An Instrument Can Make a Science: Jevons's Balancing Acts in Economics

23 May 1864
But we noticed also the letters RJ1845. There can be no doubt I think that they were carved upon the tree . . . by my excellent but unhappy brother¹ . . . I admired him & the things he made. I have some few of them yet, for instance the little set of grain weights, which he constructed for his chemical balance. The latter was ingeniously made of wood with a common knife edge; the movement & pans & weights were all complete, & he was able to make quantitative experiments with considerable exactness.

—Jevons, Papers and Correspondence I

In other branches of science, the invention of an instrument has usually marked, if it has not made, an epoch.

—Jevons, Principles of Science

Throughout his life, William Stanley Jevons showed a keen interest in scientific instruments. As witness, we may point to the eight entries he wrote for the Dictionary of Chemistry and the Allied Branches of Other Sciences (1863–68), most of them on measurement instruments: balance, barometer, hydrometer, hygrometer, thermometer, volumenometer. Jevons wrote these entries in the early 1860s, the period in which his major statistical and logical studies came to fruition and his outline for a mathematical approach to economic theory was read to Section F of the British Association for the Advancement of Science.² In his Principles of Science [1874] (1958), he even devoted a separate section to measurement instruments to point out “the general purpose of such instruments, and the methods adopted to carry out that purpose with great precision” (284).

In highlighting the importance of measurement instruments for scientific practice, Jevons differed from many of his contemporaries. Neither measurement nor instrument is, for example, an entry in John Stuart Mill's highly influential Logic (1843). This was no accident, for the Logic was a book about proof, not about discovery, as Mill's friend and biographer Alexander Bain (1882) noted. Mill had no experience with, nor knowledge of, concrete scientific practice.³ William Whewell's Philosophy of Discovery (1856) also had extremely little to say on the importance of measurement instruments, even though he included short autobiographical sketches of some of the great practical reformers of science—like Galileo, for whom instruments (such as the balance) were indispensable tools to disclose the laws governing the universe (Machamer 1998).

To depict Jevons's writings on philosophy of science as an “attempt to reconcile some of the disputes between Mill and Whewell on scientific methodology” (Schabas 1990, 54) therefore leads our attention away from the role Jevons attributed to measurement instruments in scientific discovery. Indeed, in the Principles ([1874] 1958), Jevons stressed the importance of scientific instruments for the formation of scientific disciplines, most notably for chemistry. Chemistry, he suggested, “has been created chiefly by the careful use of the balance” (272), a claim widely supported by historians of science (see, for example, Levere 1990; Wise 1993; Wise and Smith 1989a, 1989b; also Theodore Porter's article in this volume).

For Jevons, measurement instruments were an indispensable part of scientific practice: they were analytical tools of investigation as well as practical tools to give numerical precision to conclusions. Jevons's daily use of these instruments shows, even more than the entries in the Dictionary of Chemistry or the attention paid to them in the Principles, the importance of using measurement instruments. For Jevons, this was the way to access the world. There is virtually no page in his diary in which he did not record his measurements together with other descriptions.⁴ There was no walk in the country in which Jevons did not take his barometer with him to measure the height of the hills.

One instrument stood out in these accounts: the balance. Antoine-Laurent Lavoisier, Jevons's great predecessor in chemistry, had used the balance as the “most glorious weapon” in his battle against error,⁵ and...