Chapter Seven Experiment: How to Learn Things about Nature in the Seventeenth Centu~ I Reconfiguring experience Aristotle had asserted unequivocally that all knowledge has its origins in experience. He was echoed by scholastic Aristotelians, so that the aphorism "there is nothing in the mind which was not first in the senses" became a standard philosophical maxim in the later Middle Ages.! Despite this fact, many non-Aristotelian philosophers in the seventeenth century had taken to criticizing the approaches to learning about nature that were promulgated by scholastic learning for ignoring the lessons of the senses. Francis Bacon was but one among many in his stated view that Aristotle "did not properly consult experience ...; after making his decisions arbitrarily, he parades experience around, distorted to suit his opinions, a captive."2 Bacon's became a common view: Aristotelian philosophy was commonly represented during the century as being obsessed with logic and verbal subtleties, reluctant to grapple with things themselves as encountered through the senses. The rhetoric of the Baconian Royal Society came equally to incorporate such a picture of Aristotelianism, its spokesmen making frequent remarks dismissive of scholastic obsession with words instead of things. Galileo too, among many others, had attempted to dramatize what he saw as the emptiness of the official school philosophy. In Galileo's Dialogo of 1632, Simplicio (the Aristotelian character) at one point purports to explain why bodies fall by reference to their gravity. Salviati, who speaks for Galileo, replies by ridiculing the use of a word as an explanation. What is it that moves earthly things downwards? "The cause of this effect," says Simplicia, "is well known; everybody is aware that it is gravity." "You are wrong, Simplicio; what you ought to say is that everyone knows that it is called 'gravity.' What I am asking you for is not the name of the thing, but its l'ssence, of which essence yOel know not a bit more than you know about till' l:'IiSl'lwe of wh"levl'r mOVl'S the slars around.'" 132 Revolutionizing the Sciences Why was Aristotle's natural philosophy associated by its critics with a neglect of the lessons of experience and the favouring of empty words? The answers to this question will illuminate just what the new emphasis on experimental knowledge meant in the seventeenth century. As we saw in Chapter 1, section l, Aristotle's philosophy was centrally about understanding rather than discovery. Aristotle, while in practice very interested in empirical facts of all kinds (as found especially in his zoological writ. ings), wanted above all to solve the problem of how we are to understand ourselves and the world around us. Thus, in his more abstract philosophical writings, such as the Metaphysics, or in his logical writings, the specific lessons of the senses are largely sidelined in favour of analyses of how to argue, how to understand, and in what terms we must make sense of our experiences. In the Posterior Analytics especially, Aristotle attempts to show how an ideal science should be structured so that it would be able to account for empirical truths; the acquisition of those truths was not centrally at issue, and neither were any particular such truths themselves. Thus when Aristotle's followers considered what Aristotelian natural science should look like, the model that they examined was one in which empirically acquired truths were taken as given, with only their explanation being the truly important task. In a sense, therefore, an Aristotelian world was not one in which there were countless new things to be discovered; instead, it was one in which there were countless things, mostly already known, left to be explained.4 That Aristotle himself does not seem to have believed this is beside the point; it was nonetheless the lesson that his scholastic followers in medieval and early-modern Europe tended to draw from those of his writings that they found most interesting and most teachable. The typical expression of empirical fact for such an Aristotelian was one that summed up some aspect of how the world works. "Heavy bodies fall" is a typical example: it was a statement that acted as an unquestioned reference-point in a network of explanations that involved such things as the terrestrial elements and their natural motions, final causes, and the structure of the cosmos.s Such statements appeared in already generalized form, rather than in the form of singular experiences referring to historically specific events. One did not say "this heavy body fell when I dropped it"; one simply said that all heavy bodies always fall- that is how nature behaves. In the absence of the reported particular, no room was left for the denial or affirmation of a universal claim about how all heavy bodies behave. The assumption was that everyone, from everyday experience, already knows it to be true. The philosopher's job, according to Aristotle, was to show why it was true. This was a matter of giving appropriate causal explanations that would, in the ideal case, show why the fact to be explained was necessarily true given the attendant circumstances. Needless to say; ideal cases were seldom, if ever, met with. Understanding the sway, in early-modern Europe, of Aristotelian ways of formulating such questiolls involvt's seeing how l'Vt'!1 the most strongly Experiment: How to Learn Things about Nature in the Seventeenth Century 133 Figure 7.1 Cali/eo's use of the inclined plane to slow down the acceleration offree-fall, thus making it easier to measure. anti-scholastic of philosophers could still take those ways for granted, as foundational aspects of their thought. For example, the dominant scholastic-Aristotelian way of conceptualizing and handling experience forms the backdrop to Galileo's famous work on the fall of heavy bodies, finally published in the Discorsi of 1638 although reflecting work largely completed by 1609.6 Galileo tries at one point to establish the truth of his claimed experience that a falling body accelerates as it descends, its distance from the place of release increasing in direct proportion to the time elapsed. This experience takes the form of a standard Aristotelian generalization, describing how things behave in nature; Galileo does not describe a specific experiment or set of experiments carried out at a particular time, together with a detailed quantitative record of the outcomes. Instead, he simply says that, using apparatus of a kind carefully specified, he had found that the results of rolling balls down an incline and timing their passage yielded results that agreed exactly with his expectations, in trials repeated "a full hundred times." This last phrase (found frequently, in various forms, in contemporary scholastic writings) means, in effect, "countless times." Galileo wished to persuade his readers that the results amounted to common experience. His problem, however, was that the particular experience that he wished his readers to accept was not in fact one that is well known and familiar. The subsequent rise to dominance of reported experimental events as the foundations of scientific arguments would be attended by just these difficulties. When a natural phenomenon was well known, it could be adduced as part of natural philosophical reasoning with no difficulty, because no one would be likely to contest it. But if the phenomenon were not well known, and instead brought to light only through careful and unusual experimentation, how could the natural philosopher make it acceptable for use in creating philosophical explanations? Galileo wished to have his readers bl'lil'vc that things behaVl'd in nature just as hl! said they did. He could not rely 011 hjli rl'ndl'rll nll'l'ndy lx-In)/; d1lil10lil'd to nnx'pl lIw truth of 134 Revolutionizing the Sciences the foundational natural behaviours that he discussed (uniform acceleration in fa11), but at the same time he could not allow the matter to rest on nothing more than his say-so. Some people might have been prepared to accept his claims on the basis of his own personal and institutional authority, but that would not have made his arguments scientific. Galileo always adhered to a model of scientific demonstration that came straight from Aristotle: a true scientific explanation should be demonstrative, like the proofs of mathematics, and, like the mathematical theorems of Euclid, proceed on the basis of simple statements that all could accept as true at the outset. Euclid had employed starting points such as "when equals are subtracted from equals, the remainders are equal"; they were intended to be so intuitively obvious that no one could in good conscience deny them. When Aristotelian natural philosophers made arguments on the basis of empirical principles, such as "the sun rises in the east," or "heavy bodies fall," they too relied on the practical undeniability of such truths; everyone could be relied upon to accept them.7 Experimental results, however, lacked that kind of obviousness, which is why Galileo attempted, in the present case, to render them as routine as possible as quickly as possible. Claiming results that accrued from trials repeated "a full hundred times" was a way of saying "things always behave in this way," and hoping that the reader would believe it. Rene Descartes confronted similar problems. Like Galileo, Descartes finessed the problem of trust by refusing to acknowledge it as an issue. In the Discourse on the Method (1637), he invites other people to assist in his work by contributing "towards the expenses of the observations [experiences, which also means "experiments"] that he would need."s It was precisely the fecundity of his explanatory principles that required experiments, because, as Descartes himself said, for any given natural phenomenon he could usually imagine more than one possible explanation. Experiments were therefore required to determine which of them might be the true one. Descartes wanted to do all the actual work himself because, he says, receiving information about phenomena from other people would typically yield only prejudiced or confused accounts. He wanted to make the requisite experiences himself or else pay artisans to do them (since the incentive of financial gain would ensure that the artisans would do exactly what they were told). Descartes was intent only on convincing himself. He sidestepped the problem of trust by adopting a supreme selfishness: what convinced him should be good enough for anyone and everyone. II Mathematical experimentation These were issues that needed especial confrontation in the mathematical sciences. As various kinds of "physico-mathematics" sprang up in the course of the seventeenth century, the methodological impetus that had driven the emerglmc(' of tIll' catl'p;ory .~l'rV('d also to l'mphasb':l' diffil'ultit's Experiment: How to Learn Things abollt Nature in the Seventeenth Century 135 relating to experimental procedures.9 The mixed mathematical sciences had often, since their ancient inception, involved the use of specially made apparatus to investigate natural behaviours that were not obvious from everyday experience. Thus astronomy used specialized sighting instruments for measuring precise positions of bodies in the heavens (well before the appearance of the telescope, an additional instrumental resource, in the seventeenth century). Optics used special devices for measuring angles in reflection and refraction. Ptolemy had written important treatises, the Almagest and the Optics, in both sciences, and he detailed the apparatus that was required for the proper conduct of work in each. The eleventh-century Islamic philosopher known in Latin Europe as Alhazen had written the most important optical treatise used in Europe prior to Kepler's studies, and he too detailed the makeup and use of optical apparatus.lO As a result, the tradition of mathematical sciences practised by seventeenth-century Europeans involved them by its very nature in questions concerning the validation of artificially generated experience - experience that was not generally known. Consequently, the ideal of an Aristotelian science, wherein the phenomena to be explained were taken as established from the outset, did not in these cases apply. The issue became especially pressing by the beginning of the seventeenth century among people such as the Jesuit mathematicians, who wanted to show that the mathematical disciplines were genuine sciences according to Aristotelian criteria (like Galileo, they were concerned about their status as mathematicians vis-it-vis the natural philosophers). Experimental apparatus gave them trouble because of its unobviousness. Galileo's was a popular solution to this problem among mathematicians. Thus Jesuit mathematical scientists, such as the astronomer Giambattista Riccioli, reported experiments that involved dropping weights from the tops of church towers to determine their acceleration. While, unlike Galileo, Riccioli gave places, dates, and names of witnesses to underwrite his narratives, the way he used those narratives was always to turn them into authoritative assertions of how such matters always turn out. Another, especially famous, example of this presentational trick took place in 1648. The mathematician Blaise Pascal, perhaps best known for the famous "Pascal's Triangle," wrote from Paris to his brother-in-law, Florin Perier, in the Auvergne district of provincial France, requesting him to carry out an experiment. Pascal asked him to carry a mercury barometer up a nearby mountain, the Puy-de-Dome, in order to see whether the mercury's height in the glass tube would change as the trial was conducted at different altiludes. Pascal hoped and expected that it would, because he was convinced Ihat it was lh~' pressure of the air lhM sllslaini' the column of mercury in lhl' lubl', and lhtlt [Iir-pn'ssul'l' dWreai'l'S llll' hij.llwr one j.lO~'S.11 The appamtlls was ilsdf novl'l, hnvln~ bl'l'n dl'vlHl'd in till' 164llH in (llor~\I1l'I' by Jo:vill1~dlHlil 'lhrrln'lIt, who hnd hl'l'n n pml(>~(> of (:n1l11'lJ'" In llll' Inlll'r'H 136 Revolutionizing the Sciences Figure 7.2 Torricelli's experiment, in a variant by Blaise Pascal. The double arrangement is intended to demonstrate that the mercury is indeed supported by the pressure of the air. last years. Like Pascal, Torricelli ascribed the phenomenon to the weight, or pressure of the air (disputes also existed over which of the two, weight or pressure, was the correct way to speak of these matters). Pascal published a narrative account of the experiment not long afterwards, a report written by Perier with Pascal's introduction and commentary. Perier provides a detailed account of his ascent and descent of the mountain, in the company of named witnesses, and records the height of the mercury that was found each time the apparatus was set up at various stops along the way. At the end of the story, which indeed showed that the mercury stood lower in the tube the higher up the mountain it was measured, Pascal proceeds to turns Perier's narrative into the keystone of a universal philosophical truth. First of all, Pascal uses Perier's results to produce a quantitative correlation of change in height of mercury with change in altitude, already taking it for granted that what Perier had recorded held true of all such measurements. Pascal then predicts the smaller changes in mercury height to be expected if similar apparatus were to be lifted up from the ground to the much lower elevations provided by church towers found in Pnris - 11 more ev{~ryday Hdtinp; than that of P(,ril,r'H Experiment: How to Learn Things about Nature in the Seventeenth Century 137 elaborate exploit. Finally, having made specific numerical predictions of the changes to be expected, Pascal then asserts that actual trials confirm the predictions. Like Galileo with his inclined-plane experiments on falling bodies, Pascal gives no details or particularities of these ecclesiastical experiments; they just agree with expectations, as good natural regularities should. The two central difficulties raised up by experimental procedures, that of establishing trust in experimental narratives and that of establishing universality, or representativeness, for specific experimental outcomes, thus demanded answers with especial urgency in the mathematical sciences because these sciences often sought out unusual or unobvious phenomena. Opinions differed on what would happen to the height of mercury in the glass tube at increasing altitude, before Pascal's brother-in-law ascended the Puy-de-Dome in an attempt to answer the question - a question that did not already possess a generally accepted answer. The mathematical sciences (which subsumed the work of Pascal and others on mercury barometers) provided their practitioners with specialized knowledge that was hard to use as the basis for a demonstrative science because it was not rooted in universally accepted experience. Somehow, therefore, specialized knowledge had to be made into common knowledge. A frequent recourse for astronomers and other kinds of mathematicians was to rely on their individual reputations as reliable truth-tellers. In many cases (such as that of the Jesuit mathematicians), corporate reputations could also be drawn upon: professorships in universities and colleges, or, as in Galileo's case, association with powerful sources of patronage, could lend subtle weight to empirical claims: challenge the result and you were challenging the institution that implicitly certified it. Astronomers, however, had additional, more concrete ways of bolstering their claims. This is because, traditionally, astronomers did not as a rule publish their raw astronomical data. They did not present lists of observational results, such as measurements of planetary positions, which would then have required acceptance based solely on the astronomer's authority (unless, extraordinarily, similar measurements had been made by others at exactly the same times).12 Instead, astronomers used their raw data to generate predictive tables of planetary, solar, or lunar positions, using geometrical models designed to mimic apparent celestial motions. This work was presented in such a way as to efface any formal distinction between observational astronomy (writing down the numbers that were measured using observational instruments) and those parts of the enterprise centred on the calculation of predictive tables from geometrical models - models that were themselves initinlly justified by their correspondence to the data. This latter work wns tlw pilrt thnt might lw d('('ml'd Iluitnble for publication, but not till' fornll'r. Tlw Im·dkllVl' Inbh'll, mllll'r lI111n lilt' orifiinl1l rnw lIntn, Ill'rvl'lI n~ 1111' 1'1111/11' wnrrnnl 1'01' tl1I' Koodlll,tIti 01 1111' modl'IM from 138 Revolutionizing the Sciences which they were computed, since anyone could check at any time to see how accurate those predictions were. In the sixteenth century, after all, Nicolaus Copernicus's reputation as an astronomer rested on his mathematical abilities, not his presumed competence as an observer; astronomers were mathematicians. Later on in the century, Tycho Brahe, although famous as an indefatigable observer, did not publish his vast accumulation of observational results; instead, he published mathematical treatments, employing his observational data, of such things as the paths of comets, or of his new earth-centred astronomical system. Tycho hired Johannes Kepler to compute a more accurate model for the motion of Mars on the basis of his raw data, without at the same time allowing Kepler free access to his complete observational records. These records were so far from public that Kepler himself had great difficulty in gaining control of them from Tycho's widow following Tycho's death. "Experimentation" in the mathematical sciences, then, called on problems related both to trust and to the meaning of results relating to specific times and places. Astronomical practice already addressed such difficulties, as well as potential problems relating to the use of instrumentation in gathering data. In the latter case, instrumentation and apparatus, while usual for the mathematical sciences, were more problematic for areas of inquiry related to qualitative sciences. Francis Bacon's refusal to accept the legitimacy of a distinction between natural and artificial processes (as processes produced with artificial apparatus would be) thus plays an important role in the rhetoric, logic, and practice of experimental science in the seventeenth centuryY III "Baconian" experimentation As we sawin the previous chapter, Bacon's writings were used as an important resource for justifying experimental investigations, especially by the Royal Society of London. Bacon's own position on experiment as a scientific tool is, however, more ambiguous than it at first appears. Bacon, like Aristotle, stressed the importance of experience in learning the ways of nature. The examples that Bacon used to illustrate a proper use of deliberately contrived experience in making (his kind of) natural philosophical knowledge show exactly the same features of generality, or universality, that characterize the writings of scholastic philosophers. In Book II of the New Organon (1620), Bacon presents two worked examples of his new logic of investigation