Michael T. Stuart

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.

An observation of Hume’s has received a lot of attention over the last decade and a half: Although we can standardly imagine the most implausible scenarios, we encounter resistance when imagining propositions at odds with established moral (or perhaps more generally evaluative) convictions. The literature is ripe with ‘solutions’ to this so-called ‘Puzzle of Imaginative Resistance’. Few, however, question the plausibility of the empirical assumption at the heart of the puzzle. In this paper, we explore empirically whether the difficulty we witness in imagining certain propositions is indeed due to claim type (evaluative v. non-evaluative) or whether it is much rather driven by mundane features of content. Our findings suggest that claim type plays but a marginal role, and that there might hence not be much of a ‘puzzle’ to be solved.

When philosophers discuss the possibility of machines making scientific discoveries, they typically focus on discoveries in physics, biology, chemistry and mathematics. Observing the rapid increase of computer-use in science, however, it becomes natural to ask whether there are any scientific domains out of reach for machine discovery. For example, could machines also make discoveries in qualitative social science? Is there something about humans that makes us uniquely suited to studying humans? Is there something about machines that would bar them from such activity? A close look at the methodology of interpretive social science reveals several abilities necessary to make a social scientific discovery, and one capacity necessary to possess any of them is imagination. For machines to make discoveries in social science, therefore, they must possess imagination algorithms.

Sometimes we learn through the use of imagination. The epistemology of imagination asks how this is possible. One barrier to progress on this question has been a lack of agreement on how to characterize imagination; for example, is imagination a mental state, ability, character trait, or cognitive process? This paper argues that we should characterize imagination as a cognitive ability, exercises of which are cognitive processes. Following dual process theories of cognition developed in cognitive science, the set of imaginative processes is then divided into two kinds: one that is unconscious, uncontrolled, and effortless, and another that is conscious, controlled, and effortful. This paper outlines the different epistemological strengths and weaknesses of the two kinds of imaginative process, and argues that a dual process model of imagination helpfully resolves or clarifies issues in the epistemology of imagination and the closely related epistemology of thought experiments.

Philosophical conceptual analysis is an experimental method. Focusing on this helps to justify it from the skepticism of experimental philosophers who follow Weinberg, Nichols & Stich. To explore the experimental aspect of philosophical conceptual analysis, I consider a simpler instance of the same activity: everyday linguistic interpretation. I argue that this, too, is experimental in nature. And in both conceptual analysis and linguistic interpretation, the intuitions considered problematic by experimental philosophers are necessary but epistemically irrelevant. They are like variables introduced into mathematical proofs which drop out before the solution. Or better, they are like the hypotheses that drive science, which do not themselves need to be true. In other words, it does not matter whether or not intuitions are accurate as descriptions of the natural kinds that undergird philosophical concepts; the aims of conceptual analysis can still be met

Originally published in 1991, The Laboratory of the Mind: Thought Experiments in the Natural Sciences, is the first monograph to identify and address some of the many interesting questions that pertain to thought experiments. While the putative aim of the book is to explore the nature of thought experimental evidence, it has another important purpose which concerns the crucial role thought experiments play in Brown’s Platonic master argument.In that argument, Brown argues against naturalism and empiricism (Brown 2012), for mathematical Platonism (Brown 2008), and from the Platonist-friendly, abstract universals posited by the Dretske-Tooley-Armstrong (DTA) account of the laws of nature to a more general, physical Platonism. The Laboratory of the Mind is where he takes this final step.

Imagination is necessary for scientific practice, yet there are no in vivo sociological studies on the ways that imagination is taught, thought of, or evaluated by scientists. This article begins to remedy this by presenting the results of a qualitative study performed on two systems biology laboratories. I found that the more advanced a participant was in their scientific career, the more they valued imagination. Further, positive attitudes toward imagination were primarily due to the perceived role of imagination in problem-solving. But not all problem-solving episodes involved clear appeals to imagination, only maximally specific problems did. This pattern is explained by the presence of an implicit norm governing imagination use in the two labs: only use imagination on maximally specific problems, and only when all other available methods have failed. This norm was confirmed by the participants, and I argue that it has epistemological reasons in its favour. I also found that its strength varies inversely with career stage, such that more advanced scientists do (and should) occasionally bring their imaginations to bear on more general problems. A story about scientific pedagogy explains the trend away from (and back to) imagination over the course of a scientific career. Finally, some positive recommendations are given for a more imagination-friendly scientific pedagogy.

Computational systems biologists create and manipulate computational models of biological systems, but they do not always have straightforward epistemic access to the content and behavioural profile of such models because of their length, coding idiosyncrasies, and formal complexity. This creates difficulties both for modellers in their research groups and for their bioscience collaborators who rely on these models. In this paper we introduce a new kind of visualization that was developed to address just this sort of epistemic opacity. The visualization is unusual in that it depicts the dynamics and structure of a computer model instead of that model’s target system, and because it is generated algorithmically. Using considerations from epistemology and aesthetics, we explore how this new kind of visualization increases scientific understanding of the content and function of computer models in systems biology to reduce epistemic opacity.

Scientists imagine for epistemic reasons, and these imaginings can be better or worse. But what does it mean for an imagining to be epistemically better or worse? There are at least three metaepistemological frameworks that present different answers to this question: epistemological consequentialism, deontic epistemology, and virtue epistemology. This paper presents empirical evidence that scientists adopt each of these different epistemic frameworks with respect to imagination, but argues that the way they do this is best explained if scientists are fundamentally epistemic consequentialists about imagination.

Recent research shows – somewhat astonishingly – that people are willing to ascribe moral blame to AI-driven systems when they cause harm [1]–[4]. In this paper, we explore the moral- psychological underpinnings of these findings. Our hypothesis was that the reason why people ascribe moral blame to AI systems is that they consider them capable of entertaining inculpating mental states (what is called mens rea in the law). To explore this hypothesis, we created a scenario in which an AI system runs a risk of poisoning people by using a novel type of fertilizer. Manipulating the computational (or quasi-cognitive) abilities of the AI system in a between-subjects design, we tested whether people’s willingness to ascribe knowledge of a substantial risk of harm (i.e., recklessness) and blame to the AI system. Furthermore, we investigated whether the ascription of recklessness and blame to the AI system would influence the perceived blameworthiness of the system’s user (or owner). In an experiment with 347 participants, we found (i) that people are willing to ascribe blame to AI systems in contexts of recklessness, (ii) that blame ascriptions depend strongly on the willingness to attribute recklessness and (iii) that the latter, in turn, depends on the perceived “cognitive” capacities of the system. Furthermore, our results suggest (iv) that the higher the computational sophistication of the AI system, the more blame is shifted from the human user to the AI system.

John D. Norton is responsible for a number of influential views in contemporary philosophy of science. This paper will discuss two of them. The material theory of induction claims that inductive arguments are ultimately justified by their material features, not their formal features. Thus, while a deductive argument can be valid irrespective of the content of the propositions that make up the argument, an inductive argument about, say, apples, will be justified (or not) depending on facts about apples. The argument view of thought experiments claims that thought experiments are arguments, and that they function epistemically however arguments do. These two views have generated a great deal of discussion, although there hasn’t been much written about their combination. I argue that despite some interesting harmonies, there is a serious tension between them. I consider several options for easing this tension, before suggesting a set of changes to the argument view that I take to be consistent with Norton’s fundamental philosophical commitments, and which retain what seems intuitively correct about the argument view. These changes require that we move away from a unitary epistemology of thought experiments and towards a more pluralist position.

Thought experiments are a means of imaginative reasoning that lie at the heart of philosophy, from the pre-Socratics to the modern era, and they also play central roles in a range of fields, from physics to politics. The Routledge Companion to Thought Experiments is an invaluable guide and reference source to this multifaceted subject. Comprising over 30 chapters by a team of international contributors, the Companion covers the following important areas: · the history of thought experiments, from antiquity to the trolley problem and quantum non-locality; · thought experiments in the humanities, arts, and sciences, including ethics, physics, theology, biology, mathematics, economics, and politics; · theories about the nature of thought experiments; · new discussions concerning the impact of experimental philosophy, cross-cultural comparison studies, metaphilosophy, computer simulations, idealization, dialectics, cognitive science, the artistic nature of thought experiments, and metaphysical issues. This broad ranging Companion goes backwards through history and sideways across disciplines. It also engages with philosophical perspectives from empiricism, rationalism, naturalism, skepticism, pluralism, contextualism, and neo-Kantianism to phenomenology. This volume will be valuable for anyone studying the methods of philosophy or any discipline that employs thought experiments, as well as anyone interested in the power and limits of the mind.

Can an ugly experiment be a good experiment? Philosophers have identified many beautiful experiments and explored ways in which their beauty might be connected to their epistemic value. In contrast, the present chapter seeks out (and celebrates) ugly experiments. Among the ugliest are those being designed by AI algorithms. Interestingly, in the contexts where such experiments tend to be deployed, low aesthetic value correlates with high epistemic value. In other words, ugly experiments can be good. Given this, we should conclude that beauty is not generally necessary or sufficient for epistemic value, and increasing beauty will not generally tend to increase epistemic value.

While philosophers hold that it is patently absurd to blame robots or hold them morally responsible [1], a series of recent empirical studies suggest that people do ascribe blame to AI systems and robots in certain contexts [2]. This is disconcerting: Blame might be shifted from the owners, users or designers of AI systems to the systems themselves, leading to the diminished accountability of the responsible human agents [3]. In this paper, we explore one of the potential underlying reasons for robot blame, namely the folk's willingness to ascribe inculpating mental states or "mens rea" to robots. In a vignette-based experiment (N=513), we presented participants with a situation in which an agent knowingly runs the risk of bringing about substantial harm. We manipulated agent type (human v. group agent v. AI-driven robot) and outcome (neutral v. bad), and measured both moral judgment (wrongness of the action and blameworthiness of the agent) and mental states attributed to the agent (recklessness and the desire to inflict harm). We found that (i) judgments of wrongness and blame were relatively similar across agent types, possibly because (ii) attributions of mental states were, as suspected, similar across agent types. This raised the question - also explored in the experiment - whether people attribute knowledge and desire to robots in a merely metaphorical way (e.g., the robot "knew" rather than really knew). However, (iii), according to our data people were unwilling to downgrade to mens rea in a merely metaphorical sense when given the chance. Finally, (iv), we report a surprising and novel finding, which we call the inverse outcome effect on robot blame: People were less willing to blame artificial agents for bad outcomes than for neutral outcomes. This suggests that they are implicitly aware of the dangers of overattributing blame to robots when harm comes to pass, such as inappropriately letting the responsible human agent off the moral hook.

Philosophical debate about the nature and function of thought experiments would be impoverished without good historical sources. And while valuable work is being done on the history of thought experiments, a comprehensive discussion of the history of philosophical investigation into thought experiments is still absent in the literature (but see Kühne 2005; Moue et al. 2006). In what follows we take the first steps towards providing a more complete picture of the diverse attempts to shed light on thought experiments.The term “thought experiment” made its first appearance about 200 years ago in 1811. The most prolific period in the history of the philosophical investigation into thought experiments is the current ..

In this article, we analyse the evidential value of the corpus of experimental philosophy. While experimental philosophers claim that their studies provide insight into philosophical problems, some philosophers and psychologists have expressed concerns that the findings from these studies lack evidential value. Barriers to evidential value include selection bias and p-hacking. To find out whether the significant findings in x-phi papers result from selection bias or p-hacking, we applied a p-curve analysis to a corpus of 365 x-phi chapters and articles. Our results suggest that this corpus has evidential value, although there are hints of p-hacking in a few parts of the x-phi corpus.

The philosophical literatures on models and thought experiments have been developing exponentially, and independently, for decades. This independence is surprising, given how similar models and thought experiments are. They each have “lives of their own,” they sit between theory and experience, they are important for both pedagogy and cutting-edge science, they galvanize conceptual changes and paradigm shifts, and they involve entertaining imaginary scenarios and working out what happens. Recently, philosophers have begun to highlight these similarities. This entry aims at taking the idea further, by trying to systematically identify places where insights from one literature can be taken up in the other. Along the way, important differences will also be highlighted.

In the last decades it has become clear that medicine must find some way to combine its scientific and humanistic sides. In other words, an adequate notion of medicine requires an integrative position that mediates between the analytic-reductionist and the normative-holistic tendencies we find therein. This is especially important as these different styles of reasoning separate “illness” (something perceived and managed by the whole individual in concert with their environment) and “disease” (a “mechanical failure” of a biological element within the body). While the demand for an integrative view has typically been motivated by ethical concerns, we claim that it is also motivated, perhaps even more fundamentally, by epistemological and methodological reasons. Evidence-based bio-medicine employs experimental and statistical techniques which eliminate important differences in the ways that conscious humans evaluate, live with, and react to disease and illness. However, it is precisely these experiences that underpin the concepts and norms of bio-medicine. Humanistic disciplines, on the other hand, have the resources to investigate these experiences in an intersubjectively testable way. Medicine, therefore, cannot afford to ignore its nature as a human science; it must be concerned not only with disease and illness, but also with the ways in which patients as persons respond to malady. Insofar as attitudes and expectations influence the criteria of illness and disease, they must be studied as part of the genuine subject matter of medicine as a human science. In general, we urge that this is a necessary step to overcome today's trend to split evidence-based and clinical medicine.

I claim that one way thought experiments contribute to scientific progress is by increasing scientific understanding. Understanding does not have a currently accepted characterization in the philosophical literature, but I argue that we already have ways to test for it. For instance, current pedagogical practice often requires that students demonstrate being in either or both of the following two states: 1) Having grasped the meaning of some relevant theory, concept, law or model, 2) Being able to apply that theory, concept, law or model fruitfully to new instances. Three thought experiments are presented which have been important historically in helping us pass these tests, and two others that cause us to fail. Then I use this operationalization of understanding to clarify the relationships between scientific thought experiments, the understanding they produce, and the progress they enable. I conclude that while no specific instance of understanding (thus conceived) is necessary for scientific progress, understanding in general is.

Scientists imagine constantly. They do this when generating research problems, designing experiments, interpreting data, troubleshooting, drafting papers and presentations, and giving feedback. But when and how do scientists learn how to use imagination? Across 6 years of ethnographic research, it has been found that advanced career scientists feel comfortable using and discussing imagination, while graduate and undergraduate students of science often do not. In addition, members of marginalized and vulnerable groups tend to express negative views about the strength of their imaginations and the general usefulness of imagination in science. After introducing these findings and discussing the typical relationship between a scientist and their imagination across a career, we argue that reducing the number or power of active imaginations in science is epistemically counterproductive. We suggest several ways to bring imagination back into science in a more inclusive way, especially through courses on imagination for scientists, good role models, and exemplar-based learning.

Imagination is important for many things in science: solving problems, interpreting data, designing studies, etc. Philosophers of imagination typically account for the productive role played by imagination in science by focusing on how imagination is constrained, e.g., by using self-imposed rules to infer logically, or model events accurately. But the constraints offered by these philosophers either constrain too much, or not enough, and they can never account for uses of imagination that are needed to break today’s constraints in order to make progress tomorrow. Thus, epistemology of imagination needs to make room for an element of epistemological anarchy.

John D. Norton defends an empiricist epistemology of thought experiments, the central thesis of which is that thought experiments are nothing more than arguments. Philosophers have attempted to provide counterexamples to this claim, but they haven’t convinced Norton. I will point out a more fundamental reason for reformulation that criticizes Norton’s claim that a thought experiment is a good one when its underlying logical form possesses certain desirable properties. I argue that by Norton’s empiricist standards, no thought experiment is ever justified in any deep sense due to the properties of its logical form. Instead, empiricists should consider again the merits of evaluating thought experiments more like laboratory experiments, and less like arguments.

Thought experiments, models, diagrams, computer simulations, and metaphors can all be understood as tools of the imagination. While these devices are usually treated separately in philosophy of science, this paper provides a unified account according to which tools of the imagination are epistemically good insofar as they improve scientific imaginings. Improving scientific imagining is characterized in terms of epistemological consequences: more improvement means better consequences. A distinction is then drawn between tools being good in retrospect, at the time, and in general. In retrospect, tools are evaluated straightforwardly in terms of the quality of their consequences. At the cutting edge, tools are evaluated positively insofar as there is reason to believe that using them will have good consequences. Lastly, tools can be generally good, insofar as their use encourages the development of epistemic virtues, which are good because they have good epistemic consequences.

The history of the philosophy of thought experiments has touched on the work of Kuhn, Popper, Duhem, Mach, Lakatos, and other big names of the 20th century, but so far, almost nothing has been written about Paul Feyerabend. His most influential work was Against Method, 8 chapters of which concern a case study of Galileo with a specific focus on Galileo’s thought experiments. In addition, the later Feyerabend was very interested in what might be called the epistemology of drama, including stories and myths. This paper brings these different aspects of Feyerabend’s work together in an attempt to present what might have been his considered views on scientific thought experiments. I conclude by contrasting Feyerabend’s ideas with two modern currents in the debate surrounding thought experiments: 1) the claim that the epistemology of thought experiments is just the epistemology of deductive or inductive arguments, and 2) the claim that the specifically narrative quality of thought experiments must be taken into account if we want a complete epistemology of thought experiments.

This is an introduction to a special issue of Perspectives on Science, the outcome of a workshop entitled "Thought Experiments in Science: Four Blind Spots," held at the University of Toronto, March 23rd, 2012. The recent revival in philosophical study of thought experiments has been limited to fields like epistemology, science studies, and metaphilosophy. With this issue we hope to facilitate a discussion about how some other disciplinary perspectives might bear on the subject; specifically, the history of philosophy, literary studies, phenomenology and cognitive science.