Scientific understanding is a fundamental goal of science. However, there is currently no good way to measure the scientific understanding of agents, whether these be humans or Artificial Intelligence systems. Without a clear benchmark, it is challenging to evaluate and compare different levels of scientific understanding. In this paper, we propose a framework to create a benchmark for scientific understanding, utilizing tools from philosophy of science. We adopt a behavioral conception of understanding, according to which genuine understanding should be recognized (...) as an ability to perform certain tasks. We extend this notion of scientific understanding by considering a set of questions that gauge different levels of scientific understanding, covering information retrieval, the capability to arrange information to produce an explanation, and the ability to infer how things would be different under different circumstances. We suggest building a Scientific Understanding Benchmark (SUB), formed by a set of these tests, allowing for the evaluation and comparison of scientific understanding. Benchmarking plays a crucial role in establishing trust, ensuring quality control, and providing a basis for performance evaluation. By aligning machine and human scientific understanding we can improve their utility, ultimately advancing scientific understanding and helping to discover new insights within machines. (shrink)

Artificial intelligent (AI) systems that perform image classification tasks are being used to great success in many application contexts. However, many of these systems are opaque, even to experts. This lack of understanding can be problematic for ethical, legal, or practical reasons. The research field Explainable AI (XAI) has therefore developed several approaches to explain image classifiers. The hope is to bring about understanding, e.g., regarding why certain images are classified as belonging to a particular target class. Most of these (...) approaches use visual explanations. Drawing on Elgin’s work (True enough. MIT Press, Cambridge, 2017), I argue that analyzing what those explanations exemplify can help to assess their suitability for producing understanding. More specifically, I suggest to distinguish between two forms of examples according to their suitability for producing understanding. I call these forms samples and exemplars, respectively. Samples are prone to misinterpretation and thus carry the risk of leading to misunderstanding. Exemplars, by contrast, are intentionally designed or chosen to meet contextual requirements and to mitigate the risk of misinterpretation. They are thus preferable for bringing about understanding. By reviewing several XAI approaches directed at image classifiers, I show that most of them explain with samples. If my analysis is correct, it will be beneficial if such explainability methods use explanations that qualify as exemplars. (shrink)

Is understanding subject to a factivity constraint? That is, must the agent’s representation of some subject matter be accurate in order for her to understand that subject matter? ‘No’, I argue in this paper. As an alternative, I formulate a novel manipulationist account of understanding. Rather than correctly representing, understanding, on this account, is a matter of being able to manipulate a representation of the world to satisfy contextually salient interests. This account of understanding is preferable to factivism, I argue, (...) mainly for simplicity reasons. While it explains the intuitive data about understanding as successfully as factivist accounts, it is simpler by virtue of reducing the value truth bestows on understanding to that of usability. (shrink)

In this paper, I sketch a new model for the format of the content of understanding states, Compressible Graph Maximalism (CGM). In this model, the format of the content of understanding is graphical, and compressible. It thus combines ideas from approaches that stress the link between understanding and holistic structure (like as reported by Grimm (in: Ammon SGCBS (ed) Explaining Understanding: New Essays in Epistemollogy and the Philosophy of Science, Routledge, New York, 2016)), and approaches that emphasize the connection between (...) understanding and compression (like Wilkenfeld (Synthese 190(6):997–1016, 2018)). I argue that the combination of these ideas has several attractive features, and I defend the idea against some challenges. (shrink)

This paper aims to clarify the nature of understanding in medicine. The first part describes in more detail what it means to understand something and links a type of understanding (i.e., objectual understanding) to explanations. The second part proceeds to investigate what objectual understanding of a disease (i.e., biomedical understanding) requires by considering the case of scurvy from the history of medicine. The main hypothesis is that grasping a mechanistic explanation of a condition is necessary for a biomedical understanding of (...) that condition. The third part of the paper argues that biomedical understanding is necessary, but not sufficient for understanding in a clinical context (i.e., clinical understanding). The hypothesis is that clinical understanding combines biomedical understanding of a _disease_ or pathological condition with understanding _illness,_ which involves some degree of personal understanding of the patient. It is argued that, in many cases, clinical understanding necessitates adopting a particular second-personal stance and using cognitive resources _in addition_ to those involved in biomedical understanding. (shrink)

This paper aims to clarify the nature of understanding in medicine. The first part describes in more detail what it means to understand something and links a type of understanding (i.e., objectual understanding) to explanations. The second part proceeds to investigate what objectual understanding of a disease (i.e., biomedical understanding) requires by considering the case of scurvy from the history of medi- cine. The main hypothesis is that grasping a mechanistic explanation of a condi- tion is necessary for a biomedical (...) understanding of that condition. The third part of the paper argues that biomedical understanding is necessary, but not sufficient for understanding in a clinical context (i.e., clinical understanding). The hypothe- sis is that clinical understanding combines biomedical understanding of a disease or pathological condition with understanding illness, which involves some degree of personal understanding of the patient. It is argued that, in many cases, clinical understanding necessitates adopting a particular second-personal stance and using cognitive resources in addition to those involved in biomedical understanding. (shrink)

By appealing to resources found in the scientific understanding literature, I identify in what senses idealisations afford understanding in the context of the (magnetic) Aharonov-Bohm effect. Three types of concepts of understanding are discussed: understanding-what, which has to do with understanding a phenomenon; understanding-with, which has to do with understanding a scientific theory; and understanding-why, which has to do with the reason some phenomenon occurs. Consequently, I outline an account of understanding-with that is suggested by the historical controversy surrounding the (...) Aharonov-Bohm effect. (shrink)

In this paper, I give an overview of different models of understanding attribution and advance a contextualist account of understanding attribution. Whereas other contextualist accounts make the degree in which the epistemic states of the relevant agents satisfy certain invariant conditions context-sensitive, the proposed account makes the conditions themselves context-sensitive.

Machine learning methods have recently created high expectations in the climate modelling context in view of addressing climate change, but they are often considered as non-physics-based ‘black boxes’ that may not provide any understanding. However, in many ways, understanding seems indispensable to appropriately evaluate climate models and to build confidence in climate projections. Relying on two case studies, we compare how machine learning and standard statistical techniques affect our ability to understand the climate system. For that purpose, we put five (...) evaluative criteria of understanding to work: intelligibility, representational accuracy, empirical accuracy, coherence with background knowledge, and assessment of the domain of validity. We argue that the two families of methods are part of the same continuum where these various criteria of understanding come in degrees, and that therefore machine learning methods do not necessarily constitute a radical departure from standard statistical tools, as far as understanding is concerned. (shrink)

In climate science, climate models are one of the main tools for understanding phenomena. Here, we develop a framework to assess the fitness of a climate model for providing understanding. The framework is based on three dimensions: representational accuracy, representational depth, and graspability. We show that this framework does justice to the intuition that classical process-based climate models give understanding of phenomena. While simple climate models are characterized by a larger graspability, state-of-the-art models have a higher representational accuracy and representational (...) depth. We then compare the fitness-for-providing understanding of process-based to data-driven models that are built with machine learning. We show that at first glance, data-driven models seem either unnecessary or inadequate for understanding. However, a case study from atmospheric research demonstrates that this is a false dilemma. Data-driven models can be useful tools for understanding , specifically for phenomena for which scientists can argue from the coherence of the models with background knowledge to their representational accuracy and for which the model complexity can be reduced such that they are graspable to a satisfactory extent. When citing this paper, please use the full journal title Studies in History and Philosophy of Science. (shrink)

What is human understanding and why should we care about it? I propose a method of philosophical investigation called ‘function-first epistemology’ and use this method to investigate the nature and value of understanding-why. I argue that the concept of understanding-why serves the practical function of identifying good explainers, which is an important role in the general economy of our concepts. This hypothesis sheds light on a variety of issues in the epistemology of understanding including the role of explanation, the relationship (...) between understanding-why and knowledge, and the value of understanding-why. I conclude that understanding-why is valuable and yet knowledge plays more important roles in our epistemic life. (shrink)

Elgin has presented an extensive defence of reflective equilibrium embedded in an epistemology which focuses on objectual understanding rather than ordinary propositional knowledge. This paper has two goals: to suggest an account of reflective equilibrium which is sympathetic to Elgin’s but includes a range of further developments, and to analyse its role in an account of understanding. We first address the structure of reflective equilibrium as a target state and argue that reflective equilibrium requires more than an equilibrium in the (...) sense of a coherent position. On the one hand, the position also needs to be stable between a ‘conservative’ pull of input commitments and a ‘progressive’ pull of epistemic goals; on the other hand, reflective equilibrium requires that enough of the resulting commitments have some credibility independent of the coherence of the position. We then turn to the dynamics of reflective equilibrium, the process of mutual adjustment of commitments and theories. Here, the most pressing internal challenges for defenders of reflective equilibrium arise: to characterize this process more exactly and to explain its status in relation to reflective equilibrium as a target state. Finally, we investigate the role of reflective equilibrium in Elgin’s account of understanding and argue that objectual understanding cannot be explained in terms of reflective equilibrium alone. An epistemic agent who understands a subject matter by means of a theory also needs to be able to use this theory and the theory needs to meet some external rightness condition. (shrink)

To understand something involves some sort of commitment to a set of propositions comprising an account of the understood phenomenon. Some take this commitment to be a species of belief; others, such as Elgin and I, take it to be a kind of cognitive policy. This paper takes a step back from debates about the nature of understanding and asks when this commitment involved in understanding is epistemically appropriate, or ‘acceptable’ in Elgin’s terminology. In particular, appealing to lessons from the (...) lottery and preface paradoxes, it is argued that this type of commitment is sometimes acceptable even when it would be rational to assign arbitrarily low probabilities to the relevant propositions. This strongly suggests that the relevant type of commitment is sometimes acceptable in the absence of epistemic justification for belief, which in turn implies that understanding does not require justification in the traditional sense. The paper goes on to develop a new probabilistic model of acceptability, based on the idea that the maximally informative accounts of the understood phenomenon should be optimally probable. Interestingly, this probabilistic model ends up being similar in important ways to Elgin’s proposal to analyze the acceptability of such commitments in terms of ‘reflective equilibrium’. (shrink)

Carnap and Goodman developed methods of conceptual re-engineering known respectively as explication and reflective equilibrium. These methods aim at advancing theories by developing concepts that are simultaneously guided by pre-existing concepts and intended to replace these concepts. This paper shows that Carnap’s and Goodman’s methods are historically closely related, analyses their structural interconnections, and argues that there is great systematic potential in interpreting them as aspects of one method, which ultimately must be conceived as a component of theory development. The (...) main results are: an adequate method of conceptual re-engineering must focus not on individual concepts but on systems of concepts and theories; the linear structure of Carnapian explication must be replaced by a process of mutual adjustments as described by Goodman; Carnap’s condition of similarity can be analysed into two components, one securing a relation to the specific extensions of the pre-existing concepts, one regulating the transition to the new system of concepts; these two criteria of adequacy can be built into Goodman’s account of reflective equilibrium to ensure that the resulting concepts promote theoretical virtues while being sufficiently similar to the concepts we started out with. (shrink)

We might think that thought experiments are at their most powerful or most interesting when they produce new knowledge. This would be a mistake; thought experiments that seek understanding are just as powerful and interesting, and perhaps even more so. A growing number of epistemologists are emphasizing the importance of understanding for epistemology, arguing that it should supplant knowledge as the central notion. In this chapter, I bring the literature on understanding in epistemology to bear on explicating the different ways (...) that thought experiments increase three important kinds of understanding: explanatory, objectual and practical. (shrink)

A central goal of research in explainable artificial intelligence (XAI) is to facilitate human understanding. However, understanding is an elusive concept that is difficult to target. In this paper, we argue that a useful way to conceptualize understanding within the realm of XAI is via certain human abilities. We present four criteria for a useful conceptualization of understanding in XAI and show that these are fulfilled by an abilities-based approach: First, thinking about understanding in terms of specific abilities is motivated (...) by research from numerous disciplines involved in XAI. Second, an abilities-based approach is highly versatile and can capture different forms of understanding important in XAI application contexts. Third, abilities can be operationalized for empirical studies. Fourth, abilities can be used to clarify the link between explainability, understanding, and societal desiderata concerning AI, like fairness and trustworthiness. Conceptualizing understanding as abilities can therefore support interdisciplinary collaboration among XAI researchers, provide practical benefit across diverse XAI application contexts, facilitate the development and evaluation of explainability approaches, and contribute to satisfying the societal desiderata of different stakeholders concerning AI systems. (shrink)

This paper attempts to revive the epistemological discussion of scientific articles. What are their epistemic aims, and how are they achieved? We argue that scientific experimental articles are best understood as a particular kind of narrative: i.e., modernist narratives (think: Woolf, Joyce), at least in the sense that they employ many of the same techniques, including colligation and the juxtaposition of multiple perspectives. We suggest that this way of writing is necessary given the nature of modern science, but it also (...) has specific epistemic benefits: it provides readers with an effective way to grasp the content of scientific articles which increases their understanding. On the other hand, modernist writing is vulnerable to certain kinds of epistemic abuses, which can be found instantiated in modern scientific writing as well. (shrink)

In recent years, scientific understanding has become a focus of attention in philosophy of science. Since understanding is typically associated with the pragmatic and psychological dimensions of explanation, shifting the focus from explanation to understanding may induce a shift from accounts that embody normative ideals to accounts that provide accurate descriptions of scientific practice. Not surprisingly, many ‘friends of understanding’ sympathize with a naturalistic approach to the philosophy of science. However, this raises the question of whether the proposed theories of (...) understanding can still have normative power. In this paper I address this question by examining two theories of scientific understanding: Jan Faye’s pragmatic-rhetorical theory and my own contextual theory of scientific understanding. I argue that both theories leave room for normativity, despite their naturalistic tendencies. The normative power of my contextual theory is illustrated with a case study of the chemical revolution. (shrink)

In recent years, scientific understanding has become a focus of attention in philosophy of science. Since understanding is typically associated with the pragmatic and psychological dimensions of explanation, shifting the focus from explanation to understanding may induce a shift from accounts that embody normative ideals to accounts that provide accurate descriptions of scientific practice. Not surprisingly, many ‘friends of understanding’ sympathize with a naturalistic approach to the philosophy of science. However, this raises the question of whether the proposed theories of (...) understanding can still have normative power. In this paper I address this question by examining two theories of scientific understanding: Jan Faye’s pragmatic-rhetorical theory and my own contextual theory of scientific understanding. I argue that both theories leave room for normativity, despite their naturalistic tendencies. The normative power of my contextual theory is illustrated with a case study of the chemical revolution. (shrink)