Inventing Temperature

by Hasok Chang

  5. See Bolton 1900, 30-31; Mersenne described this thermometer in 1644.

  6. For a summary of the mechanical philosophers' views on heat and cold, see Pyle 1995, 558-565. For further detail, see Bacon [1620] 2000, book II, esp. aphorism XX, 132-135; on Boyle, see Sargent 1995, 203-204.

  end p.162

  Figure 4.2. Amontons's thermometer (1703, 53), with a double scale. Courtesy of the British Library.

  end p.163

  Gassendi's sort of theory seems to have enjoyed much popularity for some time, as reported in 1802 by Thomas Thomson (1773-1852), Scotland's premier "chemist breeder" and early historian of chemistry:7 There have been philosophers … who maintained that cold is produced not by the abstraction of caloric merely, but by the addition of a positive something, of a peculiar body endowed with specific qualities. This was maintained by [Petrus van] Muschenbroek [1692-1761] and [Jean Jacques d'Ortous] De Mairan [1678-1771], and seems to have been the general opinion of philosophers about the commencement of the eighteenth century. According to them, cold is a substance of a saline nature, very much resembling nitre, constantly floating in the air, and wafted about by the wind in very minute corpuscles, to which they gave the name of frigorific particles. (T. Thomson 1802, 1:339; emphasis added)

  Even by the late eighteenth century the question about the nature of cold had not been settled. The 2d edition of the Encyclopaedia Britannica in 1778 reported that there was no agreement on this question, but it came down on the side of supposing the independence of cold from heat, which led to the sort of discourse as the following: "if a body is heated, the cold ought to fly from it." This way of thinking persisted and even gathered strength by the third edition of Britannica, published at the end of the century. The author of the article on "heat" there admitted a good deal of uncertainty in current knowledge, and opined that the best way of proceeding was "to lay down certain principles established from the obvious phenomena of nature, and to reason from them fairly as far as we can." Ten such principles were offered, and the first one reads: "Heat and cold are found to expel one another. Hence we ought to conclude, that heat and cold are both positives."8

  Into this confused field came a striking experiment that seemed like a direct confirmation of the reality of cold and generated the controversy that became the last and most crucial debate in the banishing of cold from the ontology of the universe. The experiment was originally the work of the Genevan physicist and statesman Marc-Auguste Pictet (1752-1825). Pictet (1791, 86-111) set up two concave metallic mirrors facing each other and placed a sensitive thermometer at the focus of one of the mirrors; then he brought a hot (but not glowing) object to the focus of the other mirror and observed that the thermometer started rising immediately. (Fig. 4.3 shows a later version of this experiment, by John Tyndall.9 ) After a trial with the mirrors separated by a distance of 69 feet that showed no time lag in the production of the effect, Pictet concluded he was observing the radiation of heat at an extremely high speed (like the passage of light) and certainly not its

  7. For more information about Thomson, who is an important source for the history of chemistry up to his time, see Morrell 1972 and Klickstein 1948.

  8. "Cold," in Encyclopaedia Britannica, 2d ed., vol. 3 (1778), 2065-2067; the quoted statement is on p. 1066. "Heat," in Encyclopaedia Britannica, 3d ed. (1797), 8:350-353; the quoted principle is on p. 351.

  9. Tyndall (1880, 289-292) found it wondrous to have the opportunity to use this apparatus at the Royal Institution. He recalled acquiring the yearning to become a natural philosopher in his youth from the excitement of reading an account of experiments performed by Humphry Davy with the very same apparatus, which had initially been commissioned by Count Rumford.

  end p.164

  Figure 4.3. Illustration showing a version of Pictet's "double-reflection" experiment, from Tyndall 1880 (290). The spark ignited at the focus of the lower mirror causes the explosion of the hydrogen-chlorine balloon at the focus of the upper mirror. In an experiment more like Pictet's, a hot copper ball placed in the lower focus causes a blackened hydrogen-oxygen balloon at the upper focus to explode. Courtesy of the British Library.

  end p.165

  slow conduction through the air. To appreciate how remarkable this result was, we need to remember that its publication was a full decade before the discovery of infrared heating effect in sunlight by William Herschel (1800b). Unmediated heat transfer through space was a very novel concept at that time.

  This experiment was already stunning enough, but now Pictet described a variation on it that still inspires incredulity in those who read its description (1791, 116-118): I conversed on this subject with Mr. [Louis] Bertrand [1731-1812], a celebrated professor of mathematics in our academy, and pupil of the immortal Euler. He asked me if I believed cold susceptible of being reflected? I confidently replied no; that cold was only privation of heat, and that a negative could not be reflected. He requested me, however, to try the experiment, and he assisted me in it.

  When Pictet introduced a flask filled with snow into the focus of one mirror, the thermometer at the focus of the other mirror immediately dropped "several degrees," as if the snow emitted rays of cold that were reflected and focused at the thermometer. When he made the snow colder by pouring some nitrous acid on it, the cooling effect was enhanced. But how could that be, any more than a dark object could emit rays of darkness that make a light dimmer at the receiving end? Pictet was initially "amazed" by the outcome of his own experiment, which he felt was "notorious." Bertrand's suggestion seemed a frivolous one at the outset, but now it had to be addressed seriously.

  The situation here is reminiscent of the recent philosophical debates surrounding Ian Hacking's argument that we are entitled to believe in the reality of unobservable objects postulated in our theories if we can manipulate them successfully in the laboratory, for instance when we can micro-inject a fluid into a cell while monitoring the whole process with a microscope. Hacking (1983, 23) put forward a slogan that will be remembered for a long time: "[I]f you can spray them, then they're real." Hacking explains that this last insight came out of his own experience of overcoming his disbelief about the reality of the positron, the anti-particle of the electron. After learning how positrons can be "sprayed" onto a tiny niobium ball to change its electric charge (in a modern version of the Millikan oil-drop experiment), Hacking felt compelled to give up his previous notion that positrons were mere theoretical constructs. But what is Pictet's experiment, if not a successful spraying of cold onto the thermometer to lower its temperature, just as Hacking's physicists spray positrons onto a niobium ball to change its electric charge? Anyone wanting help in denying the reality of cold will have to look elsewhere, since the only answer based on Hacking's "experimental realism" will have to be that radiant cold is indeed real.

  Pictet himself dispensed with the conundrum relatively quickly, by convincing himself that he was only really observing heat being radiated away from the thermometer and sinking into the ice; the thermometer loses heat in this way, so naturally its temperature goes down. But it is clear that someone with precisely the opposite picture of reality could give a perfectly good mirror-image explanation: "Heat doesn't really exist (being a mere absence of cold), yet certain phenomena could fool us into thinking that it did. When we observe a warmer object apparently

  end p.166

  heating a colder one by radiation, all that is happening is that the colder object is radiating cold to the warmer one, itself getting less cold in the process." The Edinburgh chemist John Murray (1778?-1820) summed up the quandary arising from Pictet's and some related experiments as follows: In these experiments, then, we have apparently the emanation from a cold body of a positively frigorific power, which moves in right lines, is capable of being intercepted, reflected and condensed, and of producing, in its condensed state, its accumulated cooling power; and they appear equally conclusive in establishing the existence of radiant cold, as the other experiments are in establi
shing the existence of radiant heat. (Murray 1819, 1:359-360)

  A more generalized view was reported by the English polymath Thomas Young (1773-1829), best known now for his wave theory of light, in his Royal Institution lectures at the start of the nineteenth century: "Any considerable increase of heat gives us the idea of positive warmth or hotness, and its diminution excites the idea of positive cold. Both these ideas are simple, and each of them might be derived either from an increase or from a diminution of a positive quality."10

  Elsewhere I have examined what I consider to be the last decisive battle in this long-standing dispute, which was instigated by Count Rumford (1753-1814), who took up Pictet's experiment and developed further variations designed to show that the observations could not be explained in terms of the radiation of heat.11 Rumford is best known to historians and physicists for his advocacy of the idea that heat was a form of motion rather than a material substance, but in his own day he was celebrated even more as an inventor and a social reformer, who got rich by selling improved fireplaces and kitchens, invented the soup kitchen for the poor, reorganized the army and rounded up beggars into workhouses in Munich, and founded the Royal Institution in London.12 Rumford's scientific work was always geared toward practical applications, and his interest in radiant heat and cold was no exception. He noted that a highly reflective surface would serve to retard the cooling of objects (and people) in cold weather, since such a surface would be effective in reflecting away the "frigorific radiation" impinging on the object from its colder surroundings. Rumford tested this idea by experiments with metallic cylinders "clothed" in different types of surfaces, and late in life he enhanced his reputation as an eccentric by defying Parisian fashion with his winter dress, which was "entirely white, even his hat."13

  When the fundamental ontology of heat and cold was in such state of uncertainty, it was understandably difficult to reach a consensus about what it was

  10. Young 1807, 1:631. He continued: "[B]ut there are many reasons for supposing heat to be the positive quality, and cold the diminution or absence of that quality"; however, he did not care to state any of those "many reasons."

  11. See Chang 2002. Evans and Popp 1985 gives an informative account of the same episode from a different point of view.

  12. For brief further information, see Chang 2002, 141. The most detailed and authoritative modern account of his life and work is Brown 1979.

  13. About Rumford's winter dress, see "Memoirs" 1814, 397, and Brown 1979, 260.

  end p.167

  that thermometers measured in theoretical terms. But, as we know, the question of the existence of positive cold did get settled in the negative in the end. In the previously mentioned article, I argue that the metaphysical consensus against Rumford's frigorific radiation was reached thanks to the clear and increasing prevalence of the caloric theory of heat, into which it was difficult to fit any notion of positive cold. Therefore, the caloric theory at least produced an agreement that what thermometers measured was the degree or intensity of heat, not cold. But did it enable a complete and detailed theoretical understanding of temperature? That is the subject of the next section.

  Theoretical Temperature before Thermodynamics

  The first detailed theoretical understanding of temperature arrived in the late eighteenth century, in the tradition of the caloric theory, which was introduced briefly in "Caloric Theories against the Method of Mixtures" in chapter 2. As there were many different theories about the matter of heat, most commonly referred to as "caloric" after Lavoisier, I will speak of the "caloric theories" in the plural. Within any material theory of heat, it was very natural to think of temperature as the density of heat-matter. But as we shall see shortly, there were many different ways of cashing out that basic idea, and there were many difficulties, eventually seen as insurmountable, in understanding the operation of thermometers in such terms.

  Within the caloric tradition, the most attractive theoretical picture of temperature was given by the "Irvinists," namely followers of the Scottish physician William Irvine (1743-1787).14 The Irvinists defined temperature as the total amount of heat contained in a body, divided by the body's capacity for heat. This "absolute" temperature was counted from the "absolute zero," which was the point of complete absence of caloric. One of the most compelling aspects of Irvinist theory was the explanation of latent heat phenomena, which were seen as consequences of changes in bodies' capacities for caloric. For instance, the change of state from ice to water was accompanied by an increase in the capacity for heat, which meant that an input of heat was needed just to maintain the same temperature while the ice was melting. Irvine confirmed by direct measurement that the specific heat (which he identified with heat capacity) of water was indeed higher than the specific heat of ice, in the ratio of 1 to 0.9. The Irvinist explanation of latent heat in state changes was illustrated by the analogy of a bucket that suddenly widens; the liquid level in the bucket (representing temperature) would go down, and more liquid (representing heat) would have to be put in to keep the level where it was before (see fig. 4.4). The heat involved in chemical reactions was explained similarly, by pointing to observed or presumed differences between the heat capacities of the reactants and the products.

  Irvine's ideas were highly controversial and never commanded a broad consensus, but they were adopted and elaborated by some influential thinkers especially

  14. For further information about Irvine and his ideas, see Fox 1968, Fox 1971, 24ff., and Cardwell 1971, 55ff.

  end p.168

  Figure 4.4. An illustration of the Irvinist theory of latent heat, adapted from Dalton 1808, figure facing p. 217 .

  in Britain. Perhaps the most effective advocate of Irvine's ideas was the Irish physician and chemist Adair Crawford (1748-1795), who was active in London in later life but had attended Irvine's lectures while he was a student in Glasgow. Crawford applied Irvine's ideas in order to arrive at a new solution to the long-standing puzzle on how warm-blooded animals generated heat in their bodies (basically, the air we breathe in has a higher heat capacity than the air we breathe out). Crawford's influential treatise on animal heat (1779, 1788) served as an important conduit of Irvine's ideas to other scholars, including John Dalton (1766-1844), the originator of the chemical atomic theory. Other important Irvinists included Sir John Leslie (1766-1832), professor of mathematics and then of natural philosophy in Edinburgh, and the chemist John Murray (1778?-1820), writer of influential textbooks.

  The demise of Irvinism owed much to experimental failure. In fact what worked decisively against it was something that Karl Popper would have admired greatly: a high degree of testability. Irvine's hypothesis of heat capacity specified

  end p.169

  a precise quantitative relationship between heat capacity, temperature, and total heat, and the Irvinists were clearly agreed in how to render their theoretical notion of heat capacity measurable, by identifying it with the operational concept of specific heat. Therefore, measurements of specific heat, temperature, and latent heat in various chemical and physical processes yielded unequivocal tests of Irvinist theory. Most strikingly for thermometry, the Irvinists set out to locate the absolute zero by measuring how much latent heat was necessary to make up for known amounts of change in heat capacity while maintaining the temperature at the same level. For instance, in figure 4.4, the shaded area represents the latent heat involved in turning ice at 0°C into water at 0°C; the width of that shaded area is the difference in the specific heats of ice and water. The specific heat difference and the latent heat can be measured, and then dividing the latter by the former gives the height of the shaded rectangular area, which corresponds to the temperature difference between the melting point of ice and absolute zero. When Irvine performed this calculation, he obtained −900°F as the value of absolute zero (Fox 1971, 28). This was a great triumph for Irvinism: the "zeros" of all previous thermometric scales had been arbitrary points; with Irvinist theory one could calculate how fa
r the conventional zeros were from the absolute zero.

  However, most of the predictions made on the basis of Irvine's hypothesis were not confirmed. In the case of the measurement of absolute zero, the values obtained were so diverse that they rather indicated the nonexistence of absolute zero, in the opinion of the opponents of Irvinism. In their critique of Irvinism, Laplace and Lavoisier gave Irvinist calculations of absolute zero ranging from −600°C to −14,000°C. Even Dalton, who was doing the calculations in good Irvinist faith, admitted that the value he presented as the most probable, −6150°F, was derived from results that varied between −11,000°F and −4000°F (Cardwell 1971, 64, *127). Thus, Irvinist caloric theory stuck its neck out, and it was duly chopped off. Of course there were many other reasons why Irvinist theory fell out of favor, but the lack of convergence in the absolute-zero values was certainly one of its most important failures.

  It is a pity that Irvinism did not work out. After its demise, theoretical notions of heat and temperature have never recovered such beautiful simplicity. In addition, Irvinism was the only theory up to that point that provided a clear, detailed, and quantitative link between thermometric measurements and the theoretical concepts of temperature and heat. The caloric theorists who rejected Irvinism could only engage in defensive and untestable moves when it came to the understanding of temperature.