A research conversation with Hasok Chang and Andrea Hiott
Love & Philosophy · Research Conversations series: This episode is part one with philosopher of history and science Hasok Chang of the University of Cambridge
Episode summary
Philosopher of science Hasok Chang (University of Cambridge) joins Andrea Hiott to trace the through-line of his work — from Inventing Temperature (2004), through Is Water H₂O? (2012), to Realism for Realistic People (2022) — and to think together about what it means to treat knowledge as something we do rather than something we merely possess.
The conversation begins with why Inventing Temperature has quietly become an inspiration for neuroscientists: it shows how a felt quality (hot and cold) can become a trustworthy number (temperature) — not by discovering a pre-given fact, but through a messy, iterative, bootstrapping process of invention. From there, Chang and Hiott move through the ideas that organize his mature philosophy: epistemic iteration, operational coherence, active/activist realism, path dependence, conceptual engineering, and a pragmatist re-reading of truth that no longer leans on correspondence to an inaccessible reality.
A recurring theme: Chang describes each of his books as costing him “an article of faith” — first foundationalism, then monism, then the ideal of pure science — and how each loss opened onto a more dynamic, practice-based, and pluralist picture of knowing. Hiott connects these moves to her own navigational approach to mind and NAEC, where cognition is understood as a body’s coherent, aim-directed action in the world rather than the storage of static representations.
Hasok Chang is the Hans Rausing Professor of History and Philosophy of Science at the University of Cambridge. Trained first in physics (Caltech) and then in philosophy (Stanford), he is known for “integrated history and philosophy of science” and for what he calls complementary science — using historical and philosophical study to recover scientific questions that mainstream science has forgotten or foreclosed. His three main single-author books are Inventing Temperature (Oxford, 2004; winner of the 2006 Lakatos Award), Is Water H₂O? (Springer, 2012), and Realism for Realistic People (Cambridge, 2022). He is a founding figure of the Society for Philosophy of Science in Practice (SPSP) and the Committee for Integrated HPS.
Andrea Hiott is a philosopher and cognitive scientist developing the navigational approach to mind — the view that mind is “the body’s navigational action in all spaces of life, from the geographical to the emotional to the social to the virtual.” She is one of the hosts of the Love & Philosophy (Beyond Dichotomy) podcast and project, and her work draws on hippocampal research, 4E cognition (embodied, embedded, extended, enactive), and the ecological tradition. Her book Holding Paradox: A Navigational Approach to Mind and Consciousness is forthcoming. Her recurring interests — holding apparent contradictions open, “way-making,” constellation/kaleidoscopic thinking, and moving beyond either/or — frame her questions to Chang throughout as part of her research, whereby she references his work.
Books mentioned:
Chang, Hasok. Realism for Realistic People: A New Pragmatist Philosophy of Science. Cambridge: Cambridge University Press, 2022.
Chang, Hasok. Is Water H₂O? Evidence, Realism and Pluralism. Boston Studies in the Philosophy of Science. Dordrecht: Springer, 2012.
Chang, Hasok. Inventing Temperature: Measurement and Scientific Progress. Oxford Studies in the Philosophy of Science. New York: Oxford University Press, 2004.
Selected shorter works:
Chang, Hasok. "Operationalism." Stanford Encyclopedia of Philosophy (online), July 2009.
Chang, Hasok. "The Myth of the Boiling Point." Online hypertext paper, first posted 18 October 2007.
Key terms & concepts mentioned:
The definitions below are ordered roughly as the ideas surface in the conversation. They’re written to be listener-friendly while staying faithful to Chang’s own usage.
Epistemic iteration. Chang’s name for how inquiry actually makes progress: we don’t begin from a secure foundation, but knowingly start from an imperfect position and then use the results of inquiry to correct and refine that very starting point — a self-improving loop with no bedrock underneath. In Realism for Realistic People Chang invokes Peirce’s image that inquiry “is not standing on the bedrock of fact… It is walking upon a bog, and we can only say, this ground seems to hold for the present.” The aim is not coherence as such, but the ongoing improvement of coherence. (In the conversation, Chang stresses that he first thought iteration solved only the narrow problem of validating measurement standards, then came to see it as a very general pattern of learning.)
Operational coherence. Roughly, “a harmonious fitting-together of the elements and aspects of an activity, which is conducive to the successful achievement of the aims of that activity.” It is Chang’s practice-based replacement for truth as correspondence: instead of asking whether a proposition matches reality, we ask whether an activity’s parts hang together well enough to reliably achieve its aim. Coherence is always defined relative to an aim — without a stated or assumed purpose, the word “coherence” has no meaning. This relates to the navigational framework in that it means we always have to determine position and aim and landscape in any assessment.
Epistemic activity. “A more-or-less coherent set of mental or physical operations that are intended to contribute to the production or improvement of knowledge in a particular way, in accordance with some discernible rules (though the rules may be unarticulated).” Examples: measuring, detecting, predicting, hypothesis-testing, but also mundane actions like lighting a match.
System of practice. A coherent, interacting set of epistemic activities carried out with a view to achieving certain overall aims. (This is the unit Chang argues can be legitimately plural: several incommensurable systems of practice can coexist productively within one discipline.)
Active knowledge / “knowing how.” Chang’s term (also found in Realism for Realistic People) for knowledge that resides in ability and action rather than in propositions (this is where the pragmatism comes in) knowing how to do something rather than knowing that some statement is true. He argues proposition-based epistemology simply isn’t the right tool for evaluating methods, instruments, or institutions; for that we need to look at the coherence of what we do.
Active realism vs. activist realism (a terminology note discussed in the conversation, too). In Is Water H₂O? Chang introduced active realism: the view that realism shouldn’t be sitting back satisfied that we know a lot, but should be an attitude committed to continually learning more and better about reality. In Realism for Realistic People he renamed it activist realism, to avoid confusion with his separate notion of active knowledge (someone might wrongly hear “active realism” as “any assured possession of active knowledge”). “Activist” better captures the ever-learning, never-shrinking commitment. (The two terms are easy to run together in conversation; Chang clarifies how this has evolved for him in the conversation.)
Correspondence theory of truth (and Chang’s departure from it). The traditional view that a proposition is true when it corresponds to how the world “really is.” Chang’s two-part response: (1) For propositions, he brings correspondence “down to earth” — in real practice we only ever check the correspondence between two accessible things (a model’s prediction vs. an observed measurement; like the predicted vs. actual position of a star), never between our statements and an inaccessible metaphysical reality (Kant’s noumena, the “Platonic world”). (2) For active knowledge (”knowing how”), the very notion of correspondence “evaporates”, so the right question becomes operational coherence, not truth-as-matching. This fits well with an approach that understands any cognition as from some position in some landscape with some aim (which is the theory Hiott expresses in the navigational approach.)
Foundationalism. The idea that knowledge must rest on some ultimate, self-justifying foundation — either the fundamental laws of nature (a rationalist version) or raw sense-experience with final authority (an empiricist version). Chang says writing Inventing Temperature dissolved his attachment to foundationalism (he lost this aritlce of faith) because he saw that there is no ultimate authority in science; we start from what we’ve inherited and work our way out.
Monism (vs. pluralism). Monism is the assumption that there is a single true account of the world for science to converge on. Pluralism (Chang’s alternative, developed in Is Water H₂O?) holds that maintaining several coherent, sometimes incompatible systems of practice can be more productive than forcing premature consensus — he argues the historical evidence in the Chemical Revolution never actually compelled the single answer it’s remembered as compelling.
Operationalism. The view which was introduced by physicist Percy Bridgman in The Logic of Modern Physics (1927) is that a concept’s meaning is tied to the concrete operations (measurements, procedures) used to apply it. In Bridgman’s strict formulation, “the concept is synonymous with the corresponding set of operations.” Long dismissed as a naïve historical dead-end, it was not part of Chang’s physics or philosophy curriculum when he studied. Instead, he found Bridgman by accident. Chang recovers a non-strict, generative reading of operationalism where it is the striving to create more opportunities for learning about reality, and where a coherent multiplicity of operations is a richness of meaning rather than a problem to be eliminated.
Percy Bridgman (1882–1961). Nobel-laureate experimental physicist (high-pressure physics) at Harvard and originator of operationalism. In the conversation, Chang recounts finding a $5 used copy of The Logic of Modern Physics in a Berkeley bookshop as a graduate student and reading it on the train “like a revelation” — struck that a Nobel experimentalist could doubt everything about specific bits of physics (e.g. worrying whether the one-way speed of light there and back is really the same) with a rigor comparable to Descartes or Wittgenstein.
The one-way speed of light / relativity of simultaneity. Chang’s illustration of operational thinking: when we measure light’s speed we bounce it to a mirror and back, and simply assume the “there” and “back” speeds are equal — but we can’t independently check that. Einstein’s 1905 move (special relativity) was to show that “two events are simultaneous” is meaningless until you specify an operational procedure for determining it — and different procedures or reference frames yield different answers. This is the conceptual shock Bridgman reported having when asked to teach relativity at Harvard.
“Inventing” (why the title works). Chang notes Inventing Temperature wasn’t his title — he’d proposed The Philosophical Thermometer; an Oxford editor or reviewer suggested “Inventing.” He embraced it because invention is not “making shit up.” A genuine technological invention works only because we design it in a way nature allows to work — so invention is an “earnest engagement with nature.” We learn about the world by finding out which of our inventions (instrumental and conceptual) work out and which don’t. (Example from the book: which of William Thomson / Lord Kelvin’s competing definitions of absolute temperature survived, and why.)
Conceptual engineering. A now-popular movement in philosophy holding that our job isn’t only to analyze existing concepts but to improve them — to make them “more fit for purpose.” Chang traces its roots to Carnap’s notion of explication (tidying, sharpening, disambiguating a concept, not just describing it) and to the socially engaged strand associated with Sally Haslanger (rethinking concepts like race and gender). Chang’s twist: scientists have always been conceptual engineers — “what is science except continually inventing better ways of thinking?” — and every engineered concept is built for some epistemic aim (calculation, explanation, application, connection to other fields).
Instrumentalism (reconceived). Usually treated as the dull “opposite” of realism — the idea that theories are mere tools, not descriptions of reality. Chang wants to revalue it: if concepts and theories really are tools we do things with, that’s exciting, not deflationary — and it need not be the enemy of realism at all. As he puts it, instrumentalism is not really the opposite of realism.
Path dependence. The idea that where you can get to depends on where you started and the route you took. For Chang this is not incidental to scientific progress but essential to it: because inquiry proceeds by epistemic iteration from an inherited starting point, history is built into the structure of knowledge. “History is not just curious facts about the past… This path dependence of progress is essential to the kind of progress we have achieved.”
“The given” (reinterpreted). Not the traditional epistemologist’s “given” (raw, foundational sense-data), but what we have inherited — historically, biologically, and through the specific accumulated knowledge of a field. These are the starting points “it doesn’t make any sense to try to escape”; we begin with them and improve them.
John Dewey & the classical pragmatists. Chang credits Dewey with erasing the sharp boundary between content-learning and method-learning: we learn how to do things by doing them, and we learn method in the same process by which we learn content. This anti-dualism about knowing-that vs. knowing-how underlies Chang’s whole practice-based epistemology.
Ian Hacking — Representing and Intervening (1983). Hacking’s influential book whose very title splits science into two irreducibly different kinds of doing — representing the world and intervening in it. Chang draws on Hacking’s point that experiments are done for many different purposes (creating new phenomena as an end in itself, testing theories, clarifying what phenomena exist) — reinforcing that we must always ask what the aim of a given activity is.
Thomas Kuhn, paradigms & the descriptive/normative line. Chang is “strongly inspired by Kuhn” but deliberately avoids the word paradigm, which he thinks accumulated too much conflicting baggage. He reads Kuhn as opening up the question of the plural values scientists hold, then narrowing back to “problem-solving ability” without quite saying why that’s the point of science — leaving a “hidden normativity behind the surface descriptivity.” Chang’s alternative vocabulary — epistemic activities and operational coherence — is designed to keep aims explicit so they’re never forgotten.
Progress (aim-relative). For Chang, “we’ve made progress” means nothing until we specify what we’ve gotten better at, and to what end. Likewise “rigor”: people have “a very narrow notion of what rigor means” — the real questions are rigorous for what purpose, from what position, toward what aim.
Meaning, purpose & aim in science. A throughline: Chang argues that bringing aims explicitly into the center lets science talk rigorously about purpose and meaning without lapsing into mysticism. Hiott links this to science writer Philip Ball’s work on biology, meaning, and agency, and to her framing of these conversations under the (deliberately science-cringe-inducing) banner Love & Philosophy — the claim that what we care about in the moment legitimately shapes inquiry.
Further connections to Andrea Hiott’s “navigational approach to mind” work (whereby she references Chang and uses his ideas as inspiration, which is also the reason for these research conversations)
Hiott’s questions check with Chang’s philosophy of science to be sure she properly attributes credit to him for inspiration for her navigational approach to mind, where cognition is a body’s coherent, aim-directed movement through spaces — physical, emotional, social, conceptual. Several bridges surface explicitly in the conversation to be further explored in the future:
From felt quality to trustworthy measure. Why Inventing Temperature resonates with neuroscientists or philosophers of neuroscience: it models how something felt (hot/cold) becomes a number you can trust, without leaving the felt origin behind. Hiott is interested in Chang’s more iterative, bootstrapping picture of how a rigorous measure can grow out of lived experience.
Movement as the thing we can trust. Where losing foundationalism can feel like losing solid ground, Hiott reframes Chang’s “act of realism” as trustworthy precisely because it moves — a dynamic regularity “you find as you go,” not one “laid down at the start.” This maps with her navigational embodied approach to mind as ongoing rather than static representation.
Coherence over correspondence, applied to mind. Chang’s shift from truth of propositions to coherence of actions gives Hiott a rigorous, non-mystical vocabulary for treating mind as bodily action. In her “meta-application,” she applies operational coherence not to thermometers but to different operational systems — the aim-directed bodily motion (of an insect, a dog, a person) that, on her view, is what we are assessing as mind from some position with some aim in some landscape.
The science of colors, not thermal physics. A key distinction they draw: thermal physics can largely leave behind human bodily experience, but a science of mind (like a science of colors) cannot — subjective experience is the target, not a ladder to be kicked away. Chang suggests a relevant analogy is the physiology of temperature (which keeps the human) rather than thermal physics (which abstracts away from it).
Holding paradox / beyond either/or. Hiott’s signature move — allowing apparent contradictions and constellating from those rather than resolving them — echoes Chang’s pluralism (resisting forced monist consensus) and his refusal of the descriptive-vs-normative and realism-vs-instrumentalism either/ors. Both treat “constellation” or plural coherence as a strength rather than a failure to decide.
History / path dependence as constitutive. Hiott’s interest in each body’s “historical sensory trajectory” (her notion of phenomenological heritage) dovetails with Chang’s insistence that path dependence is essential — that where mind or knowledge can go is shaped by the route it has traveled.
Full Transcript (lightly edited)
[00:00] Andrea Hiott: A lot of people are thinking about these ideas of navigational mind now — there’s going to be a symposium, and a conference about it. It’s more neuroscience-related, and I could send you that; it’s a lot about propositional thought. But — well, thanks for being here, thanks for doing this.
Hasok Chang: You’re welcome. It’s wonderful to meet you.
Andrea Hiott: I’ve heard a lot of neuroscientists talking about Inventing Temperature, the book you wrote.
Hasok Chang: Really? That’s — what, twenty-two years ago now.
Andrea Hiott: I don’t know if you know this, but it’s a pretty inspirational book for people in neuroscience.
Hasok Chang: I didn’t know that. I know a couple of people in neuroscience who know it, but I didn’t know it was widespread at all. So what do they find useful in it?
Andrea Hiott: I think it’s because you show that you can start with something that is felt — and then actually get to a point where a felt quality can become a number you can trust, scientifically. Right now, a lot of the ways we do neuroscience use EEG or fMRI to talk to each other — and that’s not what you show happened with temperature. There’s also something about the iteration. It’s kind of messy, and the story you tell in that book — you never could have imagined it would get to where it got.
Hasok Chang: Right. And sometimes you don’t get anywhere very nice from the mess — sometimes the mess just remains messy. Other times you find ways of tying things together and tidying up, so you get a clear enough view of where you’ve got to. And then you can use that little [00:02] island of order as a platform to take the next step from.
Andrea Hiott: Exactly — and I think that’s part of the inspiration. Thinking of science that way isn’t how we necessarily learn it in school. You’re supposed to follow the linear path and find the answer, and if you don’t, you didn’t succeed.
Hasok Chang: Right. And generally — not just that book — what a lot of people find interesting in my work is that I tend to show that the allegedly simple things in the physical sciences are really not simple at all. When people say, “Oh, biology is so complex, living systems are messy” — well, yes, but something apparently simple [00:03] is also very complex once you start looking at it seriously. Temperature was one clear example.
Andrea Hiott: I wonder what’s resonated most with you over these years. The next book even — Is Water H₂O? — that title almost sounds like a paradox at first, which I like. A lot of what I’m trying to figure out is how we can hold what seems like contradiction and then see it differently. And I wonder whether that book became an invitation for you — to see the world, and what knowledge is, differently.
Hasok Chang: Yeah, certainly. [00:04] Something I’ve said to people is that every time I write a book, I lose an article of faith. In writing Inventing Temperature, what I lost was foundationalism, which I’d been very wedded to. I had this naïve notion as a physics student that we just find the fundamental laws of nature and the rest follows — that’s the foundation of knowledge. You can also have an empiricist version: the foundation of everything is experience, which has ultimate authority, and we build on it. But when you go through a story like the one in Inventing Temperature, [00:05] you realize there is no ultimate authority. We just have to get on with what we’ve got and find our way out of it to a better place.
And when I wrote the second book, Is Water H₂O?, the article of faith I lost was monism — the idea that there’s one world, one truth, and it’s the job of science to find it. So that book became a common reference point for people trying to think more pluralistically about science. And with the next one — Realism for Realistic People — [00:06] what I lost was the ideal of pure science: the idea that you can just have knowledge and then apply it to practical things. That’s why I went more positively into pragmatism there. Each major research project should be a way to teach yourself to see things differently. That’s what’s valuable about philosophy.
Andrea Hiott: It’s interesting you frame it as a loss, because as I read your books, it’s not a loss — it’s a reckoning. A clarification.
Hasok Chang: Right.
Andrea Hiott: [00:07] With foundationalism, when we lose it, it seems like we’ve lost the foundation. But your movement toward movement itself — the act of realism — actually starts to feel more like something we can trust. It’s not a foundation in a static sense, because it’s always moving, but there’s something regular about it.
Hasok Chang: Yeah — and I think it’s a more dynamic kind of regularity. Not a regularity laid down at the start; a regularity you find as you go. What you’ve been telling me about your work — this attempt to understand knowledge with a [00:08] sense of movement — is very much in line with how pragmatists have thought: knowledge is something we achieve as we go on doing things. Knowledge is what crystallizes out of coherent activities. You don’t start with a roadmap for how to achieve that coherence of action; you only learn it by twisting and turning. This is what I especially appreciate about John Dewey — he [00:09] erases the sharp boundary between content-learning and method-learning. For him, we learn how to do things by doing them, and we learn about methods in the same process. That’s a real lesson from the classical pragmatists.
Andrea Hiott: I think so too. And I wonder how you’d put grounding knowledge — I’m thinking of Percy Bridgman a little. I’m wondering when he came into your world; it seems like it was before you got into pragmatism, but maybe they were coming together.
Hasok Chang: Long before, yeah.
Andrea Hiott: I’d love to hear that, because for me it seems you’re grounding knowledge in what we do — if “grounding” is even still the right word — rather than in some higher reality. I feel Kant in you, in the framing aspect — that’s important when you get to mind-independence. And I also feel we don’t need the other world, the Platonic world, [00:10] as part of the explanation. I’d like to understand how that’s related to how you had these “losses,” and how Bridgman came into your life — this move into a more practical, active way of thinking, which is very different from Kant, and yet I can see how you could have gone through Kant to get there.
Hasok Chang: How I got where I am is a complicated story, probably not worth anyone else’s while. But Bridgman I found literally by accident. I’d heard of operationalism, and I’d heard his name, but he wasn’t part of any curriculum — neither in physics nor in philosophy. Operationalism was supposed to have been [00:11] popular in the early twentieth century, mistakenly, and something people quickly grew out of — that’s what I thought. But during the first year of my PhD in philosophy at Stanford, I happened to go to a used bookstore in Berkeley, where we sometimes went to hang out. And I came across The Logic of Modern Physics, Bridgman’s first book. I thought, “That’s going for five bucks — I’ll just buy it, why not?”
And reading it on the train back, [00:12] it was like a revelation. There are so many amazing things about Bridgman, but one that immediately popped out: here’s a Nobel Prize–winning experimental physicist who also knew how to doubt everything. It’s the same scale of resolute doubt you see at the opening of Descartes’ Meditations, or in bits of Wittgenstein — but Bridgman was doing it about really specific bits of physics. For example, he agonizes forever about the [00:13] one-way speed of light. When we measure the speed of light, we reflect it off something and back, measure the time, and assume the speed going and coming back is the same. But he says: how do we know it’s the same? We don’t.
That’s exactly the kind of thing Einstein taught us — that we don’t know what we’re saying when we say two events are simultaneous unless we specify the operational procedure for determining it. We have the metaphysical notion that if two things happen at the same time, they do, whether we know it or not. And what Einstein showed in his 1905 special relativity paper [00:14] is that such a statement is meaningless without a specific procedure — because different procedures give different answers. Einstein settles on the light-beam convention: take two points in space, put an observer at the midpoint, and if light flashes from two events arrive at the observer at the same time, the events are simultaneous. It sounds like an elaborate, pointless exercise — [00:15] but if you define it differently, or view it from a different frame of reference, you get a different result. That’s the relativity of simultaneity.
Bridgman goes through all of that. When his department at Harvard made him teach relativity — because someone was on leave — he read the original Einstein and had a conceptual shock. [00:16] We’d all naïvely believed Newton when he said that things like space and time are well known to everyone and don’t need to be defined — which is literally what he says in the Principia. Bridgman instead says: no, we have to define everything, and the most secure way of defining something is through the operations we actively perform — even if those operations are mental, or what he called “paper and pencil.” Sorry to go on —
Andrea Hiott: Please do, I love this. It says a lot about Inventing Temperature, too — that temperature is another thing we take for granted. It seems like you can’t have “inventing” in front of it.
Hasok Chang: That’s right. And in a way, philosophy is the art of doubting. If we don’t doubt [00:17] what’s taken for granted, where’s the philosophy? On the title, by the way — Inventing Temperature wasn’t my title. When I sent the manuscript to the press I’d called it The Philosophical Thermometer, which I thought was nice. When everything was accepted and we were about to go into production, the Oxford editor said, “We’re not sure about your title — what about Inventing Temperature?” And I immediately thought: that expresses precisely what I was trying to get at. To this day I don’t know whether it was my editor or a reviewer, [00:18] but it was a good idea.
When I explain the sense of that to people, I say: you may think invention is just making things up, but it’s not. An actual invention, in the technological sense, only works because we design something in a way nature allows it to work. Invention isn’t make-believe — it’s an earnest engagement with nature. We don’t know nature directly, but we learn something by learning which of our inventions work out and which don’t. So in Inventing Temperature you see many conceptual inventions as well as instrumental ones — [00:19] for instance, which concept of absolute temperature (William Thomson invents more than one) works out. The second one does, partly because it fits the new thermodynamic theory and partly because it agrees numerically with the older gas-thermometer standard. I didn’t put all of that in terms of operational coherence back in 2004, but that’s what it’s about.
Andrea Hiott: In the newer book you’re being more philosophical in a specific sense. The others are philosophical, but they’re historical — you tell it through history. This one [00:20] is about concepts and truth and reality — you’re facing the questions directly. And it makes Inventing Temperature make more sense to me. You’re putting words on things you had to assume were there.
Hasok Chang: I think that’s right. I haven’t written this up, but I recently gave a talk revisiting my work on measurement in terms of operational coherence, and it made sense — to me and to the audience. One thing I’ll add about “inventing”: through all these turns I’ve come to appreciate [00:21] conceptual engineering, as people call it in some philosophical quarters now. How much exposure have you had to that idea?
Andrea Hiott: It comes up a lot more in the past year, I’d say.
Hasok Chang: It’s becoming really popular, and for good reason. Part of the inspiration goes back to Carnap — his notion of explication, where he says that in explicating a concept we’re not only explaining what it means, but tidying it up, making it more precise, disambiguating, [00:22] improving it. He kind of leaves it there. The conceptual-engineering people picked that up, along with the more socially oriented strand from Sally Haslanger and others on race and gender. That point of view gives a more active sense of what we do in philosophy: we’re not just doing analysis, we’re trying to make our concepts more fit for purpose — which makes you ask, [00:23] what is the purpose? We often treat philosophy as a realm of eternal truths and forget to ask why we’re trying to solve a given problem.
But shift your gaze to science, and it’s obvious scientists have always been conceptual engineers. What is science except continually inventing better ways of thinking — new concepts, refinements of old ones? And scientific concepts are put to purposes: [00:24] calculation, explanation, application, connection with other fields. There are particular epistemic aims we’re trying to achieve with these concepts. This also puts a new spin on instrumentalism — usually seen as the not-very-exciting opposite of realism. But the idea that concepts are tools, theories are tools, and we do things with them — I find that really exciting.
Andrea Hiott: So do I. And you use the word inventing there, with concepts. You’re also a philosopher of history, and I think the historical aspect matters, because — as you said — [00:25] we’re not just coming up with things out of nowhere, randomly. There’s also how it relates to language itself. You brought up Carnap; there’s a confusion I’d like to see dealt with, about how much of this is about propositional knowledge, and how much is conceptual engineering in the sense of language. But either way — whether we think of it in a wider active-real sense or a more propositional, traditional sense — there’s a historical aspect, [00:26] which has a lot to do with how we’re going to use a concept in the moment we’re using it.
I bring that up because a lot of what we’ve been discussing can still sound like everything’s just floating. But with active realism and operational coherence, I feel this takes the historical aspect to heart — and gives us a way, as you were describing with Einstein, to see from different perspectives and still understand that the whole is real, is connected. That’s been a very hard tension to hold in philosophy and science. And I think you address it through — is truth correspondence? — and how you’ve eased into operational coherence instead. Does that make sense? Do you see what I’m trying to unpack — something in your work that doesn’t fit into instrumentalism or conceptual engineering as we typically think of them?
Hasok Chang: I’d hope instrumentalism or conceptual engineering could rise to meet this.
Andrea Hiott: [00:27] It could add something, expand — so we don’t have to choose between. But there’s something you’re offering that isn’t in that conception, as I’ve read it at least. Though I haven’t read everything.
Hasok Chang: Your emphasis on history is really important — a big part of my methodological work is worrying about how to bring history and philosophy of science together, but we won’t go into that now. At the end of your abstract you mention epistemic iteration, and I think that’s quite important. I didn’t realize this when I was writing Inventing Temperature: I thought epistemic iteration was [00:28] just how this particular problem — validating measurement standards — got solved. I thought it might have more general applicability, but I left it there. In the years since, I keep seeing it in other people’s work and my own. What I’ve come to realize is that there’s something very general about that iterative pattern. Whenever we go from some very imperfect starting point to something better — how else would it work? [00:29] Other than to start from what we’re given.
And this isn’t “the given” as traditional epistemologists mean it. It’s what we’ve inherited historically, what our biological constraints are, what we’ve learned so far in our specific fields. These are starting points it makes no sense to try to escape. We start with what we’re given, find clever ways to learn something, and that lets us come back and improve the starting point. Looked at this way, [00:30] history is part of what we’re trying to achieve — because without history the whole process has no place to exist. History isn’t just curious facts about the past, or incidental to what we’ve achieved. This path dependence of progress is essential to the kind of progress we’ve achieved. That’s the point.
Andrea Hiott: Very well said — path dependence is exactly the right term. I was also thinking that Is Water H₂O? starts to talk about systems of practice, which is a step toward operational coherence, isn’t it? It’s like turning anti-foundationalism into a stance. [00:31] And that’s where you start to develop active realism, right?
Hasok Chang: That’s right. There are some terminological difficulties, because I’ve been making up the terminology as my thoughts developed. That’s why in the recent book I relabeled my brand of realism as activist realism.
Andrea Hiott: May I ask you about that? It’s interesting.
Hasok Chang: What I meant by active realism already in Is Water H₂O? is that scientific realism shouldn’t be sitting back, satisfied that we know something well. It should [00:32] be an attitude, an outlook — even an ideology — committed to learning more and better all the time. Then, because in Realism for Realistic People I used the phrase active knowledge to mean, roughly, “knowing how,” I didn’t want active realism confused with that — someone might think active realism is just any assured possession of active knowledge. It’s not. So I changed it to activist to indicate more clearly that what we want is this ever-learning attitude, [00:33] rather than shrinking to preserve what we think we already know. But the sense of active realism from Is Water H₂O? is still there.
Andrea Hiott: It’s still there. And maybe this is a point where you could tell me about correspondence — because as I see it, realism isn’t about correspondence. It’s about an activity, and that’s a big difference — especially compared to traditional ways of thinking about knowledge. So in a way, that’s what the people in Inventing Temperature are doing: they’re not just finding pre-given things, they’re constructing a system that keeps helping them understand things better. [00:34] It’s not that we don’t find what’s true, in each moment from each position — but because it’s real, and because we’re all part of it, it’s going to change too. So it’s like finding ways to measure it accurately through the regularities that don’t change. I know that’s a complicated way to say it. What I’m really interested in is how you got from correspondence to — how to think about active knowledge such that we don’t need correspondence.
Hasok Chang: Right. There are two things to say about correspondence. One: even for propositional knowledge, I’m trying to get away from the standard [00:35] correspondence view of truth — correspondence between what our propositions or theories say and how the “Platonic world” out there is. Because it’s a common complaint that such correspondence is never checkable.
Andrea Hiott: And just for listeners — this is what science has kind of assumed: that we’re going to find things that are then always going to be true.
Hasok Chang: That’s how scientists often think. But if you look at their practice, yes, they’re always checking correspondence — but what are we checking? [00:36] The correspondence between two things that are both accessible. If we’re painting something in a naïve-realist way, we have the object as I see it and the object as I paint it, and we compare the two. But there’s no checking the correspondence of anything to an inaccessible reality — the kind metaphysical realists or Platonists imagine. So there’s always checking for correspondence, but it’s a mundane activity: checking the predictions of our models against [00:37] what we observe in the lab; checking predicted positions of stars against what actually appears tomorrow at 8 p.m. Both sides of the relation are accessible. So one thing I’ve tried to do is bring propositional correspondence down to earth. Correspondence is here and now — we know how to check these things. We’re not talking about correspondence between phenomena and noumena; that’s not the business of science.
That’s one side. The other: when we go into knowing how — active knowledge — the whole place of correspondence [00:38] evaporates. That’s how my thinking about pragmatism started, because I found standard proposition-based epistemology not useful for thinking about how scientific practices develop, or how we evaluate the quality of methodologies, instruments, even scientific institutions. There’s this whole other dimension of doing and knowing where a proposition corresponding to anything [00:39] isn’t going to get us far enough.
When we ask how to evaluate propositional knowledge, at least in principle we ask whether the propositions are true. But when we ask about the quality of active knowledge, that answer won’t cut it — which is why I started thinking instead about the operational coherence of what we do. So instead of the truth of propositions, we think about the coherence of actions. And when you think like that, correspondence enters only [00:40] peripherally into the notion of the quality of knowledge.
Andrea Hiott: Because one thing this is all measured against is: how do we know we’re progressing? Your artwork example is interesting — there’s the way we compare, “does it look like this?” But instead you can switch to: as I’m actively working — because we never stop creating this artwork; it’s never done, and the artwork itself is always changing — it’s like you’re trying to paint something that’s alive. So how do you think about whether you’re progressing? As you show: is each iteration achieving more coherence? It’s like we’ve shifted how we’re considering it. [00:41] And it’s incredibly helpful — instead of thinking there’s a static thing there that I have to find and replicate, which we all sort of know isn’t really how it works.
Hasok Chang: And it’s certainly not how modern people think about art. When we have the naïve view of art as reproduction of what we see, what we don’t think about is the question of purpose — why are we doing art? In that view it’s taken for granted that the purpose is to reproduce the shape we see. But we’ve left that behind in the world of art. [00:42] Likewise in science, we have to ask: what is the purpose of the activities we’re carrying out? People like Ian Hacking were very useful here — take the title of his famous book, Representing and Intervening: already there are two wholly different kinds of things scientists do. In his chapters on observation and experiment, he points out we do experiments for all kinds of reasons — sometimes creating new phenomena is a goal in itself, [00:43] sometimes we test theories, sometimes we just want a clearer picture of what phenomena there are. Same with theory, same with anything in science. The one thing I was trying to achieve with operational coherence is to bring the question of aims explicitly into the center — because coherence is only defined with respect to an aim. Without an aim that’s articulated or at least assumed, talk of coherence makes no sense.
Andrea Hiott: This is what I love. There’s a way of doing science that needs to open up and take all of that into consideration — instead of, say, in neuroscience where we’re doing [00:44] stimulus–response. It’s a lot more like the artwork example. And to do that, we really need these new ways of thinking, which I think you’re offering. I started this as research conversations, but I publish it under the title Love and Philosophy — which makes people in science cringe, on purpose, because we’re not supposed to talk about care and love. For me it has to do with what we care about in the moment. I bring it up because I talked to Philip Ball — he’s talking about biology through [00:45] meaning and aim, and it resonated with what you’re saying. A lot of people want to do that, because purpose and meaning clearly matter — but we’ve been told we shouldn’t use those words. Frameworks like operational coherence, and thinking of knowledge as active, give us a rigorous way to do that which doesn’t fall into mystical patterns.
Hasok Chang: Yeah.
Andrea Hiott: You can be mystical — but if you’re doing science, it’s like you have to be more rigorous, because you have to say from what position, toward what aim, in what context — rather than have all that assumed, which it is right now. It feels like what you’re showing with all your books: getting more rigorous about how we understand what we’re doing.
Hasok Chang: People often have a very narrow notion of what rigor means. Rigorous for what purpose, in what way? That’s the question. Same with progress: when we say we’ve made progress, that means nothing unless we specify what we’ve gotten better at, [00:46] and to what end. The whole philosophical debate on scientific progress had a sense of futility. Kuhn started opening it up when he talked about the different competing values scientists have — but then he wanted to bring it back down to problem-solving ability, without quite saying why that was the most important thing about science. [00:47] He’d have said, “Well, it just is,” and retreated to the position that he was only describing what scientists do, not dictating what they should do. That was a very fine line between the descriptive and the normative. In the end I think he was saying: science as we know it is great, so if we describe how that great enterprise works, we’re saying what’s actually good. It’s a hidden normativity behind the surface descriptivity.
Andrea Hiott: With paradigm, I feel he was opening up beyond either/or, [00:48] but it’s very hard for us to do that — so even his work gets lumped onto one side. Maybe some of that is pushback, to bring it back to the other side.
Hasok Chang: That’s one reason I’ve been strongly inspired by Kuhn but didn’t want to use the paradigm concept — it had accumulated too much baggage in all kinds of directions. With my talk of epistemic activities and operational coherence, I wanted to put aims at the center explicitly, so we never forget them.
Andrea Hiott: That’s so important — as is the way you tell what I think of as the real story of pragmatism, which is another thing that’s gotten distorted.
Hasok Chang: That’s a whole other issue.
Andrea Hiott: It depends where you are, on what position. [00:49] A lot of people have loved it forever and get it, but at different points of analytic philosophy or neuroscience it seemed like something else. That’s another discussion — but it has to do with that historical aspect: seeing things in a context, and seeing that we have to see the context each time. We don’t just assume there’s one context that always applies.
Hasok Chang: That’s right.
Andrea Hiott: That’s also the hard thing. I was going to ask you, in a general way — a lot of the frameworks start with sensation or feeling. There are many ways to use those words in neuroscience and psychology, but in Inventing Temperature it’s hot or cold, and then we get to temperature. That’s overly condensed —
Hasok Chang: That’s chapter one, yes.
Andrea Hiott: — but you don’t leave the feeling part behind.
Hasok Chang: No, never. You never leave it behind.
Andrea Hiott: It’s always part of it. But when I think about [00:50] something like mind, or consciousness, those words are a little different from hot/cold. And yet, trying to think about it through your frameworks is very helpful — until, at a point, it starts to feel like subjective experience itself is the target. There’s not going to be a ladder to kick away. So I had this thought: could there be an “inventing mind” way of thinking about this — if you know what I mean?
Hasok Chang: I guess some people would want that. All these post-humanists would think our mind was just an accidental step along the way to something greater —
Andrea Hiott: Oh, I really don’t mean it like that —
Hasok Chang: I know.
Andrea Hiott: That’s exactly the difficulty — I wouldn’t want it taken that way. It’s more almost linguistic, whereas [00:51] temperature is a measurement that’s not linguistic.
Hasok Chang: Something you said a minute ago is very important — that with the science of mind, it’s always going to be about subjective experience in some way. It has to be. With temperature, thermal physics isn’t really about our bodily experience of hot and cold. So the neuroscience you’re working in isn’t like thermal physics — it’s more like the physiology of temperature, where you’d [00:52] never want to leave behind the human experience. When it comes to thermal physics, scientists really want to leave the human origins behind, and to a large extent they can — not entirely. Another parallel would be optics: if you’re doing lasers, you can leave behind the visual experience humans started with. But if you’re doing the science of colors, it makes no sense to leave human experience behind. [00:53] I tend to think neuroscience or psychology is like the science of colors. It won’t mean anything if you leave behind the human experience that got it all started.
Andrea Hiott: I agree — and that’s why I’m trying to pick up that whole scaffold and shift it a bit, thinking about mind as a bodily action. If it already is a bodily action, you can’t leave the body behind. And you can have many different kinds of bodies — an insect, a dog, a person — each with a different bodily action that is mind. Then — and we’ll have to talk about this next time, because I’m too excited but I have to go — what happens when I use your work is almost a meta-application: I’m applying the coherence to those different operational systems, which would be the bodily motion in context, with an aim.
Hasok Chang: Yeah. Well — next time. Does that make sense? Yes. [00:54] Next time we talk I’d love to hear much more about that, and how you’re navigating those treacherous waters.
Andrea Hiott: That would be wonderful. I just want to express my appreciation. When you described finding the book in Berkeley, on the train — I’ve had a moment like that. A lot of what you’re doing is putting words on things a lot of people are trying to understand. So thanks for spending time with me.
Hasok Chang: Thank you, I appreciate that.
Andrea Hiott: Good luck with all your work. That talk you mentioned — relating operational coherence to Inventing Temperature — isn’t recorded, I guess?
Hasok Chang: I’ll ask the organizers whether it was recorded; I don’t think it was. I’d be happy to share the slides if that would be of interest.
Andrea Hiott: I’d love to see that — because the “meta” thing I described has to do with that iteration.
Hasok Chang: I’ll send that. Well — till next time.
Andrea Hiott: Till next time.
Hasok Chang: [00:55] Very good to meet you, Andrea. Thanks so much.
Andrea Hiott: Very good to meet you, and thanks so much for the time. Be well.
Hasok Chang: Take care. Bye.
Notable quotes & clips
“What is science except continually inventing better ways of thinking?” (~[00:23])
“The path dependence is essential.” (~[00:30])
“Every time I write a book, I lose an article of faith.” (~[00:04])
“Invention is not make-believe — it’s an earnest engagement with nature.” (~[00:18])
“Each major research project should be a way to teach yourself how to see things differently.” (~[00:06])
“The allegedly simple things in the physical sciences are really not at all simple.” (~[00:03])
“Philosophy is the art of doubting.” (~[00:16])
“Instead of the truth of propositions, we’re going to think about the coherence of actions.” (~[00:39])
Further reading & sources
Hasok Chang’s books
Inventing Temperature: Measurement and Scientific Progress (Oxford University Press, 2004) — winner of the 2006 Lakatos Award.
Is Water H₂O? Evidence, Realism and Pluralism (Springer / Boston Studies in the Philosophy of Science, 2012).
Realism for Realistic People: A New Pragmatist Philosophy of Science (Cambridge University Press, 2022).
Key sources referenced in the episode
P. W. Bridgman, The Logic of Modern Physics (1927) — origin of “operationalism” and the “operational definition.”
Hasok Chang, “Operationalism,” Stanford Encyclopedia of Philosophy.
Ian Hacking, Representing and Intervening (1983).
Thomas Kuhn, The Structure of Scientific Revolutions (1962).
Rudolf Carnap on explication; Sally Haslanger on conceptual engineering of race and gender.
John Dewey / classical pragmatism (Peirce, James, Dewey) — the “walking on a bog” image is Peirce’s.
Andrea Hiott’s work
Holding Paradox: A Navigational Approach to Mind and Consciousness (forthcoming).
“Radical Embodied Relation at Any Scale, from Remembering to Navigating,” Topoi (2025).
Love & Philosophy (Beyond Dichotomy) podcast & Substack; bibliography at andreahiott.com/inspirations.
