x GLASS WORKS: NEWTON'S PRISMS AND THE USES OF EXPERIMENT SIMON SCHAFFER “Instruments are in truth reified theorems. (Gaston Bachelard, 1933) ‘Perhaps, said the Marchioness, Nature has reserved the Merit of denjonstrating Truth to the English prisms; that is, to those by whose means she at first discovered herself.” (Francesco Algarotti, 1737) Experimental controversy involves contest about authority. The accep~ tance of a matter of fact on the basis of an experimental report involves conceding authority to the reporter and to the instruments used in the ‘experiment. In the seventeenth century, experimental philosophers tused a wide range of means to make authority for their work, Convie- tion was thought to result from a long series of trials or from a single decisive experiment. It might result from being present as a witness at such a tial, achieving a replication of such a trial, or by reading a report given in so much circumstantial details that such direct witnes- sing was obviated. Authority might be held to lie in the credit of a single experimenter or in the communal assent of the experimental community. In their controversies, experimental philosophers often strenuously debated these differing ways of making conviction. Such fights show some of the uses of experiment in reaching agreement among disputants. Furthermore, they show how experimental instru- ments play a central role in these usages, and are resources which experimenters deploy in their struggles to achieve authority: 1. For examples of experimenters’ ambiguities about the significance of meny tals ot Froutguely decisive ones. and about the relative importance of director ‘victual” (wunesting, see Boyle (1604), sigs. A2-Ad, Slapin (1904), or68 simon Sclaffer obscures the detailed character of ‘experimental controversy. The ‘ground of such authority was often the matter in dispute. The resolution of Glass works 69. carcer of one of Newton's experiments, a trial with two prisms which he first recorded in a notebook in 1666 as the ‘forty-fourth’ of a long list of experiments on light and colour. In significantly changed format, this trial was made into an ‘experimentumt crucis'. Newton's experimentunt crucis was the object of considerable debate during the 1670s and has remained a central topic of philosophical and historical attention. The term was not used in his notebooks, drafts or lectures before 1672, nor id it appear in the Opticks in 1704. Nevertheless, the label remained current among Newton's readers and disciples. However, the reference of the term changed markedly. There was no consensus among the experimenters on the lesson which its author intended should be taught by this trial, nor on the proper method for conducting it. Some tried their version of this experiment, obtained results different from those which they held Newton had reported and rejected Newton's account of light and colour. These experimenters treated the trial as crucal but used that cruciality to undermine Newton's theory. Others, notably Robert Hooke, replicated Newton’ trial, but then argued that the trial vas not crucial, and denied that this replication licensed Newton's account. ‘Cruciality’ was an accomplishment which varied with out- comes of attemptid replication. ‘The character of this accomplishment was intimately connected with, issues of instrumentation, specifically, with the evaluations experimen- ters gave of the quality and arrangement of their prisms. There was no uncontroversial way of making these evaluations authoritative. For one community of experimenters during one period of time, Newton's experimentum crucis could be associated with an obvious procedure, involving complex arrangements of specially crafted prisms and lenses and a self-evident matter of fact, involving the chromatic homogeneity and the fixed refrangibility of primitive colour-making light rays. In the crucial experiment, a prism was used to make ‘primitive’ rays, and then one of these rays subjected to a second refraction in second prism, Newton sometimes claimed that if white light were transmitted through a prism it could be separated into a set of ‘primitive’ colour- making rays. A properly separated ‘primitive’ ray could not then be further divided by transmission through another prism. It was a stand- ing difficulty that many of Newton's critics reported that they could split putatively ‘primitive’ rays into further colours. But for Newton and his allies, a ‘primitive’ ray could be simply defined as a ray which could not be split by a second refraction. Then experimenters who managed to split such a ray could be criticised by Newton for their failure to produce ‘primitive’ rays. This argument established a troubling,0 ‘Sime Schafer circle, akin to what HM, Collins calls the “experimenter's regress. The eriterion of a good experiment was that it produced the maticr of faet which Newion sought to establish. Experimenters had to be convinced of this matter of fact before they could share this criterion. Once con- vietion had been achieved. then this criterion seemed unchallengeable After closure, the procedures for making ‘primitive’ says becarne self. evident. This paper documents the process by which this self-evidence was accomplished.’ The unarguable meaning which individuates an experiment is not achieved without struggle. How. then, do experiments acquire thle ‘entity? Scientific instruments play a decisive role in this process Newton's arguments suggested cist good prisms were those which ‘made ‘primitive’ rays. His eitics were told they were using bad prismns Instruments help make experiments compelling. because the selfevh. ence which is attached to instrumental procediires alter closure links complex experiments t0 agreed matters of fact. This closure makes instruments into what are seen as ancomtestable transiniers of mec, sages from nature, that is, it makes them ‘transparent This process ie comparable to what Trevor Pinch calls the black boxing’ of instruments: hheargues tha aftr closure ‘the social struggle overa piece of knowledge bas become embedded in a piece of apparatus’. Such pieces can then be treated as if they ‘regularly produce reliable and uncontentious information about the natural world’. Prisms have becomeso umeonten tious that itis now hard to recapture the sense of thelr contingent and controversial use. Yeti is that contingent and controversial use which must be recovered in order to understand how “transparency” is accomplished! must also be stressed that ‘tansparency’ is not necessaitly achieved permanently, The ‘wansparency’ of instruments may vary during con: Woversy. Protagonists in disputes may engage in the “deconstrurtion of provisionally achicved ‘transparency’. On occasion i may be nected {0 emphasise the specific complexities of an instrument it ort to defend observation reports against criticism. Newion sometimes argued that the fatings of prisms explained troubled experimental resale At ther times, Newton and his critics minimised the tole of thei instres iments in order to highlight what they clalmed were basie conceptual ‘lisagreements, Furthermore, the accomplishment of ‘transparency’ sue 3. For experimenters regress and recat, se Clin (1969). pp 79-100, 129-20. or dcwsion ofthe proces by which ‘pei sng ech Sone ea step ee Tato (987) 44 Binh (986), pp. 128. ss werk n tains a realist history of experimental argument. Outcomes of eatlier dlebates are then attributed to the ‘obvious virtues’ of instrumenie that his extics had chosen the wrong instruments, this allowed the Arter claim that Nature clearly spoke ofthe ruth of Newton's theory pt light and colour. But a more considered history of the prisms weed in these arguments shows how this daim was accomplished as part of ‘the provisional, local closure ofthe optical controversy? ‘TU PRISM BECOMES “THE USEFULLEST INSTRUMENT? (16 Newton's report of vials with lenses, mirrors and prisms, first deli- vered to the public alter 1670, were connected with the active inteneee of his contemporaries in telescope and microscope design, This Insteas ‘mental context willbe considered first, before analysing the means by which the: “ucality’ of Newton's trials was connected with the troubles of such optical devices, This connection with glass working gave New, {ons reports much of their immediate impact in London, When he fated lecturing on light and colour at Cambridge in January 1670, ‘Newton began by suggesting that attemps to grind conical glases te avoid spherical aberration were as futile as efforts “0 plough the ‘{aashote He then indicated that even ifsuch conics could be produced, there was a ‘property in ent in the nature of light” which prevented Ue PerTection of dlopirics’. This property was the spell retranglblity of primitive colour-making rays," In 1672, Newton repeated hens view’, {2 is letter to the Secretary of the Royal Society, Henry Oldenburg {2 carchlly crafted reminiscence, Newton claimed that In early Tee he had been working on the grinding of non-sphericalglassee- wing ihe understanding that ‘Light il slfisaHetgogeneous mixture offer ‘frangible vay prompted his abandonment of ‘my aforessid. Ghec, Pe geiensetction oF expernental suns in comtovesy se Pickering (96), esnion (196781), 3. pp. 438-9; Newton (1984), 1.49 Nero (199-7), 1B 95. Fr Wen’ contemporiry meron nonsphercl lenses see Bennet 1962), py 34-8 C7 aera SB: Newton 939-77, 1 pp. 3, 73-76 sand arvey (0929; Newt (1984). 1 pp. 427-428