from (...) noncauses. In particular, the specificity of drugs is precisely what underwrites their value as experimental interventions on biological processes. (shrink)

When once-successful physical theories are abandoned, common wisdom has it that their characteristic theoretical entities are abandoned with them: examples include phlogiston, light rays, Newtonian forces, Euclidean space. But sometimes a theory sees ongoing use, despite being superseded. What should scientific realists say about the characteristic entities of the theories in such cases? The standard answer is that these ‘theoretical relicts’ are merely useful fictions. In this paper we offer a different answer. We start by distinguishing horizontal reduction (in which (...) a superseded theory approximates the successor theory) from vertical reduction (in which a higher-level theory abstracts away from the lower-level theory, but nonetheless can be constructed from it); these are usually regarded as having different ontological consequences. We describe a ‘verticalization’ procedure which transforms horizontal reductions into vertical reductions. The resulting verticalized theories are abstractions rather than approximations, with restricted domains. We identify a sense in which the higher-level theory describes distinct subject matters from the lower-level theory, enabling in certain cases the higher-level theory to retain distinctive explanatory power even in the presence of reduction. We suggest that theoretical entities from superseded theories should be retained in a scientific realist worldview just when, reinterpreted as higher-level abstractions, those theories and their characteristic entities continue to perform distinctive explanatory work in providing the best explanation for less-fundamental phenomena of interest. In slogan form: a good relict is an emergent relict. (shrink)

In ‘Proportionality in Causation, Part I: Theories’, I presented various ways of understanding the idea that causes which are ‘proportional’ to their effects are in some sense preferable. In this companion article, I discuss the principal applications of the resulting theories of proportionality, and the challenges they face.

Biologists and philosophers of science have recently called for an extension of evolutionary theory. This so-called ‘extended evolutionary synthesis’ seeks to integrate developmental processes, extra-genetic forms of inheritance, and niche construction into evolutionary theory in a central way. While there is often agreement in evolutionary biology over the existence of these phenomena, their explanatory relevance is questioned. Advocates of EES posit that their perspective offers better explanations than those provided by ‘standard evolutionary theory’. Still, why this would be the case (...) is unclear. Usually, such claims assume that EES’s superior explanatory status arises from the pluralist structure of EES, its different problem agenda, and a growing body of evidence for the evolutionary relevance of developmental phenomena. However, what is usually neglected in this debate is a discussion of what the explanatory standards of EES actually are, and how they differ from prevailing standards in SET. In other words, what is considered to be a good explanation in EES versus SET? To answer this question, we present a theoretical framework that evaluates the explanatory power of different evolutionary explanations of the same phenomena. This account is able to identify criteria for why and when evolutionary explanations of EES are better than those of SET. Such evaluations will enable evolutionary biology to find potential grounds for theoretical integration. (shrink)

There is a widely shared belief that the higher-level sciences can provide better explanations than lower-level sciences. But there is little agreement about exactly why this is so. It is often suggested that higher-level explanations are better because they omit details. I will argue instead that the preference for higher-level explanations is just a special case of our general preference for informative, logically strong, beliefs. I argue that our preference for informative beliefs entirely accounts for why higher-level explanations are sometimes (...) better—and sometimes worse—than lower-level explanations. The result is a step in the direction of the unity of science hypothesis. 1Introduction2Background: Is Omitting Details an Explanatory Virtue? 2.1Anti-reductionist arguments2.2Reductionist argument2.3Logical strength3Bases, Links and Logical Strength4Functionalism and Fodor’s Argument5Two Generalizations6Should the Base Really Be Maximally Strong?7Anti-reductionist Arguments Regarding the Base8Should the Antecedent of the Link Really Be Maximally Weak? (shrink)

Despite the ubiquity of “levels of organization” in the scientific literature, a nascent “levels skepticism” now claims that the concept of levels is an inherently flawed, misleading, or otherwise inadequate notion for understanding how life scientists produce knowledge about the natural world. However, levels skeptics rely on the maligned “layer-cake” account of levels stemming from Oppenheim and Putnam’s defense of the unity of science for their critical commentary. Recourse to layer-cake levels is understandable, as it is arguably the default conception (...) of levels in philosophy. However, relying on this conception of levels undermines the initial plausibility of general dismissals of the concept, because the problems skeptics identify within the “basic idea” of the concept “levels” are merely ones already widely acknowledged in this conception. I illustrate this “guilt by association” by looking at the embedded role of “levels” in articulating intertheoretical reductionism during the latter part of the 20th century. I conclude by suggesting a methodological framework focusing on local usage patterns as a promising means for future analysis of the levels concept. (shrink)

A much-discussed idea in the causation literature is that it is preferable to invoke causes which are proportional to—neither too general nor too specific for—the effect. This article presents various ways of understanding this idea. In what sense are such causal claims ‘preferable’? And what is it for one event to be ‘proportional’ to another? In a companion article, ‘Proportionality in Causation, Part II: Applications and Challenges’, I discuss the principal applications of the resulting theories of proportionality, and the challenges (...) they face. (shrink)

I raise two objections against Christian List and Peter Menzies' influential account of high-level causation. Improving upon some of Stephen Yablo's earlier work, I develop an alternative theory which evades both objections. The discussion calls into question List and Menzies' main contention, namely, that the exclusion principle, applied to difference-making, is false.

This article defends the notion of downward causation, relating it to a notion of conditional independence.

In contemporary philosophy of science, the consensus view seems to be that scientific explanations describe mechanisms responsible for the phenomena to be explained. Two kinds of explanatory relevance figure in mechanistic accounts of explanation: causal and constitutive. Following prominent accounts, it seems natural to analyze both these relations in terms of systematic interventions into some factor X with respect to another factor Y. However, such interventions are tailored to uncover causal relations only. Construing the constitutive relationship between parts and wholes (...) in terms of interventions thus raises metaphysical, conceptual, and epistemological questions. We here review the barriers that intervention-based inquiry into mechanisms encounters and consider some solutions. (shrink)

This paper sketches a causal account of scientific explanation designed to sustain the judgment that high-level, detail-sparse explanations—particularly those offered in biology—can be at least as explanatorily valuable as lower-level counterparts. The motivating idea is that complete explanations maximize causal economy: they cite those aspects of an event’s causal run-up that offer the biggest-bang-for-your-buck, by costing less (in virtue of being abstract) and delivering more (in virtue making the event stable or robust).

Interventionism is a theory of causal explanation developed by Woodward and Hitchcock. I defend an interventionist perspective on the causal explanations offered within scientific psychology. The basic idea is that psychology causally explains mental and behavioral outcomes by specifying how those outcomes would have been different had an intervention altered various factors, including relevant psychological states. I elaborate this viewpoint with examples drawn from cognitive science practice, especially Bayesian perceptual psychology. I favorably compare my interventionist approach with well-known nomological and (...) mechanistic theories of psychological explanation. (shrink)

ABSTRACT The exclusion problem is meant to show that non-reductive physicalism leads to epiphenomenalism: if mental properties are not identical with physical properties, then they are not causally efficacious. Defenders of a difference-making account of causation suggest that the exclusion problem can be solved because mental properties can be difference-making causes of physical effects. Here, we focus on what we dub an incompatibilist implementation of this general strategy and argue against it from a non-reductive physicalist perspective. Specifically, we argue that (...) incompatibilism undermines either the non-reductionist or the physicalist aspirations of non-reductive physicalism. (shrink)

This article considers the popular thesis that a more proportional relationship between a cause and its effect yields a more abstract causal explanation of that effect, which in turn produces a deeper explanation. This thesis is taken to have important implications for choosing the optimal granularity of explanation for a given explanandum. In this article, I argue that this thesis is not generally true of probabilistic causal relationships. In light of this finding, I propose a pragmatic, interest-relative measure of explanatory (...) depth. This measure uses a decision-theoretic model of information pricing to determine the optimal granularity of explanation for a given explanandum, agent, and decision problem. (shrink)

In this dissertation, I develop and defend a model of causation that allows for dualist mental causation in worlds where the physical domain is physically complete. In Part I, I present the dualist ontology that will be assumed throughout the thesis and identify two challenges for models of mental causation within such an ontology: the exclusion worry and the common cause worry. I also argue that a proper response to these challenges requires a thoroughly lightweight account of causation, i.e. an (...) account that allows for causes to be metaphysically distinct from the phenomena that produce or physically necessitate their effects. In Part II, I critically evaluate contemporary responses to these challenges from the philosophical literature. In particular, I discuss List and Stoljar’s criticism of exclusion worries, Kroedel’s alternative dualist ontology, concerns about the notion of causal sufficiency, and Lowe’s models of dualist mental causation. I argue that none of these proposals provide independent motivation for a thoroughly lightweight account of causation and therefore leave room for improvement. In the first four chapters of Part III, I develop a thoroughly lightweight model of causation, which builds on interventionist approaches to causation. First, I explain how so-called ‘holding fixed’-requirements in standard interventionist accounts stand in the way of dualist mental causation. I then argue that interventionist accounts should impose a robustness condition on causal correlations and that, with this condition in place, the ‘holding fixed’-requirements can be weakened such that they do allow for dualist mental causation. I dub the interventionist model with such weakened ‘holding fixed’-requirements ‘insensitive interventionism’, argue that it can counter the exclusion worry as well as the common cause worry, and explain under which circumstances it would predict there to be dualist mental causation. Importantly, these circumstances might, for all we know, hold in the actual world. In the final three chapters of Part III, I defend insensitive interventionism against some objections. I consider the objection that causation must be productive, the objection that causes must physically necessitate their effects, and the objection that insensitive interventionism is too permissive. I respond by drawing from the literature on causation by absences and on the relation between causation and fundamental physics. Overall, insensitive interventionism performs as well as standard interventionist accounts. I conclude that insensitive interventionism is a credible model of causation. (shrink)

Roughly put, explanatory circles — if any exist — would be propositions such that (i) each explains the next, and (ii) the last explains the first. In this paper, I give two arguments for the view that there are explanatory circles. The first argument appeals to general relativistic worlds in which time is circular. The second argument appeals to special science theories that describe feedback loops. In addition, I show that three standard arguments against explanatory circles are unsuccessful.

Some explanations in social science, psychology and biology belong to a higher level than other explanations. And higher explanations possess the virtue of abstracting away from the details of lower explanations, many philosophers argue. As a result, these higher explanations are irreplaceable. And this suggests that there are genuine higher laws or patterns involving social, psychological and biological states. I show that this ‘abstractness argument’ is really an argument schema, not a single argument. This is because the argument uses the (...) ‘is lower than’ relation, and this relation admits of different readings. I then suggest four rigorous definitions of the ‘is lower than’ relation, and show that the abstractness argument’s prospects are much brighter for some of these definitions than for others. To show this, I evaluate the so-called ‘disjunctive threat’ to the abstractness argument. (shrink)

What is it for an event not to occur? This is an urgent, yet under explored, question for counterfactual analyses of causation quite generally. In this paper I take a lead from Lewis in identifying two different possible standards of non-occurrence that we might adopt and I argue that we need to apply them asymmetrically: one standard for the cause, another for the effect. This is a surprising result. I then offer a contextualist refinement of the Lewis approach in light (...) of initial problems, and discuss how the asymmetry remained hidden until now. I then relate the non-occurrence problem to issues of transitivity and proportionality in causation, before showing that a parallel problem exists for contrastivist and interventionist approaches to causation too. (shrink)

This paper discusses the role of levels and level-bound theoretical terms in neurobiological explanations under the presupposition of a regularity theory of constitution. After presenting the definitions for the constitution relation and the notion of a mechanistic level in the sense of the regularity theory, the paper develops a set of inference rules that allow to determine whether two mechanisms referred to by one or more accepted explanations belong to the same level, or to different levels. The rules are characterized (...) as underlying level distinctions in neuroscience research. (shrink)

John Campbell has claimed that the interventionist account of causation must be amended if it is to be applied to causation in psychology. The problem, he argues, is that it follows from the so-called ‘surgical’ constraint that intervening on psychological states requires the suspension of the agent’s rational autonomy. In this paper, I argue that the problem Campbell identifies is in fact an instance of a wider problem for interventionism, extending beyond psychology, which I call the problem of ‘abrupt transitions’. (...) I then defend a solution to the problem, which replaces the surgical constraint with a weaker constraint on interventions that nevertheless does all the work the surgical constraint was designed to do. I conclude by exploring some interesting consequences of this weaker constraint for causation in psychology. (shrink)

I present a problem for theories of explanation, concerning explanations involving disjunctive properties. The problem is particular acute for the explanatory non-fundamentalist, according to whom non-fundamental scientific explanations are sometimes superior to fundamental physical explanations. I criticise solutions to the problem due to Woodward, Strevens and Sober, and Lewis, and then defend a solution inspired by an account of non-fundamental laws recently defended by Callender and Cohen.

In 2015, Tomasetti and Vogelstein published a paper in Science containing the following provocative statement: “… only a third of the variation in cancer risk among tissues is attributable to environmental factors or inherited predispositions. The majority is due to “bad luck,” that is, random mutations arising during DNA replication in normal, noncancerous stem cells.” The paper—and perhaps especially this rather coy reference to “bad luck”—became a flash point for a series of letters and reviews, followed by replies and yet (...) further counterpoints. In this paper, I critically assess Tomasetti and Vogelstein's argument, discuss the meaning of “luck” (or, better: “chance”) in the context of the debate, and use this case study to address larger questions about methodological criteria for causal explanations of population level patterns in biomedicine. (shrink)

Why do we value higher-level scientific explanations if, ultimately, the world is physical? An attractive answer is that physical explanations often cite facts that don’t make a difference to the event in question. I claim that to properly develop this view we need to commit to a type of deterministic chance. And in doing so, we see the theoretical utility of deterministic chance, giving us reason to accept a package of views including deterministic chance.

According to Interventionism, explanations cite invariant relations which hold among multiple variables. Interventionism incorrectly implies, however, that many common scientific explanations—which cite single‐variable boundary constraints—are not actually explanatory. So I propose a different account of explanation, similar in spirit to Interventionism, which gets those cases of scientific explanation right.

Advancements and deployments of AI-based systems, especially Deep Learning-driven generative language models, have accomplished impressive results over the past few years. Nevertheless, these remarkable achievements are intertwined with a related fear that such technologies might lead to a general relinquishing of our lives’s control to AIs. This concern, which also motivates the increasing interest in the eXplainable Artificial Intelligence (XAI) research field, is mostly caused by the opacity of the output of deep learning systems and the way that it is (...) generated, which is largely obscure to laypeople. A dialectical interaction with such systems may enhance the users’ understanding and build a more robust trust towards AI. Commonly employed as specific formalisms for modelling intra-agent communications, dialogue games prove to be useful tools to rely upon when dealing with user’s explanation needs. The literature already offers some dialectical protocols that expressly handle explanations and their delivery. This paper fully formalises the novel Explanation–Question–Response (EQR) dialogue and its properties, whose main purpose is to provide satisfactory information (i.e., justified according to argumentative semantics) whilst ensuring a simplified protocol, in comparison with other existing approaches, for humans and artificial agents. (shrink)

This paper is on the problem of profligate omissions. The problem is that counterfactual definitions of causation identify as a cause anything that could have prevented an effect but that did not actually occur, which is a highly counterintuitive result. Many solutions of this problem appeal to normative, epistemic, pragmatic, or metaphysical considerations. These existing solutions are in some sense substantive. In contrast, this paper concentrates on the semantics of counterfactuals. I propose to replace Strong Centering with Weak Centering. This (...) allows that the actual world is not the only world that is closest to the actual world. As a result, some counterfactuals that would otherwise have been true, turn out to be false. When these counterfactuals concern causation, fewer causal claims are true. In addition to describing steps towards solving the problem of profligate omissions, the proposal captures an abstraction that is shared by many of the existing solutions: depending on how the distance ordering underlying the Weak Centering condition is constructed and interpreted, some of these existing solutions can be recovered. (shrink)

A central argument for non-reductive accounts of group agency is that complex social entities are capable of exerting causal influence independently of and superseding the causal efficacy of the individuals constituting them. A prominent counter is that non-reductionists run into an insuperable dilemma between identity and redundancy – with identity undermining independent higher-level efficacy and redundancy leading to overdetermination or exclusion. This paper argues that critics of non-reductionism can manage with a simpler and more persuasive reductio strategy called mapping: allow (...) that group agents are causally efficacious in their own right and chart how their causal efficacy is carried out; how it relates to the causal efficacy of individual determiners; how it connects to the causal relevance of background structural factors. The focus exclusively on whether groups are or are not causally efficacious black-boxes implementation, while close attention to how causation is wired increases the visibility of individualist arguments and countenances structure-oriented explanations. (shrink)

There are several supposedly lethal objections to the view that causation is essentially proportional. The first targets an account of proportionality in terms of causal models, pointing out that proportionality is too easily satisfied in causal model accounts of causation through manipulation of the range of values that a variable can take (Franklin-Hall, 2016). The second argues that proportionality legitimizes only the most general things as causes, and proportionality thereby contravenes causal intuitions (Bontly, 2005; Franklin-Hall, 2016; McDonnell, 2018, 2017; Weslake, (...) 2013). The final, and perhaps most intractable, objection holds that proportionality counter-intuitively legitimizes disjunctive causes (Shapiro and Sober, 2012; Weslake, 2017; Woodward, 2018). This paper provides a unified response to these objections, which is best formulated in a causal model framework. I first articulate two independently plausible principles of variable selection – exclusivity and exhaustivity. I then show how the adoption of these principles responds to Franklin-Hall’s objection, and dissolves the remaining two. (shrink)

This dissertation aims to determine the optimal level of granularity for the variables used in probabilistic causal models. These causal models are useful for generating explanations in a number of scientific contexts. In Chapter 1, I argue that there is rarely a unique level of granularity at which a given phenomenon can be causally explained, thereby rejecting various causal exclusion arguments. In Chapter 2, I consider several recent proposals for measuring the explanatory power of causal explanations, and show that these (...) measures fail to track the comparative depth of explanations given at different levels of granularity. In Chapter 3, I offer a pragmatic account of how to partition the measure space of a causal variable so as to optimally explain its effect. My account uses the decision-theoretic notion of value of information, and indexes the relative depth of an explanation to a particular agent faced with a particular decision problem. Chapter 4 applies this same decisiontheoretic framework to answer the epistemic question of how to discover constitutive relationships in nature. In Chapter 5, I describe the formal details of the relationship between random variables that are meant to be coarse-grained and fine-grained representations of the same type of phenomenon. I use this formal framework to rebut a popular argument for the view that special science probabilities can be objective chances. Chapter 6 discusses challenges related to the causal interpretation of Bayes nets that use imprecise rather than precise probabilities. (shrink)

This article considers interventionist arguments for downward causation and non-fundamental level causal explanation from the point of view of inferring causation from experiments. Several authors have utilised the interventionist theory of causal explanation to argue that the causal exclusion argument is moot and that higher-level as well as downward causation is real. I show that this argument can be made when levels are understood as levels of grain, leaving us with a choice between causal explanations pitched at different levels. Causal (...) proportionality has been suggested as a principle for choosing the correct level of description for causal explanations, but this suggestion has serious problems. I offer an alternative principle, based on the idea that explanations should track actual difference-makers. I then consider claims about systems causally influencing their own parts. Such claims risk conceptual confusion for the same reason as the exclusion argument, by conflating causal and constitutive dependencies. The distinction between causation and constitution is used to give criteria for the correct use of the idea of downward causation. (shrink)