In his chapter in this volume, Christopher Pincock develops an argument for scientific realism based on scientific understanding, and he argues that Giere’s (2006) and my (2017, 2020) commitment to the context-dependence of scientific understanding or knowledge renders our views unable to account for an essential step in how scientists come to know. Meanwhile, in my chapter in this volume, I motivate a view that I call "causal pattern realism." In this response to Pincock's chapter, I will sketch a revised (...) version of Pincock’s argument for realism that is consistent with causal pattern realism. Then I will respond to Pincock’s concern that the context-dependence of understanding on my view would interfere with the scientific community’s ability to extrapolate from specific experimental and observational contexts as needed to develop knowledge. My goal is not to convince anyone to be a causal pattern realist but rather to create the space for such a view, taking into account the concerns motivating Pincock. (shrink)

Explainable artificial intelligence (XAI) is an emerging, multidisciplinary field of research that seeks to develop methods and tools for making AI systems more explainable or interpretable. XAI researchers increasingly recognise explainability as a context-, audience- and purpose-sensitive phenomenon, rather than a single well-defined property that can be directly measured and optimised. However, since there is currently no overarching definition of explainability, this poses a risk of miscommunication between the many different researchers within this multidisciplinary space. This is the problem we (...) seek to address in this paper. We outline a framework, called Explanatory Pragmatism, which we argue has two attractive features. First, it allows us to conceptualise explainability in explicitly context-, audience- and purpose-relative terms, while retaining a unified underlying definition of explainability. Second, it makes visible any normative disagreements that may underpin conflicting claims about explainability regarding the purposes for which explanations are sought. Third, it allows us to distinguish several dimensions of AI explainability. We illustrate this framework by applying it to a case study involving a machine learning model for predicting whether patients suffering disorders of consciousness were likely to recover consciousness. (shrink)

In this paper, I first outline the view developed in my recent book on the role of idealization in scientific understanding. I discuss how this view leads to the recognition of a number of kinds of variability among scientific representations, including variability introduced by the many different aims of scientific projects. I then argue that the role of idealization in securing understanding distances understanding from truth, but that this understanding nonetheless gives rise to scientific knowledge. This discussion will clarify how (...) my view relates to three other recent books on understanding by Henk de Regt, Catherine Elgin, and Kareem Khalifa. (shrink)

How can a reflective scientist put forward an explanation using a model when they are aware that many of the assumptions used to specify that model are false? This paper addresses this challenge by making two substantial assumptions about explanatory practice. First, many of the propositions deployed in the course of explaining have a non-representational function. In particular, a proposition that a scientist uses and also believes to be false, i.e. an “idealization”, typically has some non-representational function in the practice, (...) such as the interpretation of some model or the specification of the target of the explanation. Second, when an agent puts forward an explanation using a model, they usually aim to remain agnostic about various features of the phenomenon being explained. In this sense, explanations are intended to be autonomous from many of the more fundamental features of such systems. I support these two assumptions by showing how they allow one to address a number of recent concerns raised by Bokulich, Potochnik and Rice. In addition, these assumptions lead to a defense of the view that explanations are wholly true that improves on the accounts developed by Craver, Mäki and Strevens. (shrink)

This article addresses Simon Lohse’s and Daniel Little’s responses to my article “Is Social Ontology Prior to Social Scientific Methodology?.” In that article, I present a pragmatic and deflationary view of the priority of social ontology to social science methodology where social ontology is valued for its ability to promote empirical success and not because it yields knowledge of what furnishes the social world. First, in response to Lohse, I argue that my view is compatible with a role for ontological (...) theorizing in the social sciences. However, the view that results instrumentalizes social ontology. Second, in my response to Little, I argue that the same considerations I made in my article apply to naturalistic attempts to motivate a non-deflationary view, repeating some of the central issues of that article. (shrink)

We argue that the robustness analysis of idealized models can have confirmational power. This responds to concerns recently raised in the literature, according to which the robustness analysis of models whose idealizations are not discharged is unable to confirm the causal mechanisms underlying these models, and the robustness analysis of models whose idealizations are discharged is unnecessary. In response, we make clear that, where idealizations sweep out, in a specific way, the space of possibilities— which is sometimes, though not always, (...) the case—they can be holistically discharged. In turn, this can be used to show that the robustness analysis of idealized models can have confirmational force after all. (shrink)

I draw upon speech act theory to understand the speech acts appropriate to the multiple aims of scientific practice and the role of nonepistemic values in evaluating speech acts made relative to those aims. First, I look at work that distinguishes explaining from describing within scientific practices. I then argue speech act theory provides a framework to make sense of how explaining, describing, and other acts have different felicity conditions. Finally, I argue that if explaining aims to convey understanding to (...) particular audiences rather than describe literally across all contexts, then evaluating explanatory acts directed to the public or policymakers involves asking nonepistemic questions. (shrink)

Taking a cue from remarks Thomas Kuhn makes in 1990 about the historical turn in philosophy of science, I examine the history of history and philosophy of science within parts of the British philosophical context in the 1950s and early 1960s. During this time, ordinary language philosophy's influence was at its peak. I argue that the ordinary language philosophers' methodological recommendation to analyze actual linguistic practice influences several prominent criticisms of the deductive-nomological model of scientific explanation and that these criticisms (...) relate to the historical turn in philosophy of science. To show these connections, I primarily examine the work of Stephen Toulmin, who taught at Oxford from 1949 to 1954, and Michael Scriven, who completed a dissertation on explanation under Gilbert Ryle and R.B. Braithwaite in 1956. I also consider Mary Hesse's appeal to an ordinary language-influenced account of meaning in her account of the role of models and analogies in scientific reasoning, and W.H. Watson's Wittgensteinian philosophy of science, an early influence on Toulmin. I think there are two upshots to my historical sketch. First, it fills out details of the move away from logical positivism to more historical- and practice-focused philosophies of science. Second, questions about linguistic meaning and the proper targets and aims of philosophical analysis are part and parcel of the historical turn, as well as its reception. Looking at the philosophical background during which so-called linguistic philosophers also had a hand in bringing these questions to prominence helps us understand why. (shrink)

We argue that artificial networks are explainable and offer a novel theory of interpretability. Two sets of conceptual questions are prominent in theoretical engagements with artificial neural networks, especially in the context of medical artificial intelligence: Are networks explainable, and if so, what does it mean to explain the output of a network? And what does it mean for a network to be interpretable? We argue that accounts of “explanation” tailored specifically to neural networks have ineffectively reinvented the wheel. In (...) response to, we show how four familiar accounts of explanation apply to neural networks as they would to any scientific phenomenon. We diagnose the confusion about explaining neural networks within the machine learning literature as an equivocation on “explainability,” “understandability” and “interpretability.” To remedy this, we distinguish between these notions, and answer by offering a theory and typology of interpretation in machine learning. Interpretation is something one does to an explanation with the aim of producing another, more understandable, explanation. As with explanation, there are various concepts and methods involved in interpretation: Total or Partial, Global or Local, and Approximative or Isomorphic. Our account of “interpretability” is consistent with uses in the machine learning literature, in keeping with the philosophy of explanation and understanding, and pays special attention to medical artificial intelligence systems. (shrink)

Explainable artificial intelligence (XAI) seeks to produce explanations for those machine learning methods which are deemed opaque. However, there is considerable disagreement about what this means and how to achieve it. Authors disagree on what should be explained (topic), to whom something should be explained (stakeholder), how something should be explained (instrument), and why something should be explained (goal). In this paper, I employ insights from means-end epistemology to structure the field. According to means-end epistemology, different means ought to be (...) rationally adopted to achieve different epistemic ends. Applied to XAI, different topics, stakeholders, and goals thus require different instruments. I call this the means-end account of XAI. The means-end account has a descriptive and a normative component: on the one hand, I show how the specific means-end relations give rise to a taxonomy of existing contributions to the field of XAI; on the other hand, I argue that the suitability of XAI methods can be assessed by analyzing whether they are prescribed by a given topic, stakeholder, and goal. (shrink)

Scientific realism is a thesis about the success of science. Most traditionally: science has been so successful at prediction and guiding action because its best theories are true (or approximately true or increasing in their degree of truth). If science is in the business of doing its best to generate true theories, then we should turn to those theories for explanatory knowledge, predictions, and guidance of our actions and decisions. Views that are popular in contemporary philosophy of science about scientific (...) modeling and the centrality of idealization create several challenges for this traditional form of scientific realism. Yet the basic idea behind scientific realism that science has been and will continue to be epistemically successful is deeply appealing. This chapter explores the challenges posed by idealization and scientific modeling to motivate a scientific realism fully divorced from the idea that science is in the business of generating true theories. On the resulting view, the objects of scientific knowledge are causal patterns, so this knowledge only ever provides partial, simplified accounts of a complex reality. This variety of selective realism better accommodates the nature of our present-day scientific successes and offers an interpretation of scientific progress that resists the antirealist’s pessimism. (shrink)