The focus in the literature on scientific explanation has shifted in recent years towards model-based approaches. In recent work, Alisa Bokulich has argued that idealization has a central role to play in explanation. Bokulich claims that certain highly-idealized, structural models can be explanatory, even though they are not considered explanatory by causal, mechanistic, or covering law accounts of explanation. This paper focuses on Bokulich’s account in order to make the more general claim that there are problems (...) with maintaining that a structural criterion can capture the way that highly-idealized models explain. This paper examines Bokulich’s claim that the structural model explanation of quantum wavefunction scarring, featuring semiclassical mechanics, is deeper than the explanation provided by the local quantum model. The challenge for Bokulich is to show that the semiclassical model answers a wider range of w-questions, as this is her method of assessing structural information. I look at two reasonable approaches employing w-questions, and I argue that neither approach is ultimately satisfactory. Because structural similarity has preferences for more fundamental models, I argue that the local quantum model provides explanations that at least as deep as the semiclassical ones. The criterion either wrongly identifies all models as explanatory, or prefers models from fundamental theory. Either way, it cannot capture the way that highly-idealized models explain. (shrink)

Marc Lange has recently raised three objections to the account of minimal model explanations offered by Robert Batterman and Collin Rice. In this article, I suggest that these objections are misguided. I suggest that the objections raised by Lange stem from a misunderstanding of the what it is that minimal model explanations seek to explain. This misunderstanding, I argue, consists in Lange’s seeing minimal model explanations as relating special types of models to particular target systems rather than (...) seeing minimal model explanations as looking to explain robust patterns of behavior that are exhibited by a variety of physically diverse systems. (shrink)

Active materials are self-propelled non-living entities which, in some circumstances, exhibit a number of cognitively interesting behaviors such as gradient-following, avoiding obstacles, signaling and group coordination. This has led to scientific and philosophical discussion of whether this may make them useful as minimal models of cognition (Hanczyc, 2014; McGivern, 2019). Batterman and Rice (2014) have argued that what makes a minimal model explanatory is that the model is ultimately in the same universality class as the target system, which (...) underpins why it exhibits the same macrobehavior. We appeal to recent research in basal cognition (Lyon et al., 2021) to establish appropriate target systems and essential features of cognition as a target of modeling. Looking at self-propelled oil droplets, a type of active material, we do not find that organization alone indicates that these systems exhibit the essential features of cognition. We then examine the specific behaviors of oil droplets but also fail to find that these demonstrate the essential features of cognition. Without a universality class, Batterman & Rice’s account of the explanatory power of minimal models simply does not apply to cognition. However, we also want to stress that it is not intended to; cognition is not the same type of behavioral phenomena as those found in physics. We then look to the minimal cognition methodology of Beer (1996, 2020a, b) to show how active materials can be explanatorily valuable regardless of their cognitive status because they engage in specific behaviors that have traditionally been expected to involve internal representational dynamics, revealing misconceptions about the cognitive underpinnings of certain, specific behaviors in target systems where such behaviors are cognitive. Further, Beer’s models can also be genuinely explanatory by providing dynamical explanations. (shrink)

This essay develops an inferential account of model explanation, based on Mauricio Suárez’s inferential conception of scientific representation and Alisa Bokulich’s counterfactual account of model explanation. It is suggested that the fact that a scientific model can explain is essentially linked to how a modeler uses an established model to make various inferences about the target system on the basis of results derived from the model. The inference practice is understood as a two-step (...) activity, with the first step involving making counterfactual statements about the model itself and the second step involving making hypothetical statements transferring over claims derived from the model onto the target. To illustrate how this two-step activity proceeds, an agent-based simulation model is discussed. (shrink)

Detailed examinations of scientific practice have revealed that the use of idealized models in the sciences is pervasive. These models play a central role in not only the investigation and prediction of phenomena, but in their received scientific explanations as well. This has led philosophers of science to begin revising the traditional philosophical accounts of scientific explanation in order to make sense of this practice. These new model-based accounts of scientific explanation, however, raise a number of key (...) questions: Can the fictions and falsehoods inherent in the modeling practice do real explanatory work? Do some highly abstract and mathematical models exhibit a noncausal form of scientific explanation? How can one distinguish an exploratory "how-possibly" model explanation from a genuine "how-actually" model explanation? Do modelers face tradeoffs such that a model that is optimized for yielding explanatory insight, for example, might fail to be the most predictively accurate, and vice versa? This chapter explores the various answers that have been given to these questions. (shrink)

In this article, I explore the compatibility of inference to the best explanation (IBE) with several influential models and accounts of scientific explanation. First, I explore the different conceptions of IBE and limit my discussion to two: the heuristic conception and the objective Bayesian conception. Next, I discuss five models of scientific explanation with regard to each model’s compatibility with IBE. I argue that Kitcher’s unificationist account supports IBE; Railton’s deductive–nomological–probabilistic model, Salmon’s statistical-relevance model, (...) and van Fraassen’s erotetic account are incompatible with IBE; and Salmon’s causal–mechanical model is merely consistent with IBE. In short, many influential models of scientific explanation do not support IBE. I end by outlining three possible conclusions to draw: (1) either philosophers of science or defenders of IBE have seriously misconstrued the concept of explanation, (2) philosophers of science and defenders of IBE do not use the term ‘explanation’ univocally, and (3) the ampliative conception of IBE, which is compatible with any model of scientific explanation, deserves a closer look. (shrink)

Mathematical models provide explanations of limited power of specific aspects of phenomena. One way of articulating their limits here, without denying their essential powers, is in terms of contrastive explanation.

Batterman and Rice ([2014]) argue that minimal models possess explanatory power that cannot be captured by what they call ‘common features’ approaches to explanation. Minimal models are explanatory, according to Batterman and Rice, not in virtue of accurately representing relevant features, but in virtue of answering three questions that provide a ‘story about why large classes of features are irrelevant to the explanandum phenomenon’ ([2014], p. 356). In this article, I argue, first, that a method (the renormalization group) they (...) propose to answer the three questions cannot answer them, at least not by itself. Second, I argue that answers to the three questions are unnecessary to account for the explanatoriness of their minimal models. Finally, I argue that a common features account, what I call the ‘generalized ontic conception of explanation’, can capture the explanatoriness of minimal models. (shrink)

Among the many functions of models, explanation is central to the functioning and aims of science. However, the discussions surrounding modeling and explanation in philosophy have largely remained separate from each other. This chapter seeks to bridge the gap by focusing on the puzzle of model-based explanation, asking how different philosophical accounts answer the following question: if idealizations and fictions introduce falsehoods into models, how can idealized and fictional models provide true explanations? The chapter provides a (...) selective and critical overview of the available strategies for solving this puzzle, mainly focusing on idealized models and how they explain. (shrink)

This article is an experiment. Consider a minimalist model of cognition (models, means of model-building and history of their evolution). In this model, explanation could be defined as a means allowing to advance: production of models and means of model-building (thus, yielding 1st class understanding), exploration and use of them (2nd class), and/or teaching (3rd class). At minimum, 3rd class understanding is necessary for an explanation to be respected.

In this article, I consider an important challenge to the popular theory of scientific inference commonly known as ‘inference to the best explanation’, one that has received scant attention.1 1 The problem is that there exists a wide array of rival models of explanation, thus leaving IBE objectionably indeterminate. First, I briefly introduce IBE. Then, I motivate the problem and offer three potential solutions, the most plausible of which is to adopt a kind of pluralism about the rival (...) models of explanation. However, I argue that how ranking explanations on this pluralistic account of IBE remains obscure and pluralism leads to contradictory results. In light of these objections, I attempt to dissolve the problem by showing why IBE does not require a ‘model’ of explanation and by giving an account of what explanation consists in within the context of IBE. (shrink)

In Economics Rules, Dani Rodrik (2015) argues that what makes economics powerful despite the limitations of each and every model is its diversity of models. Rodrik suggests that the diversity of models in economics improves its explanatory capacities, but he does not fully explain how. I offer a clearer picture of how models relate to explanations of particular economic facts or events, and suggest that the diversity of models is a means to better economic explanations.

Agent-based models (ABMs) are one type of simulation model used in the context of the COVID-19 pandemic. In contrast to equation-based models, ABMs are algorithms that use individual agents and attribute changing characteristics to each one, multiple times during multiple iterations over time. This paper focuses on three philosophical aspects of ABMs as models of causal mechanisms, as generators of emergent phenomena, and as providers of explanation. Based on my discussion, I conclude that while ABMs cannot help much (...) with causal inference, they can be viewed as etio-prognostic explanations of illness occurrence and outcome. (shrink)

As the demand for explainable deep learning grows in the evaluation of language technologies, the value of a principled grounding for those explanations grows as well. Here we study the state-of-the-art in explanation for neural models for natural-language processing (NLP) tasks from the viewpoint of philosophy of science. We focus on recent evaluation work that finds brittleness in explanations obtained through attention mechanisms.We harness philosophical accounts of explanation to suggest broader conclusions from these studies. From this analysis, we (...) assert the impossibility of causal explanations from attention layers over text data. We then introduce NLP researchers to contemporary philosophy of science theories that allow robust yet non-causal reasoning in explanation, giving computer scientists a vocabulary for future research. (shrink)

This paper aims to contribute to the existing literature on normative cognition by providing a lineage explanation of human social norm psychology. This approach builds upon theories of goal-directed behavioral control in the reinforcement learning and control literature, arguing that this form of control defines an important class of intentional normative mental states that are instrumental in nature. I defend the view that great ape capacities for instrumental reasoning and our capacity (or family of capacities) for shared intentionality coadapted (...) to each other and argue that the evolution of this capacity has allowed the representation of social norms and the emergence of our capacity for normative guidance. (shrink)

Economists use various models to explain why it is that firms are capable of pricing above marginal cost. In this paper, we will examine two of them: the Cournot and Bertrand duopoly models. Economists generally accept both models as good explanations of the phenomenon, but the two models contradict each other in various important ways. The puzzle is that two inconsistent explanations are both regarded as good explanations for the same phenomenon. This becomes especially worrisome when the two models are (...) offering divergent policy recommendations. This report presents that puzzle by laying out how the two models contradict each other in a myriad of ways and then offers five possible solutions to that puzzle from various economists, philosophers of science, and philosophers of economics. (shrink)

A common view of how-possibly explanations in biology treats them as explanatorily incomplete. In addition to this interpretation of how-possibly explanation, I argue that there is another interpretation, one which features what I term “explanatory strategies.” This strategy-centered interpretation of how-possibly explanation centers on there being a different explanatory context within which how-possibly explanations are offered. I contend that, in conditions where this strategy context is recognized, how-possibly explanations can be understood as complete explanations. I defend this alternative (...) interpretation by analyzing the explanatory value of simple physical models the nineteenth century developmental biologist Wilhelm His constructed for animal development. (shrink)

By now a consensus has emerged that people, when subjected to high-stakes decisions through automated decision systems, have a moral right to have these decisions explained to them. However, furnishing such explanations can be costly. So the right to an explanation creates what we call the cost problem: providing subjects of automated decisions with appropriate explanations of the grounds of these decisions can be costly for the companies and organisations that use these automated decision systems. In this paper, we (...) explore whether large language models could prove significant in overcoming the cost problem. We provide an initial case for believing that they can but only with serious ethical costs. (shrink)

According to a widespread view in metaphysics and philosophy of science, all explanations involve relations of ontic dependence between the items appearing in the explanandum and the items appearing in the explanans. I argue that a family of mathematical cases, which I call “viewing-as explanations”, are incompatible with the Dependence Thesis. These cases, I claim, feature genuine explanations that aren’t supported by ontic dependence relations. Hence the thesis isn’t true in general. The first part of the paper defends this claim (...) and discusses its significance. The second part of the paper considers whether viewing-as explanations occur in the empirical sciences, focusing on the case of so-called fictional models. It’s sometimes suggested that fictional models can be explanatory even though they fail to represent actual worldly dependence relations. Whether or not such models explain, I suggest, depends on whether we think scientific explanations necessarily give information relevant to intervention and control. Finally, I argue that counterfactual approaches to explanation also have trouble accommodating viewing-as cases. (shrink)

I compare Frith and colleagues’ influential comparator account of how the sense of agency is elicited to the multifactorial weighting model advocated by Synofzik and colleagues. I defend the comparator model from the common objection that the actual sensory consequences of action are not needed to elicit the sense of agency. I examine the comparator model’s ability to explain the performance of healthy subjects and those suffering from delusions of alien control on various self-attribution tasks. It transpires (...) that the comparator model needs case-by-case adjustment to deal with problematic data. In response to this, the multifactorial weighting model of Synofzik and colleagues is introduced. Although this model is incomplete, it is more naturally constrained by the cases that are problematic for the comparator model. However, this model may be untestable. I conclude that currently the comparator model approach has stronger support than the multifactorial weighting model approach. (shrink)

Call an explanation in which a non-mathematical fact is explained—in part or in whole—by mathematical facts: an extra-mathematical explanation. Such explanations have attracted a great deal of interest recently in arguments over mathematical realism. In this article, a theory of extra-mathematical explanation is developed. The theory is modelled on a deductive-nomological theory of scientific explanation. A basic DN account of extra-mathematical explanation is proposed and then redeveloped in the light of two difficulties that the basic (...) theory faces. The final view appeals to relevance logic and uses resources in information theory to understand the explanatory relationship between mathematical and physical facts. 1Introduction2Anchoring3The Basic Deductive-Mathematical Account4The Genuineness Problem5Irrelevance6Relevance and Information7Objections and Replies 7.1Against relevance logic7.2Too epistemic7.3Informational containment8Conclusion. (shrink)

Stanford’s argument against scientific realism focuses on theories, just as many earlier arguments from inconceivability have. However, there are possible arguments against scientific realism involving unconceived (or inconceivable) entities of different types: observations, models, predictions, explanations, methods, instruments, experiments, and values. This paper charts such arguments. In combination, they present the strongest challenge yet to scientific realism.

The relationship between topological explanation and mechanistic explanation is unclear. Most philosophers agree that at least some topological explanations are mechanistic explanations. The crucial question is how to make sense of this claim. Zednik (Philos Psychol 32(1):23–51, 2019) argues that topological explanations are mechanistic if they (i) describe mechanism sketches that (ii) pick out organizational properties of mechanisms. While we agree with Zednik’s conclusion, we critically discuss Zednik’s account and show that it fails as a general account of (...) how and when topological explanations are mechanistic. First, if topological explanations were just mechanism sketches, this implies that they could be enriched by replacing topological terms with mechanistic detail. This, however, conflicts how topological explanations are used in scientific practice. Second, Zednik’s account fails to show how topological properties can be organizational properties of mechanisms that have a place in mechanistic explanation. The core issue is that Zednik’s account ignores that topological properties often are _global_ properties while mechanistic explanantia refer to _local_ properties. We demonstrate how these problems can be solved by a recent account of mechanistic completeness (Craver and Kaplan in Br J Philos Sci 71(1):287–319, 2020; Kohár and Krickel in Calzavarini and Viola (eds) Neural mechanisms—new challenges in the philosophy of neuroscience, Springer, New York, 2021) and use a multilayer network model of Alzheimer’s Disease to illustrate this. (shrink)

Philippe Huneman has recently questioned the widespread application of mechanistic models of scientific explanation based on the existence of structural explanations, i.e. explanations that account for the phenomenon to be explained in virtue of the mathematical properties of the system where the phenomenon obtains, rather than in terms of the mechanisms that causally produce the phenomenon. Structural explanations are very diverse, including cases like explanations in terms of bowtie structures, in terms of the topological properties of the system, or (...) in terms of equilibrium. The role of mathematics in bowtie structured systems and in topologically constrained systems has recently been examined in different papers. However, the specific role that mathematical properties play in equilibrium explanations requires further examination, as different authors defend different interpretations, some of them closer to the new-mechanistic approach than to the structural model advocated by Huneman. In this paper, we cover this gap by investigating the explanatory role that mathematics play in Blaser and Kirschner’s nested equilibrium model of the stability of persistent long-term human-microbe associations. We argue that their model is explanatory because: i) it provides a mathematical structure in the form of a set of differential equations that together satisfy an ESS; ii) that the nested nature of the ESSs makes the explanation of host-microbe persistent associations robust to any perturbation; iii) that this is so because the properties of the ESS directly mirror the properties of the biological system in a non-causal way. The combination of these three theses make equilibrium explanations look more similar to structural explanations than to causal-mechanistic explanation. (shrink)

Using as case studies two early diagrams that represent mechanisms of the cell division cycle, we aim to extend prior philosophical analyses of the roles of diagrams in scientific reasoning, and specifically their role in biological reasoning. The diagrams we discuss are, in practice, integral and indispensible elements of reasoning from experimental data about the cell division cycle to mathematical models of the cycle’s molecular mechanisms. In accordance with prior analyses, the diagrams provide functional explanations of the cell cycle and (...) facilitate the construction of mathematical models of the cell cycle. But, extending beyond those analyses, we show how diagrams facilitate the construction of mathematical models, and we argue that the diagrams permit nomological explanations of the cell cycle. We further argue that what makes diagrams integral and indispensible for explanation and model construction is their nature as locality aids: they group together information that is to be used together in a way that sentential representations do not. (shrink)

Philosophers have cataloged a range of genealogical methods by which different sorts of normative conclusions can be established. Although such methods provide diverging ways of pursuing genealogical inquiry, they typically converge in eschewing historiographic methodology, in favor of a uniquely philosophical approach. In contrast, one genealogist who drew on historiographic methodology is Michel Foucault. This article presents the motivations and advantages of Foucault's genealogical use of such a methodology. It advances two mains claims. First, that Foucault's early 1970s work employs (...) a distinct genealogical method, which borrows from contemporary historiographic models of explanation to expand the range of objects that are proper to genealogical accounts of historical change. I demonstrate how Foucault modifies two central commitments of Nietzsche by broadening the dimensions of genealogical inquiry and explanation. Second, that historical method has normative relevance for genealogy, insofar as different historiographic choices can lead to different normative conclusions. I motivate this second claim by explaining how Foucault's multidimensional genealogical method expands both (a) the range of objects that are subject to evaluative assessment, and (b) the set of possible prescriptive recommendations that follow from such assessment. (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)

There has been a growing trend to include non-causal models in accounts of scientific explanation. A worry addressed in this paper is that without a higher threshold for explanation there are no tools for distinguishing between models that provide genuine explanations and those that provide merely potential explanations. To remedy this, a condition is introduced that extends a veridicality requirement to models that are empirically underdetermined, highly-idealised, or otherwise non-causal. This condition is applied to models of electroweak symmetry (...) breaking beyond the Standard Model. (shrink)

This chapter provides a systematic overview of topological explanations in the philosophy of science literature. It does so by presenting an account of topological explanation that I (Kostić and Khalifa 2021; Kostić 2020a; 2020b; 2018) have developed in other publications and then comparing this account to other accounts of topological explanation. Finally, this appraisal is opinionated because it highlights some problems in alternative accounts of topological explanations, and also it outlines responses to some of the main criticisms raised (...) by the so-called new mechanists. (shrink)

Can purely predictive models be useful in investigating causal systems? I argue ‘yes’. Moreover, in many cases not only are they useful, they are essential. The alternative is to stick to models or mechanisms drawn from well-understood theory. But a necessary condition for explanation is empirical success, and in many cases in social and field sciences such success can only be achieved by purely predictive models, not by ones drawn from theory. Alas, the attempt to use theory to achieve (...) explanation or insight without empirical success therefore fails, leaving us with the worst of both worlds – neither prediction nor explanation. Best go with empirical success by any means necessary. I support these methodological claims via case studies of two impressive feats of predictive modelling: opinion polling of political elections, and weather forecasting. (shrink)

This paper offers a new account of metaphysical explanation. The account is modelled on Kitcher’s unificationist approach to scientific explanation. We begin, in Sect. 2, by briefly introducing the notion of metaphysical explanation and outlining the target of analysis. After that, we introduce a unificationist account of metaphysical explanation before arguing that such an account is capable of capturing four core features of metaphysical explanations: irreflexivity, non-monotonicity, asymmetry and relevance. Since the unificationist theory of metaphysical (...) class='Hi'>explanation inherits irreflexivity and non-monotonicity directly from the unificationist theory of scientific explanation that underwrites it, we focus on demonstrating how the account can secure asymmetry and relevance. (shrink)

Should models of consciousness be detailed _mechanistic_ models of particular types of systems, or should they be _minimal_ models that abstract away from the underlying mechanistic details and provide generalisations? Detailed mechanistic models may afford a complete and precise account of consciousness in human beings and other, physiologically similar mammals. But they do not provide a good model of consciousness in other animals, such as non-vertebrates, let alone artificial systems. Minimal models can be applicable to a wide range of (...) different conscious systems. But do they provide genuine explanations that are autonomous from explanations by detailed mechanistic models? This paper provides a taxonomy of minimal models and measures of consciousness, clarifies their relation to detailed mechanistic models, and highlights benefits of different minimalist approaches to consciousness. (shrink)

Simple idealized models seem to provide more understanding than opaque, complex, and hyper-realistic models. However, an increasing number of scientists are going in the opposite direction by utilizing opaque machine learning models to make predictions and draw inferences, suggesting that scientists are opting for models that have less potential for understanding. Are scientists trading understanding for some other epistemic or pragmatic good when they choose a machine learning model? Or are the assumptions behind why minimal models provide understanding misguided? (...) In this paper, using the case of deep neural networks, I argue that it is not the complexity or black box nature of a model that limits how much understanding the model provides. Instead, it is a lack of scientific and empirical evidence supporting the link that connects a model to the target phenomenon that primarily prohibits understanding. (shrink)

Models are indispensable tools of scientific inquiry, and one of their main uses is to improve our understanding of the phenomena they represent. How do models accomplish this? And what does this tell us about the nature of understanding? While much recent work has aimed at answering these questions, philosophers' focus has been squarely on models in empirical science. I aim to show that pure mathematics also deserves a seat at the table. I begin by presenting two cases: Cramér’s random (...) model of the prime numbers and the function field model of the integers. These cases show that mathematicians, like empirical scientists, rely on unrealistic models to gain understanding of complex phenomena. They also have important implications for some much-discussed theses about scientific understanding. First, modeling practices in mathematics confirm that one can gain understanding without obtaining an explanation. Second, these cases undermine the popular thesis that unrealistic models confer understanding by imparting counterfactual knowledge. (shrink)

This paper presents and argues for an account of objectual understanding that aims to do justice to the full range of cases of scientific understanding, including cases in which one does not have an explanation of the understood phenomenon. According to the proposed account, one understands a phenomenon just in case one grasps a sufficiently accurate and comprehensive model of the ways in which it or its features are situated within a network of dependence relations; one’s degree of (...) understanding is proportional to the comprehensiveness and accuracy of such a model. I compare this account with accounts of scientific understanding that explicate understanding in terms of having an explanation of the understood phenomenon. I discuss three distinct types of cases in which scientific understanding does not amount to possessing an explanation of any kind, and argue that the proposed model-based account can accommodate these cases while still retaining a strong link between understanding and explanation. (shrink)

The book answers long-standing questions on scientific modeling and inference across multiple perspectives and disciplines, including logic, mathematics, physics and medicine. The different chapters cover a variety of issues, such as the role models play in scientific practice; the way science shapes our concept of models; ways of modeling the pursuit of scientific knowledge; the relationship between our concept of models and our concept of science. The book also discusses models and scientific explanations; models in the semantic view of theories; (...) the applicability of mathematical models to the real world and their effectiveness; the links between models and inferences; and models as a means for acquiring new knowledge. It analyzes different examples of models in physics, biology, mathematics and engineering. Written for researchers and graduate students, it provides a cross-disciplinary reference guide to the notion and the use of models and inferences in science. (shrink)

Moral philosophers are, among other things, in the business of constructing moral theories. And moral theories are, among other things, supposed to explain moral phenomena. Consequently, one’s views about the nature of moral explanation will influence the kinds of moral theories one is willing to countenance. Many moral philosophers are (explicitly or implicitly) committed to a deductive model of explanation. As I see it, this commitment lies at the heart of the current debate between moral particularists and (...) moral generalists. In this paper I argue that we have good reasons to give up this commitment. In fact, I show that an examination of the literature on scientific explanation reveals that we are used to, and comfortable with, non-deductive explanations in almost all areas of inquiry. As a result, I argue that we have reason to believe that moral explanations need not be grounded in exceptionless moral principles. (shrink)

Scientific models typically contain idealizations, or assumptions that are known not to be true. Philosophers have long questioned the nature of idealizations: Are they heuristic tools that will be abandoned? Or rather fictional representations of reality? And how can we reconcile them with realism about knowledge of nature? Immanuel Kant developed an account of scientific investigation that can inspire a new approach to the contemporary debate. Kant argued that scientific investigation is possible only if guided by ideal assumptions—what he calls (...) “regulative ideas”. These ideas are not true of objects of nature, and yet they are not heuristic tools or fictional represen- tations. They are necessary rules governing the construction and assessment of scientific explanations. In this paper, I suggest that some idealizations can be interpreted as having necessary regulative value and as being compatible with scientific realism. I first analyze the puzzle of the nature of idealization and present the main approaches to this topic in the literature. Second, I reconsider the puzzle vis-a-vis a restricted, Kantian definition of idealization and a novel characterization of the relation between idealization and truth. Finally, I discuss in detail an example of idealization (the Hardy-Weinberg equilibrium) along the suggested Kantian lines. (shrink)

We depend upon explanation in order to “make sense” out of our world. And, making sense is all the more important when dealing with change. But, what happens if our explanations are wrong? This question is examined with respect to two types of explanatory model. Models based on labels and categories we shall refer to as “representations.” More complex models involving stories, multiple algorithms, rules of thumb, questions, ambiguity we shall refer to as “compressions.” Both compressions and representations (...) are reductions. But representations are far more reductive than compressions. Representations can be treated as a set of defined meanings – coherence with regard to a representation is the degree of fidelity between the item in question and the definition of the representation, of the label. By contrast, compressions contain enough degrees of freedom and ambiguity to allow us to make internal predictions so that we may determine our potential actions in the possibility space. Compressions are explanatory via mechanism. Representations are explanatory via category. Managers are often confusing their evocation of a representation as the creation of a context of compression . When this type of explanatory error occurs, more errors follow. In the drive for efficiency such substitutions are all too often proclaimed – at the manager’s peril. (shrink)

We introduce and discuss a knowledge-driven distillation approach to explaining black-box models by means of two kinds of interpretable models. The first is perceptron (or threshold) connectives, which enrich knowledge representation languages such as Description Logics with linear operators that serve as a bridge between statistical learning and logical reasoning. The second is Trepan Reloaded, an ap- proach that builds post-hoc explanations of black-box classifiers in the form of decision trees enhanced by domain knowledge. Our aim is, firstly, to target (...) a model-agnostic distillation approach exemplified with these two frameworks, secondly, to study how these two frameworks interact on a theoretical level, and, thirdly, to investigate use-cases in ML and AI in a comparative manner. Specifically, we envision that user-studies will help determine human understandability of explanations generated using these two frameworks. (shrink)

One of the central aims of science is explanation: scientists seek to uncover why things happen the way they do. This chapter addresses what kinds of explanations are formulated in biology, how explanatory aims influence other features of the field of biology, and the implications of all of this for biology education. Philosophical treatments of scientific explanation have been both complicated and enriched by attention to explanatory strategies in biology. Most basically, whereas traditional philosophy of science based (...) class='Hi'>explanation on derivation from scientific laws, there are many biological explanations in which laws play little or no role. Instead, the field of biology is a natural place to turn for support for the idea that causal information is explanatory. Biology has also been used to motivate mechanistic accounts of explanation, as well as criticisms of that approach. Ultimately, the most pressing issue about explanation in biology may be how to account for the wide range of explanatory styles encountered in the field. This issue is crucial, for the aims of biological explanation influence a variety of other features of the field of biology. Explanatory aims account for the continued neglect of some central causal factors, a neglect that would otherwise be mysterious. This is linked to the persistent use of models like evolutionary game theory and population genetic models, models that are simplified to the point of unreality. These explanatory aims also offer a way to interpret many biologists’ total commitment to one or another methodological approach, and the intense disagreements that result. In my view, such debates are better understood as arising not from different theoretical commitments, but commitments to different explanatory projects. Biology education would thus be enriched by attending to approaches to biological explanation, as well as the unexpected ways that these explanatory aims influence other features of biology. I suggest five lessons for teaching about explanation in biology that follow from the considerations of this chapter. (shrink)

Recently, Piccinini and Craver have stated three theses concerning the relations between functional analysis and mechanistic explanation in cognitive sciences: No Distinctness: functional analysis and mechanistic explanation are explanations of the same kind; Integration: functional analysis is a kind of mechanistic explanation; and Subordination: functional analyses are unsatisfactory sketches of mechanisms. In this paper, I argue, first, that functional analysis and mechanistic explanations are sub-kinds of explanation by scientific (idealized) models. From that point of view, we (...) must take into account the tradeoff between the representational/explanatory goals of generality and precision that govern the practice of model-building. In some modeling scenarios, it is rational to maximize explanatory generality at the expense of mechanistic precision. This tradeoff allows me to put forward a problem for the mechanist position. If mechanistic modeling endorses generality as a valuable goal, then Subordination should be rejected. If mechanists reject generality as a goal, then Integration is false. I suggest that mechanists should accept that functional analysis can offer acceptable explanations of cognitive phenomena. (shrink)

Philosophers of psychology debate, among other things, which psychological models, if any, are (or provide) mechanistic explanations. This should seem a little strange given that there is rough consensus on the following two claims: 1) a mechanism is an organized collection of entities and activities that produces, underlies, or maintains a phenomenon, and 2) a mechanistic explanation describes, represents, or provides information about the mechanism producing, underlying, or maintaining the phenomenon to be explained (i.e. the explanandum phenomenon) (Bechtel and (...) Abrahamsen 2005; Craver 2007). If there is a rough consensus on what mechanisms are and that mechanistic explanations describe, represent, or provide information about them, then how is there no consensus on which psychological models are (or provide) mechanistic explanations? Surely the psychological models that are mechanistic explanations are the models that describe, represent, or provide information about mechanisms. That is true, of course; the trouble arises when determining what exactly that involves. Philosophical disagreement over which psychological models are mechanistic explanations is often disagreement about what it means to describe, represent, or provide information about a mechanism, among other things (Hochstein 2016; Levy 2013). In addition, one's position in this debate depends on a host of other seemingly arcane metaphysical issues, such as the nature of mechanisms, computational and functional properties (Piccinini 2016), and realization (Piccinini and Maley 2014), as well as the relation between models, methodologies, and explanations (Craver 2014; Levy 2013; Zednik 2015). Although I inevitably advocate a position, my primary aim in this chapter is to spell out all these relationships and canvas the positions that have been taken (or could be taken) with respect to mechanistic explanation in psychology, using dynamical systems models and cognitive models (or functional analyses) as examples. (shrink)

In the domain of cognitive science, understanding consciousness through the investigation of neural correlates has been the primary research approach. The exploration of neural correlates of consciousness is focused on identifying these correlates and reducing consciousness to a physical phenomenon, embodying a form of reductionist physicalism. This inevitably leads to challenges in explaining consciousness itself. The computational interpretation of consciousness takes a functionalist view, grounded in physicalism, and models conscious experience as a cognitive function, elucidated through computational means. This paper (...) posits that predictive processing offers a fresh paradigm for comprehending human cognition and serves as a neural foundation for the computational elucidation of consciousness. Additionally, free energy theory, as a mathematical construct explaining the predictive processing model, furnishes a theoretical scaffold for understanding consciousness computationally. (shrink)

A finer-grained delineation of a given explanandum reveals a nexus of closely related causal and non- causal explanations, complementing one another in ways that yield further explanatory traction on the phenomenon in question. By taking a narrower construal of what counts as a causal explanation, a new class of distinctively mathematical explanations pops into focus; Lange’s characterization of distinctively mathematical explanations can be extended to cover these. This new class of distinctively mathematical explanations is illustrated with the Lotka-Volterra equations. (...) There are at least two distinct ways those equations might hold of a system, one of which yields straightforwardly causal explanations, but the other of which yields explanations that are distinctively mathematical in terms of nomological strength. In the first, one first picks out a system or class of systems, finds that the equations hold in a causal -explanatory way; in the second, one starts with the equations and explanations that must apply to any system of which the equations hold, and only then turns to the world to see of what, if any, systems it does in fact hold. Using this new way in which a model might hold of a system, I highlight four specific avenues by which causal and non- causal explanations can complement one another. (shrink)

"Explanation and Understanding" (1971) by Georg Henrik von Wright is a modern classic in analytic hermeneutics, and in the philosophy of the social sciences and humanities in general. In this work, von Wright argues against naturalism, or methodological monism, i.e. the idea that both the natural sciences and the social sciences follow broadly the same general scientific approach and aim to achieve causal explanations. Against this view, von Wright contends that the social sciences are qualitatively different from the natural (...) sciences: according to his view, the natural sciences aim at causal explanations, whereas the purpose of the social sciences is to understand their subjects. In support of this conviction, von Wright also puts forward a version of the so-called logical connection argument. -/- Von Wright views scientific explanation along the lines of the traditional covering law model. He suggests that the social sciences, in contrast, utilize what he calls “practical syllogism” in understanding human actions. In addition, von Wright presents in this work an original picture on causation: a version of the manipulability theory of causation. -/- In the four decades following von Wright’s classic work, the overall picture in in the philosophy of science has changed significantly, and much progress has been made in various fronts. The aim of the contribution is to revisit the central ideas of "Explanation and Understanding" and evaluate them from this perspective. The covering law model of explanation and the regularity theory of causation behind it have since then fallen into disfavor, and virtually no one believes that causal explanations even in the natural sciences comply with the covering law model. No wonder then that covering law explanations are not found in the social sciences either. Ironically, the most popular theory of causal explanation in the philosophy of science nowadays is the interventionist theory, which is a descendant of the manipulability theory of von Wright and others. However, this theory can be applied with no special difficulties in both the natural sciences and the social sciences. -/- Von Wright’s logical connection argument and his ideas concerning practical syllogisms are also critically assessed. It is argued that in closer scrutiny, they do not pose serious problems for the view that the social sciences too provide causal explanations. In sum, von Wright’s arguments against naturalism do not appear, in today’s perspective, particularly convincing. (shrink)

In a recent special issue dedicated to Dani Rodrik’s (2015) influential monograph Economics Rules, Grüne-Yanoff and Marchionni (2018) raise a potentially damning problem for Rodrik’s suggestion that progress in economics should be understood and measured laterally, by a continuous expansion of new models. They argue that this could lead to an “embarrassment of riches”, i.e. the rapid expansion of our model library to such an extent that we become unable to choose between the available models, and thus needs to (...) be solved to make ‘model pluralism’ viable. Drawing on Veit’s (2019a) ‘model pluralism’ account, this paper argues that model pluralism as a thesis about the relationship between science and nature undermines the very idea of a general model selection framework for policy making. (shrink)

A common kind of explanation in cognitive neuroscience might be called functiontheoretic: with some target cognitive capacity in view, the theorist hypothesizes that the system computes a well-defined function (in the mathematical sense) and explains how computing this function constitutes (in the system’s normal environment) the exercise of the cognitive capacity. Recently, proponents of the so-called ‘new mechanist’ approach in philosophy of science have argued that a model of a cognitive capacity is explanatory only to the extent that (...) it reveals the causal structure of the mechanism underlying the capacity. If they are right, then a cognitive model that resists a transparent mapping to known neural mechanisms fails to be explanatory. I argue that a functiontheoretic characterization of a cognitive capacity can be genuinely explanatory even absent an account of how the capacity is realized in neural hardware. (shrink)