Inventing Temperature: Measurement and Scientific Progress

by Hasok Chang

What is temperature, and how can we measure it correctly? These may seem like simple questions, but the most renowned scientists struggled with them throughout the 18th and 19th centuries. In Inventing Temperature , Chang examines how scientists first created thermometers; how they measured temperature beyond the reach of standard thermometers; and how they managed to assess the reliability and accuracy of these instruments without a circular reliance on the instruments themselves.

In a discussion that brings together the history of science with the philosophy of science, Chang presents the simple eet challenging epistemic and technical questions about these instruments, and the complex web of abstract philosophical issues surrounding them. Chang's book shows that many items of knowledge that we take for granted now are in fact spectacular achievements, obtained only after a great deal of innovative thinking, painstaking experiments, bold conjectures, and controversy. Lurking behind these achievements are some very important philosophical questions about how and when people accept the authority of science.

Genres: Science, Philosophy, History, Nonfiction, Physics, History Of Science, Academic

304 pages, Paperback

First published January 1, 2004

Reviews

[Alex · Rating 4 out of 5]

heard about from berna devezer on twitter

i am a huge fan of this style of scholarship, which chang calls "complementary science" (and distinguishes from sociology of scientific knowledge, internal history, methodology, and naturalistic epistemology). i am especially appreciative of the detailed discussion of measurement, operationalism, and coherentism. in statistics we often talk about the relationship between the real world and theoretical world in vague hand wave-y terms, and the additional exploration of this link was fascinating. this book really brings home the theory-ladenness of measurement, and also suggests iterative improvement of knowledge as more important than justification of existing knowledge.

followups for myself after reading:

- buy a copy of this book for my shelf

- read bridgman's logic of modern physics (1927), or preferably some less extreme coherentist response. the idea that two measurements of the length of an object by different operations are actually measuring differing underlying concepts is fascinating to me

- i would love to see some work on DAGs and operationalism. in a DAG do you encode theoretical model of the world, or an observational model of the world? or both at the same time, where theoretical constructs are unobserved and data is observed?

- explore what is lost when we give up on justification of knowledge in favor of iterative improvement of knowledge (i.e. 10% better each analysis rule)

- read chang's 1997 review of mayo and ampliative inference

- read stephen stigler and c.s. peirce and nancy cartwright

- in an applied sense, consider the measurement of structural properties of social networks and what kind of theory-ladenness is inherent in things like spectral clustering

- hennig and gelman suggest a constructive basis for statistics; foundationalism is out; it'd be interesting to have a catalog of currently accepted philosophies of science adapted to discuss statistical knowledge. mayo's severe tests i think are a useful tool but something about them doesn't feel like what statisticians actually do

- read more about measurement

[Jeffrey]

Read this slowly (let's say deliberately) over a year and a half. Each chapter can be savored, though the first chapter turned out to be the most compelling.

Chang takes something that seems so simple and ubiquitous -- the measurement of temperature -- and "shake[s] us out of theoretical complacency" by walking through the (shockingly recent and very long-lasting) scientific debates around the issue.

Chapter 1 is all about how calibration of fixed points was done for early thermometers. This chapter has some serious narrative momentum buoyed by many surprising facts. Early thermometers were calibrated based on things like body temperature and the freezing point of water. Maybe you think the freezing point of water is simple and straightforward (of course not, as one learns over and over and over again in this book) but clearly body temperature is too crude. Ok, so let's use boiling water. Even knowing -- as early scientists in this area did -- that pressure is an important factor for the boiling point, there are so many variables to account for. What water should be used? Distilled water will boil at a different temperature than sea water. What vessel should be used? Water in a metal pot will boil at a different temperature than in a glass beaker. How should the vessel be cleaned? Water boiled in a glass beaker will boil at substantially higher temperatures if the beaker is first cleaned with a strong acid. All of this was complicated tremendously by the lack of any coherent physical theory of heat (much less boiling).

In the end, scientists decided to use the temperature of steam formed just above boiling water as a repeatable, seemingly universalizable fixed point. But oh the irony! The temperature of steam only seems so reliable because the water vapor is condensing around particles in the air, and those just happen to be pretty universally dense in the air all around the places where scientists were doing these measurements. Only by being ignorant of this uncontrolled factor were scientists able to make progress on a seemingly intractable problem of infinite variation.

All of that work just to get one fixed point on a thermometer. Similar efforts had to be undertaken to clarify the appropriate freezing water fixed point. So, phew, we now have two fixed points. But that just takes us to the end of Chapter 1 because, of course, the next problem is how to generate sensible gradations between those two fixed points to indicate intermediate values. Do different materials -- air, mercury, water, etc. -- expand uniformly with temperature between those two fixed points? How would you even know? That's all treated in chapter 2.

Naturally, chapter 3 relates the history of thermometry above and below the freezing/boiling fixed points. What do you do at temperatures below the freezing point of mercury or above the melting point of your apparatus?

Chapter 4 talks about theory that finally came toward the end of the 19th century to try to make sense of the bewildering array of different attempts to measure temperature. This chapter contains a fascinating discussion of the ontological status of cold.

The history of cold is worth examining in some more detail ... There have been a number of perfectly capable philosophers and scientists through the ages who regarded cold as real as heat---starting with Aristotle. who took cold and hot as opposite qualities on an equal footing, as two of the four fundamental qualities in the terrestrial world. The mechanical philosophers of the seventeenth century were not united in their reactions to this aspect of Aristotelianism. Although many of them subscribed to theories that understood heat as motion and cold as the lack of it, the mechanical philosophy did not rule out giving equal ontological status to heat and cold. In the carefully considered view of Francis Bacon (1561-1626), heat was a particular type of expansive motion and cold was a similar type of contractive motion; therefore, the two had equal ontological status. Robert Boyle (1627-1691) wanted to rule out the positive reality of cold, but had to admit his inability to do so in any conclusive way after honest and exhaustive considerations, The French atomist Pierre Gassendi (1592-1655) had a more complex mechanical theory, in which "calorific atoms" caused heat by agitating the particles of ordinary matter; Gassendi also postulated "frigorific atoms, whose angular shapes and sluggish motion made them suited for clogging the pores of bodies and damping down the motions of atoms.

This reminds me of a public lecture I attended where a scientist was talking about climate change. An audience member, during the Q&A, made the comment that "it seems like the heart of the issue is that we need to figure out how to manufacture cold." "Oh how silly!" we all thought. But no less than Celsius himself might have had this same view!

The common attribution of the centigrade thermometer to the Swedish astronomer Anders Celsius (1701-1744) is correct enough, but his scale had the boiling point of water as 0° and the freezing point as 100°. ... My own hypothesis is that those who designed upside-down thermometers may have been thinking more in terms of measuring the degrees of cold than degrees of heat. If that sounds strange, that is only because we now have a metaphysical belief that cold is simply the absence of heat, not a real positive quality or entity in its own right. Although the existence of the upside-down temperature scales does not prove that their makers were trying to measure degrees of cold rather than heat, at least it reveals a lack of a sufficiently strong metaphysical commitment against the positive reality of cold.

Overall, the book was fascinating. The history parts were fantastic and quite readable. I would recommend skipping or skimming the more philosophical sections that end each chapter. They largely repeat the same material from the history, but in a less compelling way. I also think someone with a more pop-science bent could take this same material and really make a banger out of it.

Final quote that didn't fit in my review, but which I really liked because I see this fallacy all the time.

Bridgman warned that "our verbal machinery has no built-in cutoff": it is easy to be misled by the use of the same word or mathematical symbol in various situations into thinking that it means the same thing in all of those situations. (Bridgman focused most strongly on the unwarranted jump from one domain of phenomena to another, but his warning applies with equal force to the jump between the abstract and the concrete.) (Pg. 200)

[Alex Lipp · Rating 5 out of 5]

As good as everyone says! A historical & philosophical investigation of early thermometry. How do you create a thermometer when you don’t yet have a way to measure temperature? How do you “know” that boiling water always boils at the same temperature? Introduces “epistemic iteration” as a method to building scientific understanding. Excellent read for practising scientists… makes you realise how much of “scientific knowledge” is held together by string and tape.

[Floris · Rating 5 out of 5]

Chang’s Inventing Temperature is a classic in history and philosophy of science, so it was high time to try it out. It’s essentially a book about the value of history and philosophy of science pretending to be a book about the history of temperature. It’s based around six chapters. The first four are chunky history chapters, each centred around a central narrative and complemented with a shorter philosophical analysis section. The next two are two shorter programmatic philosophy chapters, the first concluding the findings from the previous four, whilst the last sets out a programme for a new style of scholarship Chang calls “complementary science”, of which his book is ultimately an example.

Chapter 1 is an account of how the two-point fixity (although mainly only boiling point) idea of temperature developed from the 17th century to the 19th century. The analysis section discusses what Chang calls “epistemic iteration” as reaching a particular standard in a step-by-step fashion without having a “correct” answer. Chapter 2 discusses attempts to find a scale for temperature. The analysis section examines “observability” as a philosophical problem (quotes classics like Van Fraassen, Hacking, Maxwell). Chapter 3 discusses the problem of defining temperature beyond the melting/freezing points of water, and the different suggestions that were made in the 19th century to solve it. The analysis underlines how accepting plurality in science involves accepting imprecision (158). Chapter 4 mainly looks at the work of William Thomson (Lord Kelvin) and James Joule, and asks how temperature changed from an abstract concept to a “real”, measurable thing. Chang uses the concept of epistemic iteration (Chap 1) to show how moving between the abstract-real scale can take place.

Chapter 5 summarises the arguments Chang has been making with respect to the history and philosophy of temperature. His main point is for science to accept circularity in measurement and that empirical science must be “coherentist”, i.e. scientific beliefs are ultimately not self-justifying but are justified in so far as they are part of a mutually supportive system of beliefs (citing Foley, 223). Coherentism and epistemic iteration is still progress, he argues (224). He also lists a range of epistemic values/virtues philosophers have identified as being indicative of the strength of scientific knowledge (quantity, variety, precision, simplicity, support by more general theories, ability to predict unknown phenomena, credibility relative to background knowledge, accuracy, consistency, scope, fruitfulness, elegance, completeness, unifying power, explanatory power, testability, fertility, neatness, conservativeness) , taking these as potential identifiers of whether epistemic iteration is leading to progress or not (227-8). He ends by providing a brief reflection on the unity of history and philosophy of science. Agreeing with Imre Lakatos that all historiography of science is philosophical by virtue of their abstract ideas, which are necessary ingredients of narratives, he dismisses the idea that philosophers only deal with the “abstract” whilst historians only deal with the “concrete” (233-4). Although he does warn against confusing “abstract” with “universal”.

Chapter 6 expands on his reflection on the merits of his HPS approach by outlining the nature and purpose of what he calls “complementary science”. His programmatic aim for complementary science is to act as a critical counterpart to modern-day science. Since scientists need to take many fundamentals and conventions for granted in order to do their work, they cannot spend much time reflecting critically on them. On the other hand, not reflecting critically on what one takes for granted is as Popper puts it a “danger to science” (237). Hence, complementary science is needed to do this critical work whilst letting the scientists do their day jobs, by kind of red-teaming science. The analogy Chang uses is that of capitalism: “it is the best-known economic system for ensuring high productivity and efficiency which, in the end, translate into the satisfaction of human needs and desires. At the same time, hardly anyone would deny the need for philanthropy or a social welfare system that ameliorates the inevitable neglect of certain human needs and the unreasonable concentration of wealth in a capitalist economy” (237-8). I’m not sure this analogy does him that many favours, but the essential principle is there: we operate a system whilst also trying to improve that system.

Doing complementary science is not particularly revolutionary, as it works with the familiar logic of making the familiar strange and the strange familiar. Chang describes how inquiry in complementary science can be initiated through reconsidering things that have been taken for granted in current science (his own approach in this book) and through focussing on unusual or puzzling elements in past science (239-40). But he goes further, arguing that complementary science can also contribute to scientific knowledge generation itself (through recovery, critical awareness, and making new developments known) (240-7). He also addresses the concern that complementary science might be too internalist (in its focus on also contributing to scientific knowledge generation), but dismisses this by pointing to its critical rather than prescriptive role for science (247-50).

Whilst the final 30 pages of this book are perhaps the most influential and contain the most fruitful passages for subsequent scholars, the first 220 are still excellent in providing the space to delve deep into historical and philosophical analysis. I love the way each chapter is centred around a deceptively simple question (what is temperature, how do you fix it, etc.), which is answered through a thoroughly researched “narrative” section and cogent “analysis” section. Each narrative is quite orthodox in terms of its subject material (which is composed primarily of well-established, white men of science from France and Britain in the 18th and 19th centuries), and quite “internalist” in the way it leaves out much of the social/cultural/visual/economic/etc. “externalities” of science that scholars are so used to considering part of science that it now looks strange when they aren’t mentioned. But, given that this book was published over two decades ago, had the explicit aim of reconsidering things taken for granted, and was designed to introduce a new kind of scholarly programme, I think it doesn’t deserve to be marked down for its orthodoxy or internalism.

Complementary science hasn’t caught on amongst historians and philosophers of science as much as Chang might have hoped (from what I can tell). Perhaps it’s the utilitarian (and perhaps somewhat subordinate) perspective it takes on modern-day “science” that has put off scholars who want their work to exist on its own merits? Maybe it’s the onward march of disciplinary specialisation that has not only made working with and alongside researchers from other disciplines harder but also less attractive (from a career point of view)? Maybe it’s not a problem with complementary science (or HPS, iHPS, etc.) per se, but with any kind of fundamentally interdisciplinary work? Max Dresow and Chang offer some not-overly optimistic nor pessimistic thoughts on the development of HPS and its prospects in Debating Contemporary Approaches to the History of Science, for those who want to add more maybes to this list.

[Ari · Rating 5 out of 5]

One part history, one part science, one part philosophy, mostly excellent. The history in particular is fascinating and gripping. I had never focused on how confusing and tricky the foundations of thermometry are. Suppose you're an 18th century scientist and you're making a thermometer. You have a volume of alcohol or mercury in a linearly graduated tube, calibrated to have freezing at 0 and boiling at 100. How do you know that the thing that makes the thermometer read 50 degrees is halfway between boiling and freezing for water? This isn't just a philosophical puzzler. The alcohol and mercury thermometers -- calibrated identically at the endpoints -- don't agree in the middle. How do you compensate?

It turns out that thermometry is *full* of tricky problems like this. When we say that water boils at 100 degrees, this turns out to be mostly wishful thinking. Boiling is a gradual process, not a sharp one (as anybody who has made pasta can picture). Further, it's possible to superheat water well above 100 degrees -- especially under laboratory conditions with clean glassware and a minimum of interruption.

This book has a bit of the feeling of a good whodunit. You get to watch smart people solve hard problems by thinking about the available evidence and interventions in highly unobvious ways. In the late 18th century, defining the boiling point was a sufficiently serious and urgent problem that the Royal Society appointed a blue-ribbon panel, chaired by Henry Cavendish, to solve it. Their approach was to use the condensation temperature of saturated steam, which turns out to be a more repeatable process than boiling.

The historical parts of the book are better than the philosophy. The philosophical sections are a bit heavy going -- ponderous, slightly self-important, and painfully explicit in asserting the obvious. I don't think I learned anything new about the scientific process in general. I could have told you going in that many of our definitions are somewhat circular but that we can use criteria like consistency, agreement between different modes of measurement, etc etc to assess our work.

Still, the book is well worth reading and the topic ought to be better known.

[Nicholas Martinez · Rating 5 out of 5]

By presenting the maturation of temperature and heat from naturalistic philosophy into scientific theory, Chang offers a healthy skepticism and genuine respect to evaluate the history of contemporary scientific virtues. He competently reports the assumptions and limitations—read: errors—of earlier investigations, and he reflects on their conditions that they were still able to generate productive results with a net progress toward modern science. Chang also challenges the current community to consider the merit of this history, in all scientific disciplines, in order to navigate the frontier of discovery, given that the earliest scientists often obtained their results despite reliance on premises and principles that were not recognizably scientific but characteristically underpin scientific values. It’s an especially important message in the climate of today’s scientific criticism, and Chang approaches it with care and candor.

[Ave Mets · Rating 5 out of 5]

Philosophy of science based on an extremely thorough, detailed history of measuring temperature.

[Kushal · Rating 4 out of 5]

Great demonstration of how precarious our certainties are, though it does belabour the point more than was probably needed

[James · Rating 4 out of 5]

Chang describes his goal as "complementary science" - a historical/philosophical effort to comb through the records of science, look at the questions asked and conclusions drawn, and retrieve knowledge or programs of study that were abandoned for historically contingent or epistemically unjustified reasons. Chang demonstrates "comp sci" on the invention of the modern thermometric scale - the difficulties in getting a proper fixed point with a reproducible temperature, the difficulties in extending temperature past where simple thermometers stop working, the invention of the theoretical Kelvin temperature and its empirical operationalization. The book is a detailed history of real science and concept-formation going down, Chang is a tremendously interesting thinker and writer, and successfully cracked open some new lines of interest in philosophy of science for me. I've had chances to use the concepts from this book in my own research thinking already in the time I spent reading it. Wish I knew more thermodynamics to follow the details of Kelvin's theory.

Notes:
• Is trying to do "complementary science". Like an ex-post distillation step for the concepts invented in the course of science, which can be performed without as much specialist training as the initial discovery phase.
○ “The conclusion of each episode takes the form of a judgment regarding the cogency of the answers proposed and debated by the past scientists, a judgment reached by my own independent reflections—sometimes in agreement with the verdict of modern science, sometimes not quite.”
○ “The intended audience closest to my own professional heart is that small band of scholars and students who are still trying to practice and promote history-and-philosophy of science as an integrated discipline.”
• “Today we tend to be oblivious to the great challenges that the early scientists faced in establishing the familiar fixed points of thermometry, such as the boiling and freezing points of water.”
• There are all sorts of edge cases of different boiling modes like the bubbling off of dissolved gases, hissing in large vessels, and bumping/superheating under high purity or high vessel smoothness. The "clean cases" were actually the most unclear, due to extreme conditions. However, luckily, the temperature maintained by ordinary-purity water boiling in ordinary-cleanness vessels turned out to be highly consistent. Steam temperature was also highly consistent, but there was much theoretical debate about why.
• The epistemic problem of fixed points was approached using ordinal thermoscopes, which were invented based on subjective temperature plus a presumption of independence.
• Principle of respect - simplicity of the instrument can eventually overrule sensory data. See: thermometer vs hands that have acclimated to various temperatures, distorting sensation.
○ I think the principle of respect becomes a lot less mysterious if you bring in parsimony. In the hands-switching case, others' sensations will disagree with mine, wax will still melt in one vs the other, all the other temperature-coupled properties stay the same except for my sensation with my hands. So sure, "the derived measurement can correct the grounding standard", but more because it's proven itself parsimonious with all other temperature-coupled phenomena, not because it's earned some mysterious "respect".
• There were some axes of variation, such as "degree of boiling" (aka how much is it boiling), which were widely observed in the 19c and dropped thereafter. Bolton tries to explain this as probably just a matter of other factors like solutes and vessel, but the author doesn’t buy it, thinks it’s just sweeping details under the rug
○ Rlly? Like yes it's sweeping details, but either it replicates or it doesn't man. Idk what was going on back then but if it doesn’t replicate in present day, clearly something weird was happening, doesn't matter exactly what it was except for history
• Alcohol and mercury thermometers disagreed in the middle range, even when calibrated using the same fixed points at the end. The conclusion was that their thermal expansion followed different curves.
○ Some people proposed a relativist standard, in which one thermometer was arbitrarily taken as standard, and thermometers using the other substance would have to be calibrated in between to match.
○ Another more abstract position was to take "whichever metric made the laws of thermal phenomena as simple as possible"
○ But in practice, the intuition all the experimentalists settled on was that there existed some "real" temperature, and one thermometer would be right and all others wrong
• This opens up the Problem of Nomic Measurement: how do you know which fluid is the one that expands linearly, if you have no way to apply uniform increments in temperature?
○ I predict this will be resolved via parsimony
• Several eminent experimentalists spent "some time" making ungrounded/unreflective assertions about which fluids expanded uniformly and which didn't
• Method of mixtures is a great approach, and initially showed mercury to be the most linear. However, it doesn’t actually evade the PoNM, because it assumes uniform thermal capacity across the temperature range
• The chemical theory of caloric also had an objection: latent heat was known, but it wasn't known that latent heat was absorbed only at the phase transition temperatures. So the chemical calorists needed to know what fraction of the caloric was measurable by thermometer at any given temperature.
○ Huge unobservable degree of freedom, provides lots of room for epicycles
• “Regnault's career is worth examining in some detail, since the style of research it shaped is directly relevant to the scientific and philosophical issues at hand.”
○ Known for assiduous detail and expensive measurement devices
○ Positivist-flavored bent, tried to make measurements without routing through theoretical assumptions.
• Stripped thermometer calibration down to comparability, and took that to the extreme. Most experimentalists were happy to assume that once property endpoint-calibrated, thermometers would behave equivalently.
• Regnault's comparability analysis found huge variations up to 5C between mercury thermometers, depending on the expansion behavior of the glass, which was too finicky to standardize at the time. Decided to use air over mercury, because expansion coefficient was much larger, enough to swamp that of the glass
• Also found that while air and many gas thermometers were very consistent, sulfuric acid gas was not, ruling out the linear-expansion arguments made from first principles of gas kinetics. Air thermometers were useful, but not theoretically privileged.
• Epistemological aside on Regnault and ontological principles
○ Examines Regnault's method for attacking the problem of nomic measurement
○ Says Regnault recognized that the ordinal stage-2 standard wouldn't be sufficient to define a numerical stage-3 standard, and that any proposal which required a s3 standard to verify would be circular. So he arrived at comparability.
○ Says that the virtue of comparability is that it basically makes only one assumption - that temperature must have only a single fixed value in any given circumstance
○ Goes on to ask why this assumption is justified, and ends up considering it an "ontological principle", justified by neither logic nor experience
○ Here I disagree. Single-valuedness of temperature comes from the Stage 1 standard being single-valued. The S3 standard is supposed to be a formalization of sensory perception of hot/cold, which only has a single value at a time, in experience. If the proposed formalization has multiple values, it's clearly not a very good formalization of the relevant property, unless it's collapsible to a single-valued prediction of the sensory experience somehow, in which case just use that
• Holism problem: “It is not that we propose a theory and Nature may shout NO. Rather, we propose a maze of theories, and Nature may shout INCONSISTENT.” This means that whenever we assume XY and falsify Z, the falsification can be redirected at X^Y. Regnault avoids this by minimizing (but not eliminating, as the positivists would want) assumptions needed.
○ Some assumptions still needed: ability to read thermometers correctly, for example.
• Notes that Regnault got sort of lucky - if multiple disagreeing standards had each proved consistent, he wouldn't have been able to identify a "best" standard.
• Wait I literally don't understand how Regnault solved the problem of getting a numerical standard. Consistency is required even for a stage 2 standard. And he would still have no way of knowing if there were some simple laws that involved multiplicative scaling with temperature, but wouldn't appear so if measured by a consistent but nonlinear thermometer.
○ It still seems to me like PoNM requires some sort of parsimony argument
• At some point bimetallic thermocouples become available???
• Low-temperature extension found by consistency+agreement of air thermometers, alcohol thermometers, and the bismuth-copper thermocouple in the 0 to -80 C range
• Attempted high-temperature extension with metallic expansion and Newton's law of cooling, but still led to circularity/holism problems. Extension of governing behavior to high temperatures couldn't be verified.
○ Inter-method agreement would do it, wouldn't it?
• Wedgwood's pyrometer was highly successful. Worked by measuring the shrinkage of a ceramic caused by vitrification, which seemed to produce a consistent result that was a function of only the highest temperature to which it had been exposed for at least three minutes. Produced a stage-2 ordinal scale, but could not yet be compared with numerical temperature.
○ Used expansion of silver as an intermediate, calibrating a silver-to-fahrenheit and silver-to-wedgwood conversion ratios with two-point calibration, assuming linearity of mercury, silver, and his ceramic
○ Bruh
• “I will show that each of the temperature standards favored by Wedgwood's critics was as poorly established as Wedgwood's own. Their main strength was in their agreement with each other.”
• Bridgman's "operationalist" mindset considers measurement procedure as constitutive of the quantity being measured. If two different measurement operations produce the same values within a particular domain, this can sometimes justify the linguistic convenience of using the same word to refer to their values.
○ Very positivist, maximally cashed out in terms of empirical predictions.
○ I think this is actually the way I would defend quantities like length, if pressured for rigor - what it means for an object to be X long is that if we applied Y procedure of repeated yardstick translation to it, we'd observe outcome Z
• Bridgman says that a light-year is a very different kind of thing from a meter, because they're measured differently. Surely there's some room here for predictions about in-principle measurement procedures using yardsticks, though?
○ Similarly, bridgman says that the diameter of an electron being 10^-13 cm is meaningless, or means something very different from conventional lengths, because we can't use rulers to measure that - only the tiny units of length that are only used in electrodynamic contexts.
○ But, like, bruh surely you could just do some linking procedure and still have your notion of length have an experimentally cash-out-able meaning
○ Or, if you insist on the procedure being "uniform", maybe the smallest scale of length can in principle be scaled up to measure longer lengths, but in practice we find it convenient to anchor the concept in our day-to-day range?
• Validity is only worth debating if the meanings involved are not exhausted by their definitions. Otherwise, they're all true by tautology.
• Foundationalism (standard X justified by standard Y) vs coherentism (standards X&Y justified over standard Z due to their agreement) vs mutual grounding (X/Y/Z are allowed to coexist, and can prove respective virtues as we go)
• Watt improved the steam engine by noticing three places where the "work capacity" of the steam wasn't being used correctly. Too much water was being injected, cooling the cylinder too much and wasting work on the next cycle. The condenser had to be separated to prevent the high-pressure steam from resisting the compression of the piston. And finally, Watt noticed the steam actively rushing into the condenser, suggesting that not all of the expansive work in the heated steam had been used to push the piston, and extended the push phase.
• Whoa Rumford's experiment seems great! Falsifies theory of caloric by generating heat via friction. Problem is, it could be evaded by this slippery "latent caloric" dodge, so it didn't topple the caloric theory
• Existence of Carnot cycle let quantities of heat and intervals of temperature be tied to quantities of work, since a Carnot engine put out a constant amount of work per heat at a given temperature diff.
• Note to self: I keep thinking the thing to do is "go with the scale that makes the math cleanest", and thinking that "go for the true scale" is metaphysically confused. But the "cleanest math" formulation is kind of hard to work towards in practice, and maybe the practical way to do it is something like "pull on the intuitions that get at the most fundamental-seeming aspects of what temperature is, and try to formalize those into a numerical scale", which looks a lot like the "go for the true scale" attitude
• Since Kelvin's temperature was defined in terms of the Carnot cycle, there was a question about how to measure it. You could try building a Carnot engine, but that's intractable
○ Related thoughts I had as a kid: when you say "X is 1m long", presumably what you mean is something like "if you put X next to a meter stick, they would line up at both ends." But then how do you measure the length of a fire? A beam of light? A black hole? Can you actually come up with a generalized procedure for measuring length that handles all edge cases?
• The Joule-Thompson experiment found a way to measure Kelvin temperature, using the properties of ideal gases. A novel theoretical result showed that the expansion of an ideal gas should be isothermal. But non-ideal gases have IMFs to break, so they're slightly endothermic. Joule-Thompson experiment started by just using the mercury thermometer, measured the cooling of air, and then made a correction to the air thermometer based on the non-idealness of air (because an ideal gas thermometer shows Kelvin temperature).
• If it stopped there, it would just be another circularity. But when they used the corrected air thermometer to perform the experiment again, they could get a correction to their correction. And this process empirically seemed like it was very quickly convergent!
○ Whoaaa what the hell
• Chang re-explains foundationalism (start with a self-justifying base, deduce consequences) vs coherentism (a belief is justified iff it belongs to a set of self-consistent beliefs)
○ Obvious critique of coherentism jumps out at me which is like "ok well I believe that I can fly" or smth, and then I just make more and more ridiculous claims to stay consistent, and you're forced to say my beliefs are "justified"
○ Obvious response to that critique is that the real thing to do is just abandon "justified" as a standard and instead just call coherent beliefs coherent beliefs, and say the goal of epistemology is a set of coherent and reality-consistent beliefs, rather than "justification". Ontology not truth-apt.
○ This seems to me now obviously correct, whereas coherentism was a bit counterintuitive to begin with
• Chang then goes into a whole bunch of discussion of "epistemic virtues" which seems to me shockingly sloppy and semantic, but makes a good point (or rather, cites literature whose conclusion I intuitively agree with) that "truth" might be the end point of epistemology but it's not a good bar for progress, because it's hard to come up with good standards of partial truth, so you need to think about progress bars like these "virtues". Damn that's an interesting problem, I can give the "virtues" a pass as early-stage intuitive hooks.
• Talks about enrighment and self-correction, citing sensation->thermoscopes->thermometers as an example of enrichment and the Joule-Thompson experiment as an example of self-correction. The discussion seems to me hand-wavy but I have to agree that both of those examples are compelling.
• In discussing the source of interest in philosophy of science, a great organic articulation of the feeling of tension! Sazen-y tho. "What drove me initially into this field and still drives me on is a curious combination of delight and frustration, of enthusiasm and skepticism, about science."
• Damn Chang is just an incredibly interesting and sophisticated thinker. Reading this section on complementary science feels like reading something that a particularly flourishing alternate version of me would have written.
• Complementary science feels like it's not going to take off, because precisely the thing that it does is comb through old science for "the facts/avenues that got de-emphasized in the search of the most productive set of simplifying assumptions" and focuses on those. It's not the fastest way to the most knowledge. But from a pure epistemological standpoint, complementary science seems like a totally natural "completionist" activity. Seems like Chang has picked out something that's simple, real, and in a sense important.

[DW Rowlands · Rating 5 out of 5]

This is a dense, academic text, but I still found it extremely interesting. Its primary topic is the history of how natural philosophers and physicists developed an understanding of what temperature meant and, more importantly, an operationalization for measuring it in an increasingly broad range of circumstances. This is, itself, a fascinating story of boot-strapping: how do you calibrate a thermometer to fixed points (like the boiling point of water) when you have no way to measure whether water truly always boils at the same temperature and, in fact, due to phenomena like super-heating, it doesn't always boil at the same temperature, especially for different investigators using different experimental set-ups?

However, Hasok Chang uses this story for something more: an attempt to demonstrate the value of what he calls "complementary science," a mode of history and philosophy of science that endeavors to examine questions that, due to specialization, modern science cannot directly approach, and to understand the causes and natures of phenomena that have been eliminated as confounding factors by cutting-edge experimentation. (This includes a paper he wrote two years before the book was published, regarding some 19th Century experiments that were seen at the time as the evidence of the radiant spread of cold, which seem not to have been fully followed-up on or explained because the underlying theoretical model of "caloric" on which they were based was rejected.)

[Lakmus]

Interesting telling of the details of how temperature measurement was developed. The philosophy parts got me shrugging, because it seemed like a fairly obvious description of how science proceeds (retrospectively, hindsight is 20/20). Like the idea of 'complementary science' being done within history/philosophy of science because sometimes it's nice to sit down and have some guy already summarise the field for you, so you don't have to find out nine years into your PhD about some set of studies in 1954 that practically were onto the same thing as you, and also pause and reflect on what's already been done to an extent that isn't typically captured by review papers or textbooks (in my field at least).

Could also have been a solid hundred pages shorter tbh. It ain't quite *that* deep.

[Stephen Perry · Rating 5 out of 5]

This was an amazing book in the history and philosophy of science (I especially enjoy the author’s conception of HPS performing a complementary role to specialist science). It is an incredibly illuminating text on the relatively shaky foundations of modern science, a significant blow to anyone with a strictly foundationalist conception of scientific knowledge. As a person attempting to study the history and philosophy of quantification of physical properties, it was immensely helpful, and I am infinitely indebted to Dr Chang. 10/10 recommend to anyone with even a remote interest in science.

[Kevin Orrman-Rossiter · Rating 4 out of 5]

An interesting and thorough book on the lively narrative of scientists attempts from the 17th to mid-19th century to solve the puzzle of temperature and how to measure it. This book is made valuable by a fascinating last chapter that poses how the history and philosophy of science could complement or be a continuation of science by 'other means'. This final chapter was captivating by itself.

[Sierra Finlinson · Rating 4 out of 5]

Has my basic understanding and acceptance of science been taken to the core? Yes. Am i now a wild skeptic for all things "scientific" measurement? Yes. Will I re-read it? Of course. Did I enjoy it? Absolutely not.

[Popup-ch · Rating 2 out of 5]

Chang writes well, but this book has an extremely narrow focus, with several chapters focusing on the different means of boiling water. (Only looking at the actual bubble formation, and how that is correlated with surface treatment and dissolved impurities.)

[Zachary Shifrel · Rating 5 out of 5]

The subject matter is interesting in itself but Chang’s practicing of phil sci taught me a lot about how to think about science in general

[Pat Nelson · Rating 5 out of 5]

So good, really original lines of thinking mixed with facinating history of measurement. I am still thinking about this a month after finishing it

[Conor Bennett · Rating 4 out of 5]

Lots of interesting insights into the history of thermodynamics, particularly of little-known things.

[Kevin · Rating 5 out of 5]

This book provides a detailed history of the measurement of temperature, a problem that few people bother to think about today. But what at first glance seems to be trivial turns out to be an incredibly difficult and complicated problem that forces us to consider a deep epistemological dilemma: how can we verify our measurements of an unobservable physical property in the absence of any previously accepted value? Chang argues that 19th century physicists' solution to this problem illustrates how science operates much more broadly. Rather than progressing through a process of systematic falsification, as a naive understanding of science might suggest, thermometry developed through an iterative process by which poorly justified ideas were built upon and corrected. At each point in this process, scientists relied on untested hypotheses and uncritical dismissals of alternative views. Yet while this history undermines the notion that science is a perfectly rigorous process, it is ultimately optimistic, suggesting that scientific progress does not rely on the impossible task of establishing indubitable truths but instead occurs through a continuous process of building coherent theories.

[Rachel Williams · Rating 5 out of 5]

Chang's book is a brilliant and endlessly fascinating example of how to do history and philosophy of science in a rigorous but interesting fashion. Chang's discussion of operationalism, coherentism, and epistemic iteration have had a huge impact on my recent thinking in regards to how to evaluate the prospects for current and future scientific approaches to consciousness.