Eleanor Helms is Professor of Philosophy at Cal Poly, in San Luis Obispo, CA. She uses tools from the history of philosophy (especially phenomenology and existentialism) to study things we can’t see, especially in literary fictions, thought experiments, religions, and conspiracy theories.
A post by Eleanor Helms
Philosophy has been using thought experiments from its earliest beginnings. Recall Zeno’s Achilles and the tortoise, Plato’s ring of Gyges, and Lucretius’ spear thrown at the edge of the universe. In all that time, you might think philosophers would have come to some agreement about how they work. (Then again, have you met philosophers?) In my recent book Kierkegaard and The Structure of Imagination (Cambridge, 2025), I propose a new explanation of what thought experiments do. The book takes a deep dive into the history of the concept, all the way back to Kant, whose views also inspired the Danish thinkers Hans Christian Ørsted and Søren Kierkegaard. I argue that thought experiments provide what Kant called “cognition.” Kant famously worried that concepts without perceptions were empty (CPR A51/B75). Ordinarily, a concept like “cat” gets filled in by experiencing (wait for it . . .) a cat. But there are some concepts, like the idea of the universe as a whole, that can be thought but not cognized because there are no instances of “universe” that could appear in our experience. (We can experience objects in the universe, but not the universe itself.)
Unlike knowledge, cognition does not require taking anything to be true. Instead, cognition is a way of connecting concepts with sensations. It’s about meaning, not belief. But I argue in the book that cognition is still a valuable epistemic good. It’s a good thing all around for us to be minutely aware of how (or whether) our concepts attach to perceptible objects. However, there are some useful concepts that don’t map onto sensory objects directly. To cognize abstract concepts like the universe (or the self, or God, or freedom, or justice, or altruistic love . . .), we need some new and creative sources of cognition. But are other bases for cognition even allowed? Cognition is about filling in concepts with sensory content, specifically. Can we have a relation to sensory material that doesn’t belong to a perceived object? As strange as it might seem, Kant describes not one but three non-perceptual sources of cognition. In my book, I show how thought experiments can facilitate all three. They are: practical cognition, analogy, and construction (a method borrowed from geometry). The book focuses mainly on construction, but I’ll say a little about each one.
Thought experiments facilitate practical cognition when they describe a situation and ask me what I would do (e.g., Robert Nozick’s Experience Machine) or what someone should do. Practical cognition for Kant ultimately aims at morality, as the way to achieve the highest possible good. Many of the familiar ethical thought experiments fit here, as we have to choose what to prioritize (e.g., trolley problems). While it’s common in imagination debates to emphasize “central” perspective-taking or simulation, Kant doesn’t think we need to place ourselves within a situation imaginatively in order to judge it: what matters is just the filling in of abstract rules by representing them in a concrete context. We draw on practical cognition when we show children moral exemplars, which Kant thinks “elevate the soul and awaken a wish to be able to act in like manner oneself” (4:412n, and see Fremstedal 2022, Kaftanski 2024). What makes these exemplary actions moral (rather than, say, self-interested) won’t be outwardly visible, but the cases still lend their concrete material to moral principles by motivating us to reach for them in specific (real or imagined) situations.
Thought experiments can also work as analogies. One obvious contemporary case is Thomson’s sick violinist, where she builds an analogy between this kidnapping case and abortion. The thought experiment proposes that these materially dissimilar situations have analogous formal structures. Kant’s example of analogy is a hand mill as analogous to a despotic state (CPJ 5:352, see Pillow 2001). A hand mill does not instantiate a despotic state, but it turns our attention to its relevant structures in a perceptible (non-despotic) object.
Lastly, I argue that thought experiments in philosophy do the same work as construction in geometry. Hans Christian Ørsted, who was the first to use the term “thought experiment” in any language, proposes this connection. Geometers “construct” figures when they draw according to rules, using tools like a compass and straight edge. The resulting diagram does not perfectly instantiate the ideal figure (its lines will have some thickness, for example), but the process of drawing it still matters cognitively for the geometer. More than the result, the process of constructing the diagram exhibits some of its features as essential. I argue that thought experiments “construct” when they highlight relevant features of a situation to an imaginer and show why they have to be (or don’t have to be) a certain way. For example, Gettier problems reveal that not just any justification counts as knowledge. Frankfurt cases identify what matters for autonomy. As constructions, thought experiments guide attention to certain features of a target concept. Comprehending concepts in these three ways fills them in with concrete material, even when we cannot perceive instances of what we have in mind.
My overall claim is that thought experiments are mainly about achieving cognitions and (therefore) meaningful concepts. Alternatively, some contemporary accounts treat thought experiments as second-best placeholders for real (e.g., laboratory) experiments. That view is more common than one might think (Friedman 2002, Norton 2004, Buzzoni 2019), perhaps because so much of the recent scholarship on thought experiments has taken place in philosophy of science journals. The paradigmatic cases aren’t Plato, Thomson, and Gettier but Galileo, Einstein, and Schrödinger. In the sciences, it makes a lot of sense to say that thought experiments stand in for real experiments. But what about philosophy? Ethicists, epistemologists, and metaphysicians can’t usually confirm their hypotheses with empirical studies. Consider Frank Jackson’s Mary. Even if we could contrive a similar scenario, we would be left with Mary’s report of what she feels and whether, in her opinion, what she feels is new. In the same way, sending real trolleys down tracks would tell us what people tend to choose, but we’d be no better able to make the kinds of conceptual distinctions that Foot and Thomson want us to make.
Early work on thought experiments characterized philosophical uses of thought experiments as derived from scientific ones (Sorensen 1992). Until astonishingly recently (into the 1990s, and I found some as late as 2010), it was common to misattribute the first use of the term “thought experiment” (Gedankenexperiment) to Mach. The story told from there is that thought experiments act as (1) placeholders for future empirical experiences (Buzzoni 2019) and (2) access points for past experiences stored implicitly in memory (Gendler 2013). But H.C. Ørsted had used the term 80 years earlier—in 1812, in both German and Danish (Tanke-experiment)—to describe mental experiments in science, math, and philosophy. And unlike Mach, who was dismissive of Kant, Ørsted had a deep, life-long interest in transcendental philosophy, including the problem of cognizing abstract concepts.
Fair warning, the book takes a deeper dive into the history of philosophy than might appeal to everyone: Kant, Ørsted, and Kierkegaard—Ørsted’s contemporary, who also used the term (Tanke-experiment) well before Mach. And of course, the fact that Ørsted and Kierkegaard used the term first doesn’t guarantee a Kantian foundation is better. But I think it’s valuable to chart different paths through the history of philosophy. Roads not taken (or not taken by many) offer new ways to think about familiar things. In the Kant-inspired account I put forward in the book, I aim to shift debates away from how (or whether) thought experiments provide knowledge toward an appreciation for cognition as its own kind of epistemic good. Cognition allows us to establish common and durable points of reference. And it keeps being valuable even when we’re talking about things that don’t exist and never could. Thought experiments aren’t the only tool for communicating abstract ideas. Elijah Chudnoff argues that philosophy papers—even the ones without thought experiments—make complex mental experiences repeatable (2021: 201-03). But re-thinking a thought isn’t the same as cognizing it. Thought experiments are the best tools philosophers have, I argue, for tethering our thoughts to the material stuff of reality.
References
Buzzoni, Marco. 2019. “Thought Experiments in Philosophy: A Neo-Kantian and Experimentalist Point of View.” Topoi 38(4): 771–9.
Chudnoff, Elijah. 2021. Forming Impressions. New York: Oxford University Press.
Fremstedal, Roe. 2022. Kierkegaard on Self, Ethics, and Religion: Purity or Despair. Cambridge, UK: Cambridge University Press.
Fremstedal, Roe. 2014. Kierkegaard and Kant on Radical Evil and the Highest Good: Virtue, Happiness, and the Kingdom of God. London: Palgrave Macmillan.
Friedman, Michael. 2002. “Kant, Kuhn, and the Rationality of Science.” Philosophy of Science 69(2): 171–190.
Gendler, Tamar Szabó. 2013. Thought Experiment: On the Powers and Limits of Imaginary Cases. New York: Routledge.
Kaftanski, Wojciech. 2024. “Imagination, Mental Representation, and Moral Agency: Moral Pointers in Kierkegaard and Ricoeur.” Phenomenology and the Cognitive Sciences 23:179–198.
Kant, Immanuel. 1998. Critique of Pure Reason. Trans. Paul Guyer and Allen W. Wood. Cambridge: Cambridge University Press.
Norton, John D. 2002. “Why Thought Experiments Do Not Transcend Empiricism.” In Contemporary Debates in Philosophy of Science, pp. 44-66. Edited by Christopher Hitchcock. Oxford: Blackwell.
Pillow, Kirk. 2001. “Jupiter’s Eagle and the Despot’s Hand Mill: Two Views on Metaphor in Kant.” Journal of Aesthetics and Art Criticism 59(2): 193–209.
Sorensen, Roy. 1992. Thought Experiments. Oxford: Oxford University Press.