Inventing Temperature Read online

  As indicated by its inclusion in the Oxford Studies in the Philosophy of Science, this book is intended to be a work of philosophy. However, the studies presented here are works of philosophy, science, and history simultaneously. I am aware that they may cross some boundaries and offend the sensibilities of particular academic disciplines. And if I go into explanations of various elementary points well known to specialists, that is not a sign of condescension or ignorance, but only an allowance for the variety of intended readership. I fear that professional philosophy today is at risk of becoming an ailing academic discipline shunned by large numbers of students and seemingly out of touch with other human concerns. It should not

  end p.6

  be that way, and this book humbly offers one model of how philosophy might engage more productively with endeavors that are perceived to be more practically significant, such as empirical scientific research. I hope that this book will serve as a reminder that interesting and useful philosophical insights can emerge from a critical study of concrete scientific practices.

  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. More broadly, discussions of epistemology and scientific methodology included in this book will interest philosophers of science, and perhaps philosophers in general. Discussions of physics and chemistry in the eighteenth and nineteenth centuries will be of interest to historians of science. Much of the historical material in the first four chapters is not to be found in the secondary literature and is intended as an original contribution to the history of science. I also hope that the stories of how we came to believe what we believe, or how we discovered what we know, will interest many practicing scientists, science students, and non-professional lovers of science. But, in the end, professional labels are not so relevant to my main aspirations. If you can glimpse through my words any of the fascination that has forced me to write them, then this book is for you.

  end p.7

  1. Keeping the Fixed Points Fixed

  Narrative: What to Do When Water Refuses to Boil at the Boiling Point

  Abstract: This chapter draws attention to the challenges faced by the early scientists in establishing the fixed points of thermometry such as the boiling and freezing points of water. It provides a historical account of the difficulties in establishing the boiling point of water. It then discusses the meaning and achievement of fixity.

  Keywords: fixed points, boiling point, freezing point

  Hasok Chang

  The excess of the heat of water above the boiling point is influenced by a great variety of circumstances.

  Henry Cavendish, "Theory of Boiling," c. 1780

  The scientific study of heat started with the invention of the thermometer. That is a well-worn cliché, but it contains enough truth to serve as the starting point of our inquiry. And the construction of the thermometer had to start with the establishment of "fixed points." 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. This chapter is an attempt to become reacquainted with those old challenges, which are no less real for being forgotten. The narrative of the chapter gives a historical account of the surprising difficulties encountered and overcome in establishing one particular fixed point, the boiling point of water. The analysis in the second half of the chapter touches on broader philosophical and historical issues and provides in-depth discussions that would have interrupted the flow of the narrative.

  Blood, Butter, and Deep Cellars: The Necessity and Scarcity of Fixed Points

  Galileo and his contemporaries were already using thermometers around 1600. By the late seventeenth century, thermometers were very fashionable but still notoriously unstandardized. Witness the complaint made about the existing thermometers in 1693 by Edmond Halley (1656-1742), astronomer of the comet fame and secretary of the Royal Society of London at the time:

  end p.8

  I cannot learn that any of them … were ever made or adjusted, so as it might be concluded, what the Degrees or Divisions … did mean; neither were they ever otherwise graduated, but by Standards kept by each particular Workman, without any agreement or reference to one another. (Halley 1693, 655)

  Most fundamentally, there were no standard "fixed points," namely phenomena that could be used as thermometric benchmarks because they were known to take place always at the same temperature. Without credible fixed points it was impossible to create any meaningful temperature scale, and without shared fixed points used by all makers of thermometers there was little hope of making a standardized scale.

  Halley himself recommended using the boiling point of alcohol ("spirit of wine") as a fixed point, having seen how the alcohol in his thermometer always came up to the same level when it started to boil. But he was also quick to add a cautionary note: "Only it must be observed, that the Spirit of Wine used to this purpose be highly rectified or dephlegmed, for otherwise the differing Goodness of the Spirit will occasion it to boil sooner or later, and thereby pervert the designed Exactness" (1693, 654). As for the lower fixed point, he repudiated Robert Hooke's and Robert Boyle's practice of using the freezing points of water and aniseed oil, either of which he thought was "not so justly determinable, but with a considerable latitude." In general Halley thought that "the just beginning of the Scales of Heat and Cold should not be from such a Point as freezes any thing," but instead recommended using the temperature of deep places underground, such as "the Grottoes under the Observatory at Paris," which a "certain Experiment of the curious Mr. Mariotte" had shown to be constant in all seasons (656).1

  Halley's contribution clearly revealed a basic problem that was to plague thermometry for a long time to come: in order to ensure the stability and usefulness of thermometers, we must be quite certain that the presumed fixed points are actually fixed sharply, instead of having "a considerable latitude." There are two parts to this problem, one epistemic and the other material. The epistemic problem is to know how to judge whether a proposed fixed point is actually fixed: how can that judgment be made in the absence of an already-trusted thermometer? This problem will not feature prominently on the surface of the narrative about the history of fixed points to be given now; however, in the analysis part of this chapter, it will be discussed as a matter of priority (see especially "The Validation of Standards" section). Assuming that we know how to judge fixedness, we can face the material problem of finding or creating some actual points that are fixed.

  Throughout the seventeenth century and the early parts of the eighteenth century, there was a profusion of proposed fixed points, with no clear consensus as to which ones were the best. Table 1.1 gives a summary of some of the fixed

  end p.9

  Table 1.1. Summary of fixed points used by various scientists Person

  Year

  Fixed points ("and" indicates a two-point system)

  Source of information

  Sanctorius

  c. 1600

  candle flame and snow

  Bolton 1900, 22

  Accademia del Cimento

  c. 1640?

  most severe winter cold and greatest summer heat

  Boyer 1942, 176

  Otto Von Guericke

  c. 1660?

  first night frost

  Barnett 1956, 294

  Robert Hooke

  1663

  freezing distilled water

  Bolton 1900, 44-45; Birch [1756] 1968, 1:364-365

  Robert Boyle

  1665?

  congealing oil of aniseed or freezing distilled water

  Bolton 1900, 43

  Christiaan Huygens

  1665

  boiling water or freezing water

  Bolton 1900, 46; Barnett 1956, 293

  Honoré Fabri

  1669

  snow and hig
hest summer heat

  Barnett 1956, 295

  Francesco Eschinardi

  1680

  melting ice and boiling water

  Middleton 1966, 55

  Joachim Dalencé

  1688

  freezing water and melting butter or ice and deep cellars

  Bolton 1900, 51

  Edmond Halley

  1693

  deep caves and boiling spirit

  Halley 1693, 655-656

  Carlo Renaldini

  1694

  melting ice and boiling water

  Middleton 1966, 55

  Isaac Newton

  1701

  melting snow and blood heat

  Newton [1701] 1935, 125, 127

  Guillaume Amontons

  1702

  boiling water

  Bolton 1900, 61

  Ole Rømer

  1702

  ice/salt mixture and boiling water

  Boyer 1942, 176

  Philippe de la Hire

  1708

  freezing water and Paris Observatory cellars

  Middleton 1966, 56

  Daniel Gabriel Fahrenheit

  c. 1720

  ice/water/salt mixture and ice/water mixture and healthy body temperature

  Bolton 1900, 70

  John Fowler

  c. 1727

  freezing water and water hottest to be endured by a hand held still

  Bolton 1900, 79-80

  R. A. F. de Réaumur

  c. 1730

  freezing water

  Bolton 1900, 82

  Joseph-Nicolas De l'Isle

  1733

  boiling water

  Middleton 1966, 87-89

  Anders Celsius

  by 1741

  melting ice and boiling water

  Beckman 1998

  J. B. Micheli du Crest

  1741

  Paris Observatory cellars and boiling water

  Du Crest 1741, 8

  Encyclopaedia Britannica

  1771

  freezing water and congealing wax

  Encyclopaedia Britannica, 1st ed., 3:487

  points used by the most respectable scientists up to the late eighteenth century. One of the most amusing to our modern eyes is a temperature scale proposed by Joachim Dalencé (1640-1707?), which used the melting point of butter as its upper fixed point. But even that was an improvement over previous proposals like the "greatest summer heat" used in the thermometers of the Accademia del Cimento, a group of experimental philosophers in Florence led by Grand Duke Ferdinand II and his brother Leopold Medici. Even the great Isaac Newton (1642-1727) seems to have made an unwise choice in using what was often called "blood heat,"

  end p.10

  namely human body temperature, as a fixed point in his 1701 scale of temperatures.2

  By the middle of the eighteenth century, a consensus was emerging about using the boiling and freezing of water as the preferred fixed points of thermometry, thanks to the work of the Swedish astronomer Anders Celsius (1701-1744), among others.3 However, the consensus was neither complete nor unproblematic. In 1772 Jean-André De Luc (1727-1817), whose work I shall be examining in great detail shortly, published these words of caution: Today people believe that they are in secure possession of these [fixed] points, and pay little attention to the uncertainties that even the most famous men had regarding this matter, nor to the kind of anarchy that resulted from such uncertainties, from which we still have not emerged at all. (De Luc 1772, 1:331, §4274 )

  To appreciate the "anarchy" that De Luc was talking about, it may be sufficient to witness the following recommendation for the upper fixed point given as late as 1771, in the first edition of the Encyclopaedia Britannica: "water just hot enough to let wax, that swims upon it, begin to coagulate" (3:487).5 Or there is the more exotic case of Charles Piazzi Smith (1819-1900), astronomer royal for Scotland, who proposed as the upper fixed point the mean temperature of the King's Chamber at the center of the Great Pyramid of Giza.6

  The Vexatious Variations of the Boiling Point

  In 1776 the Royal Society of London appointed an illustrious seven-member committee to make definite recommendations about the fixed points of thermometers.7 The chair of this committee was Henry Cavendish (1731-1810), the reclusive

  2. See Newton [1701] 1935, 125, 127. Further discussion can be found in Bolton 1900, 58, and Middleton 1966, 57. Blood heat may actually not have been such a poor choice in relative terms, as I will discuss further in "The Validation of Standards" in the analysis part of this chapter. Middleton, rashly in my view, berates Newton's work in thermometry as "scarcely worthy of him." According to modern estimates, the temperatures of healthy human bodies vary by about 1 degree centigrade.

  3. On Celsius's contributions, see Beckman 1998. According to the consensus emerging in the late eighteenth century, both of these points were used together to define a scale. However, it should be noted that it is equally cogent to use only one fixed point, as emphasized in Boyer 1942. In the one-point method, temperature is measured by noting the volume of the thermometric fluid in relation to its volume at the one fixed point.

  4. In citing from this work, I will give both the paragraph number and the page number from the two-volume edition (quarto) that I am using, since there was also a four-volume edition (octavo) with different pagination.

  5. Newton ([1701] 1935, 125) had assigned the temperature of 20 and 2/11 degrees on his scale to this point. It was not till the 3d edition of 1797 that Britannica caught on to the dominant trend and noted: "The fixed points which are now universally chosen … are the boiling and freezing water points." See "Thermometer," Encyclopaedia Britannica, 3d ed., 18:492-500, on pp. 494-495.

  6. The information about Piazzi Smith is from the display in the Royal Scottish Museum, Edinburgh.

  7. This committee was appointed at the meeting of 12 December 1776 and consisted of Aubert, Cavendish, Heberden, Horsley, De Luc, Maskelyne, and Smeaton. See the Journal Book of the Royal Society, vol. 28 (1774-1777), 533-534, in the archives of the Royal Society of London.

  end p.11

  aristocrat and devoted scientist who was once described as "the wisest of the rich and the richest of the wise."8 The Royal Society committee did take it for granted that the two water points should be used, but addressed the widespread doubts that existed about their true fixity, particularly regarding the boiling point. The committee's published report started by noting that the existing thermometers, even those made by the "best artists," differed among themselves in their specifications of the boiling point. The differences easily amounted to 2-3 degrees Fahrenheit. Two causes of variation were clearly identified and successfully dealt with.9 First, the boiling temperature was by then widely known to vary with the atmospheric pressure,10 and the committee specified a standard pressure of 29.8 English inches (roughly 757 mm) of mercury, under which the boiling point should be taken. Drawing on De Luc's previous work, the committee also gave a formula for adjusting the boiling point according to pressure, in case it was not convenient to wait for the atmosphere to assume the standard pressure. The second major cause of variation was that the mercury in the stem of the thermometer was not necessarily at the same temperature as the mercury in the thermometer bulb. This was also dealt with in a straightforward manner, by means of a setup in which the entire mercury column was submerged in boiling water (or in steam coming off the boiling water). Thus, the Royal Society committee identified two main problems and solved both of them satisfactorily.

  However, the committee's report also mentioned other, much less tractable questions. One such question is represented emblematically in a thermometric scale from the 1750s that is preserved in the Science Museum in London. That scale (shown in fig. 1.1), by George Adams the Elder (?-1773), has two boiling points: at 204° Fahrenheit "water begins to boyle," and at 212°F "water boyles vehemently." In other words, Adams recognized a temperature interval as wide as 8°F in which various stages of
boiling took place. This was not an aberrant quirk of an incompetent craftsman. Adams was one of Britain's premier instrument-makers, the official "Mathematical Instrument Maker" to George III, starting from 1756 while the latter was the Prince of Wales.11 Cavendish himself had addressed the question of whether there was a temperature difference between "fast" and "slow" boiling ([1766] 1921, 351). The notion that there are different temperatures associated with different "degree of boiling" can be traced back to Newton ([1701] 1935, 125), who recorded that water began to boil at 33° of his scale and boiled vehemently at 34° to 34.5°, indicating a range of about 5-8°F. Similar observations were made by

  8. This description was by Jean-Baptiste Biot, quoted in Jungnickel and McCormmach 1999, 1. Cavendish was a grandson of William Cavendish, the Second Duke of Devonshire, and Rachel Russell; his mother was Anne de Grey, daughter of Henry de Grey, Duke of Kent. See Jungnickel and McCormmach 1999, 736-737, for the Cavendish and Grey family trees.