the matter of fact 2

7/31/2019 The Matter of Fact 2

1/68


2/68


3/68

The Matter of Fact 2.0


4/68


5/68

The second exhibition by the Historyof Science 126: The Matter of Fact class.

Fall 2009, History of Science DepartmentHarvard University


6/68

Copyright 2009 by listed authorsAll rights reserved.

Catalog designed by Grace Kim. Cover design by Kevin Davies and Grace KimCover art from iStockPhoto. Photographs on back cover by Evangelos Kotsioris

Photograph by Evangelos Kotsioris


7/68

TABLE OF CONTENTS

Introduction Jimena Canales

Instruments in ContextHow American Culture Shapes the Produc-tion of Scienti c KnowledgeKevin Davies

To See the Invisible in Late Nineteenth-Century PhysicsInstruments, Actors, FactsConnemara Doran

Fact: Blood is Red

Fact: MAX(blood) 645nmA history of systematizing and quantifyingcolor from the late 19 th century to the 1950sGregor Jotzu

Technology and Fact Jay Forrester and Magnetic Core MemoryHelen Keefe

Heads on DisplayInstruments to Convince the PublicGrace Kim

The Father and Mother of AngiogenesisWarring against the Scienti c Community

Joseph Pomianowski

1

3

10

20

31

37

48


8/68

1 The Matter of Fact 2.0

For this exhibit, we have decided to put facts in their proper place. Wehave locked them in three glass exhibitcolumns in the Department for theHistory of Science, to observe them,think about them and control their of-ten inexplicable hold on our society.

In the seventeenth century RobertBoyle and Thomas Hobbes debatedabout the existence of the vacuum alleg-edly revealed by the air-pump. Despitetheir disagreements, Boyle claimedthat both he and his contradictor hadone thing in common: a respect for matters of fact. Mr. Hobbes , claimedBoyle, does not deny the truth of anyof the matters of fact I have delivered.1 While Hobbes may not have disagreedabout the experimental facts at hand,his philosophy did, however, contestthe very category of fact that Boylesought to establish.2

Since 1662 facts have continued towage a battle against their makers, lim-iting the importance of interpretationand theory in their world. They havesuccessfully led us to forget that even

1 (Boyle 1772 (1662)) Cited in (Shapin 1994)2 (Shapin 1994; Shapin and Schaffer 1985)

the word fact comes from the Latin factum , a noun derived from facere which means to do or to make. In an-other ght they have perhaps been evenmore successful: facts have foughtagainst their own precarious material-ity. Yet even the hardest, the strongestof facts cannot survive in a vacuum.

A fact is a fact, explained the philosopher and mathematician HenriPoincar to point out that they werenot the be-all and end-all of science.3 The fabric of our world is not black with fact and white with conven-tion, claimed the philosopher W.V.O.Quine, but rather pale gray.4 Facts,claimed the sociologist of scienceHarry Collins, are like ships in a bot-tle, painstakingly constructed to seemas if no one could have made them.5 Facts, insisted the historian of sci-ence Bruno Latour, are like frozenvegetables, they need a bevy of support networks to survive and thrive.6 Facts, reminded us the historian of science Lorraine Daston, are nothing3 (Poincar 1902)4 (Quine 1966 (1956))5 (Collins 1985)6 (Latour 1987)

I ntroductIonPutting Facts In Their Proper Place

Jimena Canales


9/68

Introduction 2

like rocks, combating a century-long portrayal of them as hard, obstinateand even brutish.7

In our exhibit we no longer ask Dofacts exist? but instead examine howcertain facts come to be and strengthenwhile others wither and wane. It dis- plays a world that is no longer black with fact and white with convention, but that is also certainly not palegray either. It reveals how facts have become sacred. And, by bringing theminto a space of common use (and plac-ing them next to the scienti c instru -ments with which they are so closelyassociated), we explore how they canalso be profaned.8

7 (Daston 2005)8 For the profanation of technology see (Ag-amben 2009)

Agamben, Giorgio. 2009. What is an Apparatus?In What is an Apparatus? : and Other Essays.Stanford, Calif.: Stanford University Press.

Boyle, Robert. 1772 (1662). An Examen of Mr. T.Hobbes His Dialogus Physicus de Natura Aris.In The Works of the Honourable Robert Boyle,

edited by T. Birch. London: J. & F. Rivington.

Bibliography

Collins, H. M. 1985. Changing Order:Replication and Induction in Scienti c Practice.Chicago: University of Chicago Press.

Daston, Lorraine. 2005. Hard Facts. In MakingThings Public: Atmospheres of Democracy,edited by B. Latour and P. Weibel. Cambridge,Mass.: MIT Press.

Latour, Bruno. 1987. Science in Action: How toFollow Scientists and Engineers through Society.Cambridge, Mass.: Harvard University Press.

Poincar, Henri. 1902. La science et lhypothse,Bibliothque de philosophie scienti que. Paris:E. Flammarion.

Quine, W.V.O. 1966 (1956). Carnap and LogicalTruth. In Ways of Paradox and Other Essays. New York: Random House.

Shapin, Steven. 1994. A Social History of Truth:Civility and Science in Seventeenth-CenturyEngland. Edited by D. L. Hull, Science and ItsConceptual Foundations. Chicago: University of Chicago Press.

Shapin, Steven, and Simon Schaffer. 1985.Leviathan and the Air-Pump: Hobbes, Boyleand the Experimental Life. Princeton: Princeton

University Press.

Photograph by Joseph Pomianowski


10/68

I nstruments In c ontext

How American Culture Shapes the Production ofScienti c Knowledge

A culture is madeor destroyedby its articulate voices.

Ayn Rand

Kevin Davies

Scienti c knowledgethose factsthat scientists have produced/de-duced by crafting theory, conductingexperiments, and creating instrumentsto aid studyhas long interested so -cial scientists and scholars of the hu-manities. Challenged to explain howscientists come to consensuses abouttheories and how such theories come to be accepted as fact by the general pub-lic, academics have articulated a number of theories about the complicatedrelationship between the producers of scienti c knowledge and the publicthat consumes such information.1 Thesciences of the human mind and humanintelligence are a particularly interest-ing kind of knowledge to study, becausethe clear divisions between scientistsand nonscientists that exist in some

elds are not so readily apparent here.Whereas to study (or simply create)X-rays one needs certain kinds of tech-

nologies and other material resources,and whereas to study microphysics onemay need a rigorous understanding of

1 (Kuhn 1996)

mechanics or mathematics, the studyof human intelligence does not imme-diately appear to be characterized bysuch gaps.

In this paper, I seek to build uponthe scholarship that analyzes the rela-tionship between scientists and nonsci-entists by looking at some of the waysin which knowledge about the humanmind and human intelligence was pro-duced in turn of the 20th century Amer-ica. Human intelligence, the body,and the mind have been at the center of philosophical and scienti c inqui -ries long before the year 1900; whatmakes this particular period interest-ing to study, however, is the emergenceof new technologies used to study themind and intelligence and the relation-ship those technologies have with the prevailing culture of the day. The anal-ysis of three such instrumentsa mod -el of a human head used a psychology

lab, an anthropometric device used inanthropometric studies, and an IQ testfrom the middle of the 20th century will be a key component of this paper.



11/68

Though these instruments were used indifferent times and by scientists work-ing in different elds, I will argue thatthey are connected by the fundamentalassumptions about the mind/body thatundergird their use. Ultimately, I willsuggest that historians of science woulddo well to critically interrogate the re-lationship between these basic, nearlyinvisible assumptions about the worldand the production of scienti c knowl -edge if they are to gain a fuller pictureof how scienti c knowledge spreads.

Background: Why study thestudy of the mind?

Though the three instruments at theheart of this project were certainly notavailable to everyday citizens, intelli-gence and the human mind itself as ob-

jects of study and public inquiry were.As such the boundary we presume toexist between scientists and nonscien-tists was blurred, and both were able toengage in debates about the nature of intelligence and the mind that ultimate-ly shaped and informed the productionof scienti c knowledge.

If everyday people are involved inthe production of knowledge about in-telligence and the mind, then histori-ans of science might begin to questionwhether and how the everyday assump-tions that such people make about thenature of people and the nature of in-telligence in uence the creation of sci -enti c instruments and the knowledgesuch instruments ultimately produce.To bolster this analysis, I will discussthe overall context in which these in-struments were used, paying particular

attention to the invisible assumptionsthat enable us to use each instrument before discussing the production of sci-enti c knowledge more generally.

Contextualizing the Instruments

The three instruments at the heart of this project are: an anthropometric kitincluding calipers and an elaborate in-strument designed to measure various body parts; models of a human heads based on corpses with half of the brainexposed, used to study the relation-ship between the brain and illnesses byidentifying differences across an arrayof models; and an IQ test that, thoughnot from the era in question, shares thesame basic structure as tests from thatera. These instruments are diverse, andat rst it might be tempting to discussthem separately. Each was used by dif-ferent scientists, working in differentdisciplines, and operating in differentregions and times; thus the linkages thatwe imagine between scientists workingon similar problems and experimentsmight seem to be missing at rst. Yet

there are number of compelling rea-sons to examine the instruments jointly.Though the scienti c community didnot necessarily connect the instrumentseither theoreticallythat is, in terms of what each instrument allowed scien-tists to see about a common subject or experimentallythat is, throughactual usethe instruments certainlycan be connected theoretically if oneis attentive to how they each producea form of knowledge about the body.Furthermore, the instruments must be

Instruments in Context 4


12/68

applied to actual human beings in order for knowledge to be produced; as such,we must account for how cultural no-tions of the body that dominated larger American culture in uenced the way the

people studied using these instrumentswere perceived. When that knowledgeis valuable not merely in terms of its ab-stract scienti c utility, but also in termsof how it changes the compelling culturaldialogues of the day about race and hu-man identity, the question of popular cul-ture becomes even more important.

One such cultural dialogue centeredon Darwinism, by which I mean the body of knowledge inspired the labor byCharles Darwin and his contemporariesthat described history in terms of the evo-lution of animals driven by the statisticalselection of those better able to survive.Historian of science Janet Browne has

highlighted the extent to which Darwin-ism facilitated the spread of evolutionaryideas about human society and relations.2 Browne speaks extensively about thevarious scientists who, using evolution-ary theory, took up arguments on bothsides of many popular debates of the day, particularly about relationship betweenraces. The intersection of evolutionaryideas with the political/social movementsin turn of the century America did notnecessarily lead the emergence of anyone particular argument; instead, Brownesuggests that people on each side of thedebate came to incorporate evolutionarythinking into their arguments. Brownediscusses this synthesis as the spread of scienti c knowledge to ordinary Ameri -cans. It would be equally valid, however,to discuss Darwinism not as a precursor 2 (Browne 2006)

to a phenomenon like imperialism, butinstead, like such phenomena, as a re-

ection of the tension between races andsocioeconomic groups that came to in u -ence the knowledge about race and classthat scientists produced.

The study of the human body is pro-foundly tied to American popular culture because it can entail an analysis of twohuge components of that culture: the na-ture of the human body and the nature of social relations between people. Decadesafter the American Civil War ended, the

question of where African-Americans(and, for that matter, immigrants moregenerally) t within American societyraised questions about how Americansof different races related to one anoth-er.3 These questions were important tothe American political apparatus, which played a large role in mediating the rela-

tionship between white and black Ameri-cans and in controlling the American population at large by restricting the owof immigrants into the country to a fewnarrow channels from a few speci c re -gions, in addition to managing severalcolonies. Darwinism provided a newframework for these kinds of culturalconcerns to emerge in the practice of science, and its in uence is clear uponanalysis of how the instruments centralto this project were used.

Producing Scientifc Knowledge/Reproducing

Popular Knowledge

The model of the human head hasno clear function. Made from delicate

3 (Richardson 2000)



13/68

materials and imported from Europe,the head was used in Hugo Mnster- bergs Harvard lab; photographic

evidence reveals that the heads weredisplayed together (while few remainin the collection, tens existed in years past) and that when they were used inexhibits they traveled together as well(Figure 1). What I infer from this isthat the heads were used to study hu-manity comparatively, particularly therelationship between the brain andwhat today we might call mental ill-ness. The heads vary in terms of howthey are constructed: while they all ex- pose the brain and the expressions onthe faces of each model re ect pain,stress and anguish (evidence of their use in the study of mental illness),they vary in terms of how the brainis depicted physically. These differ-ences are, seemingly, the point (as theyemerged from differences in the peopleupon whom the models are based):

the heads demonstrate the connection between differences in the brain anddifferences in the symptoms that was

fundamental to the study of these pa-tients. Such a use of these instrumentsrequires certain a priori assumptionsabout the ways in which we study themind: in a particular, they teach us tosee traits of the mind and the body as insome way re ective of and reducible to physical differences in the brain matter of individuals.

If this way of looking at brainsand their relationship does not seemradical, consider what a departure it isfrom conceptions of human differencethat center upon relative moral differ-ences, or that describe human differ-ence in terms of a soul rather than justa mind and body. This profoundly bio-logical, brain-oriented conception of identity relates to that which the use of the anthropometric device reproduces.While popular culture has reduced de-


Figure 1. Models of heads in a laboratory. Courtesy of the Harvard University Archives, call# HUPSF Psychological Laboratories (BP 1).


14/68

pictions of anthropometry to the studyof the size of the head or of bumps onthe head, the device itself highlights tothe point articulated by Steven Gouldin his text The Mismeasure of Man ,which is that anthropometric deviceswere used to link all sorts of measure-ments of body parts and the relation-ship between body parts (e.g. the ratioof ones forearm to ones bicep) andcertain human traits such as criminali-ty, aggression, and intelligence (Figure2). Gould highlights that in the produc-

tion of information on these data, sci-entists personal views regularly biasedthe observations they made; in particu-lar he highlights how one scientist sawhuge disparities between black andwhite subjects when he knew the raceof the person he was studying but notwhen such information was kept from

him.4

Yet I argue that the relationshipof the scientists personal views andthe data they created goes even further than that, for the very categories of traits they were investigating are traitsthat are de ned culturally: there is nosingle conception of crime, nor of whatconstitutes aggression nor intelligence,and thus to study these traits scienti -cally is both to embed a particular con-ception of these traits into the practiceof science and to ensure that this con-ception spreads along with the knowl-edge that it helped to produce. Theanthropometric devices, by requiringthe rei cation of certain traits into ab -stract qualities that could be measuredon the body, allowed scientists to link the biology of the entire body to iden-tity; in the context of Darwinism and4 (Gould 1981, Davies 1955)

given the particular political climate of the day, this meant that scientists werecomplicit in the production of knowl-edge that suggested that non-whiteswere less intelligent, valuable, and tthan their European-American counter- parts, and that these differences were

not only measurable, but measurablespeci cally by studying the physicalmaterial of the human body. Whether that assumption of difference of thescienti c evidence of difference came

rst is to some extent a question of thechicken vs. the egg: both phenomenonemerged in concert with one another,and together they characterize a part of the popular conception of humanity inAmerica at the time.

Only a decade or so after these in-struments were used, the IQ test trav-eled to America, and there it was usedto buttress certain arguments madeabout human evolution and the t -ness of some people relative to othersthat are also important in the descrip-tion of this eras paradigm of thought.5

* All subsequent HCHSI images and informationabout instruments in this catalog are from the

Figure 2. A set of three anthropometric in-struments. Courtesy of the Harvard Collec-tion of Historical Scienti c Instruments. *



15/68

Though the IQ test, created in France inorder to study and correct for differenc-es outcomes of the educational system,was not meant to identify or measurean individuals inherent intelligence,as the instrument traveled to Americathat was precisely the use to which itwas put, very often in studies of immi-gration or of differences in intelligenceacross races.6 In fact, the modern ideathat these differences exist is still jus-ti ed by many with reference to IQ. 7 Here we see another norm that exists in

relation to the two previously de ned:the idea that human intelligence is anabstract trait, measurable not even interms of the body itself, but in termsof a test created and administered by ascientist.

Conclusions: The Intersection

of Culture and Science

The reproduction of these normsin the form of scienti c instrumentsand in the practice of scientists is thecentral phenomenon this paper seeksto highlight; in the conclusion, I hopeto explore the implications of such

reproduction has for our conceptionof scienti c knowledge and how suchknowledge worksthat is, how it trav -els from person to person, whether itchanges form as it travels, who is em- powered to create such knowledge.

Several scholars have provided use-ful conceptualizations of the spread of scienti c knowledge. Social scientist

HCHSI database, Waywiser.5 (Richardson 2000)6 (Gould 1981)7 (Herrnstein and Murray 1996)

Bruno Latour described the spread of science not in terms of revolutions or paradigms, but instead by using thetrope of a network.8 Latour used theterm metrology to refer to the exten-sion of these networks through toolsof calculation and measurement this gigantic enterprise to make theoutsidethat is, the world outside anetworka world inside which factsand machines can survive. While nota conception of the spread of scienti cknowledge per se, historian of science

Hasok Changs ontological principleshighlight an important aspect of scien-ti c knowledgethose features thatare commonly regarded as essentialfeature of reality within an epistemiccommunityand, like Latour, seeksto make explicit the shared conceptionof the world that is a key component of

the social connections between scien-ti c thinkers 9.Each of these conceptions captures

aspects of the story told aboveChangmight discuss the norms identi edabove as ontological principles thatundergird the discourse on intelligenceand the mind, and Latour might see, for instance, the use of IQ tests to Americaas the extension of a network to includea new domain. Yet what each leavesout to an extent is the degree to whichnonscientists can change the natureof a network and broad cultural ideassalient in the places where science ismade can be as fundamental a part of the production of scienti c knowledge

as any other ontological principle, evenif these norms come from nonscientists

8 (Latour 1988)9 (Chang 2004)



16/68

and are nearly invisible until criticallyanalyzed. Critical attention to the roleof nonscientists and sustained analysisof the culture in which science madeare sometimes necessary in order todescribe the dissemination of scienti cknowledge, and thus provide two are-nas of investigation which might yield promising returns in future studies.

BibliographyBrowne, Janet. Darwins Origin of the Species. New York: Grove Press, 2006.

Chang, Hasok. Inventing Temperature. Oxford:Oxford University Press, 2004.

Davies, John. Phrenology: Fad and Science. New Haven: Yale University Press, 1955.

Gould, Stephen. The Mismeasure of Man. NewYork: Norton, 1981.

Herrnstein, Richard and Charles Murray. TheBell Curve. New York: Free Press, 1996.

Kuhn, Thomas. The Structure of Scienti cRevolutions. Chicago: University of ChicagoPress, 1996[1962].

Latour, Bruno. Science in Action. Cambridge:Harvard University Press, 1988.

Richardson, Ken. The Making of Intelligence. New York: Columbia University Press, 2000.


Photograph by Grace Kim


17/68

t o s ee the I nvIsIble In l ate n Ineteenth -

c entury P hysIcsInstruments, Actors, Facts

It is a very small fact, but it is a fact giving a great return.

Henri Poincar, Science and Method (1908)

Connemara Doran

How does one measure the great-ness of the return given by ascienti c fact the usefulness or depthof insight it gives us into the workingsof nature? In his popular scienti c es -says written in the rst decade of the

twentieth century, the great Frenchmathematician, physicist, and cosmol-ogist Henri Poincar challenged hisreaders to consider the dilemma facingthe scientist because of the hierarchy of facts. At the bottom of the hierarchy,undoubtedly, are facts that tell us noth-ing beyond themselves, but less clear is

how the scientist can know which factsactually contain hidden clues. For it isonly by choosing which facts to pursuefurther that the scientist will be led,step by step, to a fact at the top of thehierarchy, a fact that teaches us a newlaw. In this sense, the scientist mustconspire with nature in determining thereturn of a scienti c fact.

Amidst the huge array of ndingsPoincar could have offered his read-ers to assess, Poincar chose a ndingso ordinary that one would not expect

greater signi cance. When the geod -esist [seeking to determine the size andshape of the earth] nds that he has toturn his glass a few seconds of arc inorder to point it upon a signal that hehas erected with much dif culty, this

very small fact reveals much more thanthe existence of a little hump upon theterrestrial geoid.1 This little fact abouta few seconds of arc on his instrumentgives the scientist a great return becausethe hump contains secrets of immensescope in time and place, which areclues to the laws of the universe be -

cause this hump gives him indicationsas to the distribution of matter in theinterior of the globe, and, through that,as to the past of our planet, its future,and the laws of its development.2 Thelittle fact established by the geodesist provides crucial data with which thegeologist can see the deep sunk rocksthat affect the measure of weight at dif-ferent points on the surface of the earth

geodesy weighing them at a distance1 (Poincar 1908, 572)2 (Poincar 1908, 572)

To See the Invisible in Late Nineteenth-Century Physics 10


18/68

so to speak.3 From little fact to bigger fact, the return of a fact to science canextend far beyond itself.

In a sense, all science is the effortto see the invisible. But during thelate nineteenth century, this was literal-ly what much of science entailed. Herewe will consider two entirely differenttypes of questions that challenged sci-entists confronting the invisible, andthe ways in which their instrumentsgave them a means to conceptualize,explore, experiment, and interpret.

Conceiving the Inconceivably Small: J. J. Thomson and the

Electron-Corpuscle

We should examine with the utmost care the roleof hypothesis; we shall then recognize not onlythat it is necessary, but that in most cases it islegitimate. We shall also see that there are sev-eral kinds of hypotheses; that some are veri able,and when once con rmed by experiment becometruths of great fertility; that others may be usefulto us in xing our ideas.

Henri Poincar, Scienceand Hypothesis (1902), 4.

How can an experiment hope todetect the existence of something so

small that the very concept strains our abilities to conceptualize? How canwe even conceive the inconceivablysmall? Yet so amazing was the streamof surprises coming out of the laborato-ries during the late nineteenth century radio waves transmitting sound acrossmiles of space, lightwaves created in

evacuated tubes that scientists wereneither shy of speculating about whatother secrets nature might be hiding

3 (Poincar 1908, 559)

from us, nor of daring to nd them anduse them.

In tracing how scientists conceivedthe inconceivably small, we will un-cover the interplay of evidence andworking hypotheses in the creationof a scienti c fact at the end of thenineteenth century. We experience thedrama that occurs every day among theinstrument, the scientist, and the elu-sive fact in the scientists search for ev-idence. For the kind of evidence thatwe, as scientists, are able to amass withany instrument depends in large part onthe conceptual categories by meansof which we structure our thought, pattern our arguments and proofs, and

certify our standards for explanation.4

In striking contrast to the insis-tence on absolute precision which wasdominant in the search for standards of measure, nineteenth-century scientistsalso used a variety of speculative con-ceptual categories to guide their study both of radiant heat and of electric dis-

charge (light produced by electric cur-rent) in air, in puri ed gases, and in theCrookes tube (a vacuum tube). Longexperience with the apparatus and the4 (Daston 1991, 282)

Figure 1. Photograph of J.J. Thomson dem-onstrating his m/e experiment in a lecture (c.1900). (Falconer 1997, 226).



19/68

phenomena under investigation en-abled the British physicist J.J. Thomson(Figure 1) to obtain a great return fromapproximate measures of the ratio of mass-to-charge of cathode ray par-ticles (beams of electrons produc-ing heat and light in a vacuum tube).Precision is not always a necessity inthe search for scienti c understanding.

The thermopile (Figure 2) proveda very cooperative instrument inThomsons experiments. Used to mea-sure the difference of temperature

of radiating surfaces, the thermopile(Figure 3) was characterized as prob-ably the most celebrated instrumentever designed for the study of radiantheat, for . . . it is to the services of thethermopile that we owe the researchesof Melloni and Tyndall [for all types of gases], as well as nearly all the advanc-

es that have since been made in the

study of radiation.5 What made thethermopile so useful in the study of rar-e ed gases and infrared radiation, andin so many experimental contexts, wasa simple physical principle: two wiresof different metals are joined end toend so as to form a closed circuit, [and]then when one of the junctions is heat-ed, or cooled, an electric current passesround the circuit due to the differencein temperature between the two ends.6

In using this instrument, both theBritish physicist John Tyndall (c.1850-1885) and J.J. Thomson (from c.1880)went far beyond seeking increased pre-

cision. Each structured his thinking,and planned his entire experimentalresearch program, based on stimulat-ing possibilities suggested by roughresults ( ndings which, although ap - proximate, are taken to be factual) andconceptual working hypotheses (a particular model or mechanism that

could account for such ndings).Yet these two scientists diverged inthe use to which they put their results.5 (Preston 1894, 491)6 (Preston 1894, 491)

Figure 2. Double Thermopile. Courtesy ofthe Harvard Collection of Historical Scien-ti c Instruments.

Figure 3. Tyndalls experimental setup us-ing the thermopile. (Tyndall 1872, Frontis-piece).



20/68

Tyndalls early insights about the mo-lecular theory of gases ultimately becamea platform for advancing the mechanicaltheory of matter for explaining every -thing in terms of molecules of matter inmotion. Opposed to such an approach

which made the hypothesis a meta - physical assumption Thomson insteadused his rough results to lead to a more

nuanced working hypothesis to guide himtoward a more complete and fully consis-tent theory. Thomson made a rough esti-mate of the ratio of mass-to-charge of thecathode ray particle to be of the order of 1/1000 the ratio for the hydrogen ion,and from this identi ed it as a minisculecomponent of the atom indeed a cor-

puscle, a fundamental constituent of allmatter.What rough results served as evidence

for each scientist, and what working hy-

potheses guided their research?Tyndall developed a thermopile tech-

nology which provided measures (lessthan perfect but) more than suf cient toestablish the radiant-heat absorption of various gases, to show the greenhouseeffect created by the earths atmosphere,to establish the identity of radiant heatand visible light, and to demonstrate thescattering of light by tiny particles in gas-es.7 But Tyndall also drew upon JamesClerk Maxwells kinetic theory of gasesas evidence for the hypothesis that heat isa mode of motion in accordance withthe mechanistic goal of explaining all phenomena in terms of matter in motion.Maxwells kinetic theory of gases was based on the statistical theory of molecu-lar velocities he had introduced in 1859,the rst use of statistical argument in physics, and a mode of analysis that grewin importance as new technologies madeincreasingly possible experimental accessto the invisible and the miniscule.

The Crookes tube (Figure 4) greatlyexpanded the use of thermopile technol-ogy. In an 1879 public demonstration published in Philosophical Transactions ,the British physicist William Crookesexplained how the cathode ray tube(Figure 5) generates heat when im- pacting a metallic target.8 Crookesdescribed how molecular bombard-ment of particles within the glass tube,and the molecular impacts at the endof the tube produced so much heat thatwhen he touched it with his nger, itimmediately raised a blister, and thespot where the focus fell was nearly7 (Tyndall 1982)8 (Crookes 1979, 135)

Figure 4. Crookes Tube to demonstrateheating effect. Courtesy of the Harvard Col-lection of Historical Scienti c Instruments.



21/68

red hot.9 But how might the scientistexplain the unusual phenomena cre-ated in the Crookes tube? Crookes presents his theoretical speculationsduring the experiments as workinghypotheses, useful, perhaps necessary,in the rst dawn of new knowledge, butonly to be retained as long as they areof assistance.10 Crookes working hypothesis was of an ultra-gaseous stateof matter in which the propertieswhich constitute gaseity are reducedto a minimum, and the matter becomesexalted to an ultra-gaseous or molecular state, in which the very decided but hith-erto masked properties now under inves-tigation come into play.11 Gaseous dis-charge the light emitted from gas atomssubjected to an electric current was atthe core of J.J. Thomsons research andthe stimulus for his own working hypoth-esis.

Thomson began his work on gaseousdischarge immediately upon being electedin 1884, at age 28, as Cavendish Professor of Experimental Physics at Cambridge.His goal was to explain the relation be-tween chemical atoms and the ether

the presumed medium through whichelectromagnetic waves traveled. By1886, he had established in experimentsutilizing X-rays, that electric conductionthrough gases took place by splitting thegas molecules into oppositely chargedions.12 Since the X-rays were producedwhen cathode rays struck metal in the

Crookes tube, Thomson immediately9 (Crookes 1979, 135-136)10 (Crookes 1979, 137)11 (Crookes 1979, 137)12 (Falconer 1997, 227)

studied the cathode rays and showed thatthey not only carried an electric charge(as shown by Jean Perrin in 1895) but alsothat an electric charge was an indispens-able property of the rays.13

Within a few months, in a March 30,1897 lecture demonstration at a RoyalInstitution Friday Evening Discourse,Thomson gave his rst estimate of mass/charge for cathode rays produced in aCrookes tube. His experiment relied onthe heating effect of the rays so that theestimate was based on temperature dif-ferences (determined by means of ther-mopile technology) and statistical rea-soning.14 Thomson inferred that cath-ode rays were minute charged corpus-cles that were a constituent part of allatoms. While Thomson was criticizedfor both the roughness of the estimateand the overly-hasty conclusion, andwhile he did follow immediately withmore re ned experiments measuringthe de ection of the rays in electricand magnetic elds, the published re - port (October 1897) stated even morestrongly his speculations about the

13 (Falconer 1997, 227)14 (Falconer 1997, 227)

Figure 5. Cathode Ray Tube, Crookes style.Courtesy of the Harvard Collection of His-torical Scienti c Instruments.



22/68

corpuscular nature of the electron.15 Moreover, Thomson never let the in-evitable roughness of the measuresthwart his further research on gaseousdischarge (the light emitted from gasatoms excited by an electric current) based on that working hypothesis. Hiscathode-ray work was merely one of many steps towards his longstandinggoal to establish a coherent theory of gaseous discharge, for which he wonthe Nobel Prize in 1906.

Seeing the Light: From Phosphorescence to

Radioactivity

[Error] is the part of mans personal collabora-tion in the creation of scienti c fact. If I an -swered amiss, it was because I chose to reply tooquickly, without having asked nature, who aloneknew the secret.

Henri Poincar,The Valueof Science (1905), 326

This prediction once made, it evidently does notdepend upon him whether it is ful lled or not.In sum, facts are facts, andif it happens that they

satisfy a prediction, this is not an effect of our free activity .

Henri Poincar,The Value

of Science (1905), 328-329 15 Philosopher and historian of science TheodoreArabatzis emphasizes how many scientists weresimultaneously working with cathode rays in this period, and suggests that historians of scienceexamine how scienti c facts such as the cor -

puscular nature of cathode rays (electrons) areoften agreed upon through consensus in scienti ccommunities rather than being discovered byan individual. Although I focus on J.J. Thomsonhere, it is important to note that he was one of many who contributed signi cantly to the ac -ceptance of the belief that electrons denote realentities. (Arabatzis 1996, 432)

Phosphorescent materials obtaintheir luminosity from prior expo-sure to light. Two nineteenth-century physicists, a father and son, sought touncover what lay behind this piece of common knowledge. In attemptingto capture the evanescent glow of lu-minescent crystals with photographic

plates, their research ultimately led toan understanding of spontaneous ra-dioactivity. For that phenomenon, anaccidental absence of light fortuitouslyresulted in seeing the light that is,understanding what was unique aboutspontaneous radioactivity.

In 1859, Edmond Becquerel, pro-

fessor of physics at the Paris MuseumdHistoire Naturelle, invented the phosphoroscope (Figure 6, Figure 7)to study the luminescence of crystals.The phosphoroscope enabled him tomeasure the period of phosphorescence

noting the times at which the mate -rial absorbs the light (entering the holeson one rotating disk) and subsequentlyemits the light through the holes in thesecond rotating disk (not aligned withthe holes in the rst disk). Since thespeed of rotation could be adjusted very

Figure 6. Hand-held Becquerel Phosphoro-scope, with phosphorescent sample in theglass holder. Courtesy of the Harvard Col-lection of Historical Scienti c Instruments.



23/68

precisely, Becquerel was able to mea-sure the very brief time intervals duringwhich phosphorescence occurred in thedifferent substances examined.

In 1872, Becquerel turned his atten-tion to uranium salts together with hisson, Henri Becquerel (Figure 8) thethird generation of the Becquerel fam-ily to hold the Museum chair in phys-ics who continued this work after hisfathers death in 1891. Henris experi-ments with uranium salts would even-tually lead him to co-discover (with the

physicists Pierre and Marie Curie) theeven more startling property of spon-taneous radioactivity, for which they jointly won the 1903 Nobel Prize in

Physics.Triggering Henri Becquerels re-

search was the German physicistWilhelm Roentgens November 1895observation that invisible penetratingrays (X-rays) accompany the phosphorescence of accelerating cathoderays (observed fact 1).16 A stream of cathode rays in an evacuated tube cre-ated bright green phosphorescence onthe glass, emitting invisible rays whichpenetrated objects and left their sub-surface image on a photographic

plate. In January and February 1896,Becquerel attacked the related question:do phosphorescent bodies emit X-raysif exposed to high enough intensity of light (question A)? He experimented by placing phosphorescent crystals on photographic plates wrapped in light-tight black paper on his windowsill at

the museum, where sunlight wouldstimulate the mineral to glow.17 In his rst test on the various phos -

phorescent materials in his cabinets(Figure 9), Becquerel concluded: the phosphorescent substance emits raysthat penetrate black paper that is opaqueto light and reduce the silver salts of a photographic plate.18 One week later,on 2 March 1896, he reported to theAcademy of Sciences his ndings onuranium salts which hadnot been exposed to light. Heavily overcast skiesfor two days had made Becquerel puthis prepared plates in a drawer; after another three days without sun (so thatno sunlight was absorbed by the ura-nium), the plates were as black as when16 (Roentgen 1895)17 (Badash 1996, 22)18 (Becquerel 1896, 420)

Figure 7. Condensing Lens for a hand-heldBecquerel Phosphoroscope. Courtesy ofthe Harvard Collection of Historical Scien-ti c Instruments.



24/68

the uranium had been exposed to fullsun.

Becquerels preliminary observa-tion of these properties is sometimesdescribed as an accident. But wasthis really an accidental discovery?While it is a historical contingency

that the uranium salts happened to be placed with a photographic plate awayfrom sunlight for ve days, allowingBecquerel to observe the impact of the uranium alone on the photographic plate, Becquerel had a very well-prepared mind ready to interpret and testhis observation. His long, extensive

work with his father determining thecharacteristics of luminescent phenom-ena had prepared him to ask crucialnew questions of his material, and nd

new ways to use the available instru-ments, such as the photographic plate,to probe them.19

As historian of science Peter Geimer argues, Becquerels experiments in-volved photography of the invisiblewhich depended integrally on the sci-entists interpretative stance in the con-text of his current work:

Henri Becquerel opened a drawer in his labo-ratory and found images of uranium salts that hadregistered themselves on a photographic plate inthe dark. In all these cases the performers dis-

covered only afterwards what they themselveshad arranged before or what they themselves hadregistered but not seen.20

By his previous experience with the photography of luminescent materials, by arranging the photographic platein the dark drawer with the uraniumsalts, by developing the photographic

plate to see if there was an image of the crystals despite their being in thedark, and by recognizing the implica-tions of the exposed plate immediately,even reporting it the very next day tothe Academy of Sciences, Becquerelwas far from making an accidentaldiscovery.21

Thus, Becquerel had observedthat uranium salts emit penetratingrays evenwithout any prior exposureto light (observed fact 2). From thisBecquerel inferred that visible phosphorescence is not connected to theradiation from uranium (inferred fact2).22 Furthermore, after continued ex- perimentation, he observed that ura-

19 (Badash 1996, 22)20 (Geimer 2000, 120)21 (Badash 1996, 23)22 (Badash 2005, 37)

Figure 8. Photograph of Henri Becquerelwhen he was a young researcher. (Badash

1996, 22).



25/68

nium salts could substitute for a cath-ode ray tube, since both move the goldleaves in an electroscope in the sameway (observed fact 3).23 Becquerelinferred from this that the penetratingradiation of uranium is the same as the penetrating radiation of cathode rays(inferred fact 3).

Two questions arose from thesefacts. First, where does the penetrat-ing radiant energy emitted by uraniumcome from (question B)? Becquerelturned to work on the Zeeman effect

(the splitting of the spectral lines ina magnetic eld) to seek an answ er.Second, do any other substances giveoff this radiation, and, if so, howmuch (question C)? In investigatingthis question, Marie Curie discov-ered radium and polonium, two newelements which were both highly ra-

dioactive. Becquerel then used thesenew, more intense radiation sources,observed their de ection in a mag -netic eld, and discovered that therewere two kinds of rays spontaneouslyemitted. The kind that could be devi-ated, he equated with cathode rays(which were later recognized as high-speed electrons).

Becquerel uncovered two break-through physical and chemical factsabout radiation through this sequenceof collaborations with nature: thatspontaneous radiation occurs in somematerials (uranium, radium, poloni-um); and that spontaneous radiation isof two types, cathode rays and light.Moreover, in coming to see what wasunique about spontaneous radiation,Becquerel also came to understand the23 (Badash 1996, 24)

light of phosphorescent materials in anew light.

Conclusion

In these stories about scientists at-tempting to perceive and analyze theinvisible identifying radiant heatwith visible light, measuring proper-ties of miniscule subatomic particles,capturing the brief luminescence of crystals, probing the nature of thenewly recognized phenomenon of spontaneous radioactivity we wit -ness how late nineteenth-century physicists explored the invisiblewith novel instruments and work-ing hypotheses. Even rough resultsof measurements, when consistentwith conceptual categories guidingThomsons research, were able to bemarshaled in support of the electron-

corpuscle. Contingent occurrences even one so trivial as a week of cloudy skies could lead to a con -ceptual breakthrough when inter-

Figure 9. Four tubes of phosphorescent

chemicals. Courtesy of the Harvard Collec-tion of Historical Scienti c Instruments.



26/68

preted by an experienced researcher,such as Becquerel, in his work withuranium salts. Even the smallest of inferred facts can eventually yield amonumental return.

BibliographyArabatzis, Theodore. 1996. Rethinking theDiscovery of the Electron. Studies in Historyand Philosophy of Modern Physics 27, (4): 405-435.

Badash, Lawrence. 1996. The Discovery of Radioactivity. Physics Today 49: 21-26.

Badash, Lawrence. 2005 Becquerels blunder.Social Research 72: 31-62.

Becquerel, Henri. 1896. The Radiation FromUranium. Comptes Rendus 122: 420. Reprintedin A Source Book in Physics. By William FrancisMagie, 1935. New York, McGraw Hill: 610.

Crookes, William. 1979. On the Illumination of Lines of Electrical Pressure, and the Trajectoryof Molecules. Philosophical Transactions. PartI.

Daston, Lorraine. 1991. HistoricalEpistemology. In Questions of Evidence: Proof,Practice, and Persuasion across the Disciplines.James Chandler, Arnold I. Davidson, HarryHarootunian, ed.: 282-289. Chicago: TheUniversity of Chicago Press.

Falconer, Isobel. 1987. Corpuscles, Electrons,and Cathode Rays: J. J. Thomson and theDiscovery of the Electron. British Journal for

the History of Science 20: 241-276.

Falconer, Isobel. 1997. J. J. Thomson and theDiscovery of the Electron. Physics Education 32, (4): 226-231.

Geimer, Peter. 2000. Noise or Nature?Photography of the Invisible around 1900. InShifting Boundaries of the Real: Making theInvisible Visible. Helga Nowotny and MartinaWeiss, ed: 119-135. Zurich: AG an der ETH.

Preston, Thomas. 1894. The Theory of Heat. Edition 2. London: Macmillan and Co.

Roentgen, Wilhelm Konrad. 1895. The RoentgenRays. Sitzungsberichte der Wrzburger Physikalischen-Medicinischen Gesellschaft.Reprinted in A Source Book in Physics. ByWilliam Francis Magie. 1935. New York:McGraw Hill: 600-610.

Smith, George E. 2001. J. J. Thomson andthe Electron: 1897-1899. In Histories of theElectron: The Birth of Microphysics, edited by Jed Z. Buchwald and Andrew Warwick.Cambridge, MA: MIT Press, 21-76.

Thomson, J. J. 1897. Cathode Rays. TheElectrician. (May 12): 104-109. Also, Thomson,J. J. 1897. Cathode Rays. PhilosophicalMagazine. Vol. 44, Series 5.

Tyndall, John. 1882, rst published 1872.Contributions to Molecular Physics in theDomain of Radiant Heat.

Walton, Harold F. 1992. The Curie-BecquerelStory. Journal of Chemical Education 69, (1):10-15.


Photograph by Evangelos Kotsioris


27/68

Fact: Blood is Red 20

F act : b lood Is r edF act :

max (blood ) 645 nmA history of systematizing and quantifying colorfrom the late 19 th century to the 1950s

Lying on his back, he gazed up now into the high cloudlesssky. Do I not know that that is in nite space, and that it isnot a rounded vault? But, however I screw up my eyes andstrain my vision, I cannot see it but as round and nite, andin spite of my knowing about in nite space, I am incontest -ably right when I see a rm blue vault, far more right thanwhen I strain my eyes to see beyond it.

Leo Tolstoy, Anna Karenina (1877)

Did everyone see color in the sameway? A disastrous train crash in1875 was attributed to the fact that thetrain engineer did not, and hence failedto stop at a red light.1 On the other hand, the industrialist Joseph Lovibondshowed that the market price for our

could be accurately correlated to itscolor measured using the Tintome -ter he sold from 1887.2Color was alsoan important subject of philosophicalinvestigations, as well as an examplefrequently used to illustrate and exam-ine philosophical concepts. By the turnof the century, the question of color vi-

sion had been tackled by great thinkersincluding Aristotle, Descartes, Newton,Goethe, Maxwell and Helmholtz, who1 (Jennings 1896, 5)2 (Johnston 1996)

sometimes ercely rejected previoustheories.

In this paper I shall outline the histo-ry of measuring color from the late 19th century to the 1950s a period whichincludes the establishment of an inter-national standard for quantifying color

and the appearance of commercialcolorimeters based on using photoelec-tric cells rather than the human eye. Iwill examine the interactions betweenindustrial interests, different groups of scientists, theories about color and theinstruments used for measurement in particular two colorimeters from the

Harvard Collection of Historical Sci-enti c Instruments. I will conclude byconsidering different perspectives onthe history of science and how they can

Gregor Jotzu


28/68

be used to look at colorimetry during

the period covered in this paper.While philosophers were ponder-ing the question whether a color was a property of an object or something thatexists only in our perception, a very di-verse range of groups were looking for precise and authoritative methods for quantifying and systematizing color:manufacturers attempting to designmore ef cient lamps, pointillist paintersinvoking color theories to defend their art (and using them to make it morestriking), the textile industry lookingfor new dyes and general practitionersusing the redness of blood to deter-mine the health of their patients.3 Their views and requirements were crucialfor the developments that followed. In particular, whilst artists and artisanswere most concerned about the effectsof their works and hence focused on psychological aspects, medical doctorsand chemists cared less about their ownemotions when looking at a sample of blood, urine or solutions containing ti-tanium, and only used color as an easyand precise way to determine the con-

3 (Gage 1999; Kuehni and Schwarz 2008; John-ston 2001)

centrations of certain substances in asolution.

Color Quantifed by Eye

Von Fleischls Haemometer (Figures1&2), manufactured by C. Reichert inVienna from about 1880, was a popu-lar tool amongst medical doctors of the time.4 It was used to determine thequantity of haemoglobin in a sampleof blood. To do so an amount of blood(precisely speci ed by the use of a spe -

cially designed, calibrated pipette) wasmixed with a certain amount of water and placed in a semi-cylindrical vessel.The vessel was placed next to a trans- parent red glass wedge, tinted withCassiuss golden-purple. Upon thisthe light from a gas or oil lamp (day-light is not admissible) is projected by

a plate of plaster of Paris. The depthof color of the wedge increases fromone end to the other and is marked witha scale. Looking through both bloodand glass at the same time, the wedgewas moved until they both showedan equal intensity of red colour. Thenumber indicated on the scale was thenread off, where a reading of 100 should be found for a healthy person, or 14grams of haemoglobin in 100 grams of blood, and other scale readings should be converted linearly.

One book on clinical diagnosiswritten by a pathologist gave detailedinstructions on how to use this instru-ment, including the conditions of illu-

4 Inventory number 1996-1-0530 in the HarvardUniversity Collection of Historical Scienti c In -struments. The description cited here is: (Jaksch1899, 22-24)

Figure 1. The von Fleischl Haemometer.Courtesy of the Harvard Collection of His-torical Scienti c Instruments.



29/68


mination required for precise measure-ment.5 Although it did mention dif -culties which may arise because humancolor vision was used in the processof quanti cation, the author consid -ered these problems in a very techni-cal manner, as if he were speaking of an additional instrument. Errors werecaused by the way the retina functions,not because there was a fundamental problem with using human color vi-sion as a means of quanti cation. For example, the manual recommended

using both eyes and rest them everyfew seconds, preferably by looking atsomething green, the complementarycolor to red, thus increasing the sen-sibility to the latter color. In fact, the problems of color vis ion were summa-rized by saying that low retinal color sensibility for red may add 5 per cent

more to the overall error. Althoughthere were other methods for determin-ing the concentration of haemoglobin

5 (Wood 1905)

which were not based on using color,the colorimetric approach was gener-ally recommended as being faster andmore accurate.

Another physician recommendedcolorimetric methods to all those whoare interested in giving to their patientsadvice founded upon facts rather thanfads or fancies.6 He assigned great importance to making measurements of quantities such as the density of hae-moglobin in a patients blood, assertingthat in practice the name of the disease

is not so important as accurate knowl-edge of the physiology of the patient.

Constructing a Universal System for Color

Quantifcation

The underlying assumption of theclinical methods presented above (aswell as similar methods used in chem-istry and biology) was that proper-ties such as the redness of blood werequanti able. For optical physicists thesuccess of these colorimetric methodswould likely have served as an indica-tion that all of color vision should be

quanti able in a similar way. Amongthem the three-color theory developed by Young, Maxwell and Helmholtzwas generally accepted.7 This theoryexplained the phenomenon of color mixture: a combination of wavelengthsof the red and green part of the opti-cal spectrum could make light look the

same as light containing only wave-lengths from the yellow part. All pos-sible colors could be described by the6 (Stone 1923)7 (Sherman 1981)

Figure 2. The von Fleischl Haemometer.(Jaksch 1899).


30/68

stimulus of the three different types of

cones in the retina. Three coordinateswere hence suf cient to completelyspecify a color.

Many psychologists disagreed. Butthey also disagreed amongst each other.Some argued that color was not a puresensation (which could be described physiologically), but a perception,

which involved mental processes suchas memory. Therefore it was inaccessi- ble to complete quanti cation, at leastwithout a more advanced knowledge

of the human brain. Others insisted

that more than three coordinates werenecessary for a satisfactory descriptionof color aspects such as lustre, glow,gloss, transparency and body color should be considered. Furthermore,time dependent effects such as glitter,sparkle and icker, but also after-imag -es and adaptation had to be included to

complete the picture. The experimentalstudies of David Katz and others cor-roborated this view.8 Some scientists8 (Katz, MacLeod, Robert Brodie,1907-1972, tr.,and Fox, Charles Warren,1904- joint tr. 1935) The

Figure 3. The color coordinate system of the OSA report. R=Red, G=Green, V=Violet. Smallnumbers indicate wavelengths in nm. (Troland 1922, 527-591).



31/68


considered these effects to be unwantednoise, others saw them as essential.

After the rst World War, these on -going discussions became problematicdue to increasing pressure from vari-ous industries, which demanded thatnational standards institutions provide binding norms. The vast amount of col-or charts in use at the time (later calledan individualistic anarchy9) somearranged in circles, others in trianglesor simply in lists and tables were con -sidered to damage the economys ef -

ciency and competitiveness.10

Further-more, most of these charts could only be used for direct comparison with the products themselves, not for preciselyquanti ed measurement. Finding a so-lution which was generally agreed onwas seen as more important than whaton exactly this solution was based.

The Report of Committee on Colo-rimetry, published in 1922, was later seen as a crucial step in the develop-ment of an international color stan-dard.11 This committee had been set uponly three years after the foundation of the Optical Society of America (OSA)in 1916 and had ve members. Four of them were physicists working in indus-try, and the chairman had a backgroundin psychology. Their report named bril-

rst German edition, (Katz 1911), was publishedin 1911. See also (Niedere 1998, 5) and (John-ston 2001) for other studies. Of course this di-vision into physicists and psychologists wasnot complete. Some physicsists such as LudwigPilgrim also looked into the effects of adaptation,and some scientists had mixed backgrounds.9 (Guild 1934, 69-78)10 (Johnston 1996) gives a detailed analysis of the role of industrialists in the formation of color standards.11 (Troland 1922, 527-591)

liance, hue and saturation as the threeattributes of color, which can betreated as quantities and speci ed nu -merically and arranged into a systemsuch that neighbouring members differ from one another in each of the threeattributes by just noticeable degrees.

In this report,color was de ned assomething radically different in kindfrom its stimul[us], something entirely psychological, but something that wasstill quanti able. The (purely physical)

stimulus was seen as being completely

speci ed by [the] spectral distributionof a sample of radiant energy. The con-nection between these twoquantities (as they were now de ned) was madeusing psychophysicalthree-color excitation curves, which mapped thein nitely-dimensional space of spectraldistributions into the three-dimensional

space of colors. The authors admittedthat it was improbable that the curvesin [this report] faithfully represent theactual resonance functions of the el-ementary chromatic mechanisms in theretina. However this was not seen as problematic, as there was little doubtthat the results of future measurementswere potentially convertible into termsof such actual response functions.However, to ensure that these psy-chophysical data could be consideredconstant, a standard observer had to bede ned. This de nition included: limit -ing the eld of view to three degrees indiameter, restricting admissible inten-sities of light, specifying previous ex- posure and excluding a considerablenumber of individuals [who] possessrods in the centre of the retina. Thisvery limited standard observer was


32/68

the psychologists main point of criti-cism against the committees report.However, the system presented in thereport was compatible with measure-ment apparatus available at the time,and could be represented graphicallyin an understandable way (Figure 3).Furthermore, physicists being a morecoherent group, the impediment thatdetermining a spectral distribution(without using human color vision) wasactually no easy matter was success-fully glossed over.12 The fact that both

available technologies (the thermopileand the photocell) were considered in-adequate for such measurements wasnot a problem brought up in this OSAreport.13 The approach chosen in thisreport was hence in uenced more bya belief that precise spectroscopic mea-surements not involving humans eyes

should be possible in principle, rather than by a practical reality of such mea-surements.

The proposed system presented atthe 1931 International Commissionon Illumination (CIE), an organisationwhich met every three years, was verysimilar to that developed by the OSA

12 (Wright 1944; Guild 1934, 69-78)13 Even in 1934, physicist J. Guild (who contributed some of the data to the 1931 CIE standard)states: One thing seems certain: we are never likely to get rid of the dif culty of individualvariations. Thermopiles, while very constant intheir properties, are not suf ciently sensitive for incorporation in reasonably robust colorimetricapparatus, while photo-electric cells, far from being more constant in their properties than hu-man observers, are notoriously more variable,and I am told by those who are competent tospeak on such matters that there are inherent dif-

culties which make it unlikely that this defectwill ever be overcome. (Guild 1934, 69-78)

committee.14 The main difference wasthat new data was available for thethree-color excitation curves. It was based on measurements taken on only17 British men.15 The system passedthe vote, and although the Germanrepresentatives were more skeptical of the general approach, when the role of presenting a proposal for a colorimetricstandard was passed on to them at thenext meeting, they failed to add any-thing but very minor alterations to the1931 system. The 1931 CIE standard

unleashed a ood of activity [] inthe color industries and was used tospecify the requirements on road, railand aviation signals.16 Although someadditional standards (e.g. including alarger eld of view) have been intro -duced since, CIE 1931 is still in usetoday.

Color Black-Boxed

The Klett-Summerson photoelectriccolorimeter (Figures 4&5) was rst patented in 1940 and served a similar purpose as the von Fleischl haemom-eter.17 On rst sight, its shape and me -tallic exterior make it look like a trainengine, its appearance is quite differentcompared to the elegance of the Vien-nese instrument. It was more versa-tile, as it was not limited to measuring14 (Anonymous2007, 1-8; Judd 1950; OpticalSociety of America, Committee on Colorimetry1953, 100-152)15 (Guild 1934, 69-78)16 (Wright 2007, 9-22)17 Inventory number 2006-1-0003 in the Har-vard University Collection of Historical Scien-ti c Instruments; Summerson, William H. USPatent 2193437 (1940); Clinical Manual: (An-onymous1950)



33/68


blood, but can measure a variety of substances.

18Most importantly though,

the process of color-measurement has been (physically) black-boxed. A photo-electric cell was used instead of the hu-man eye, the source of illumination wasincluded in the apparatus (and could

be controlled by choosing a color l -ter) and the results were read off froma dial. All the user had to do was ll the

colorimeter tube with a known amountof the solution to be measured. It waslater marketed as taking the drudgeryand error out of the counting of bacterialcolonies. Nevertheless, its success wasnot immediate: in 1947 manufacturer E. Leitz still marketed models which usedcomparison by eye as high-end instru-

ments.19

Although the instrument was de-signed to give a reading directly pro- portional to the concentration of a givensubstance, when it became more widely

18 However, this was not completely new, asmodels as old as the von Fleischl, e.g. the Du- bosq model, also had this ability. There, insteadof using tinted glass, the unknown solution iscompared with a well-known solution. For anoverview of different colorimeters, see appen-dix titled TYPES OF COLORIMETERS ANDTHEIR USE (Myers 1924)19 (Leitz 1929)

used, many researchers published their results in Klett-units, along withspecifying which one of the providedcolor lters was used. 20 This highlightsthe in uence instrument makers had on

the science of the time. Furthermore, re-minds us thatuniversality is (and was)not a clearly de ned concept. This also becomes apparent in a debate which took place at the 1924 CIE meeting: French,British and American representativessuggested the use of electric lamps, astheir radiation did not depend on hu-

midity and temperature. The Germannational laboratory instead defended theamyl-acetate-based Hefner lamp, ar-guing that it could easily be built in anylab, whilst electrical lamps could not bemanufactured in a reproducible way.21

In any case, the international stan-dards created by the CIE did not seem

to nd their way into medical and bio -chemical colorimetry, as human color vision had already been dropped out of the equation by the successful imple-mentation of photo-electric cells. Thistype of colorimetry had now becomequite separate from the universal sys-tem of color which was developed un-der its in uence.

Conclusions: Looking at the History of Color

The development of colorimetry provides a very interesting, but slightlyunusual, case study for historians of science. Many studies have unveiled20 See e.g. (Westland and Beamish 1955, 1776-1778; Bello and Ginsberg 1967, 843-850; Wald-ron and Lacroute 1975, 855-865)21 (Johnston 1996)

Figure 4. The Klett-Summerson Photoelec-tric Colorimeter. Courtesy of the HarvardCollection of Historical Scienti c Instru -ments.


34/68

the process of decision-making in-volved in the establishment of scien-ti c facts the (temporary) settlementof a debate through various forms of authority where it was not obvious. In

this case it is much more apparent. Theminutes of committee meetings clearlyshow how this process worked, and therole of seemingly external actors suchas industrialists or instrument makerscan be traced more easily. Interestingly,the question of (different types of)vi-

sion , often an approach in current sci-

ence studies, here becomes theobject of study as well.22How can we further illuminate the

development of color measurement?Thomas Kuhns description of sci-

22 (Jasanoff 1998, 713-740)

enti c revolutions is dif cult to applyhere, as there was no closed commu-nity of colorimetrists, no paradigm to be changed.23 Some of the dif cultiesthat arose when physicists took up dis-cussions with psychologists could beexplained by the fact that their viewswere incommensurable, but this approach does not get us much further.

The concept of mechanical objec-tivity was developed by Lorraine Das-ton and Peter Galison mainly from theanalysis of images in scienti c atlases

so it requires some translation to be applied in this case.24 We may ask: werescientists worried about the projec-tion of their own preconceptions andtheories onto data? Were they willing

23 (Kuhn 1996)24 (Daston and Galison 2007)

Figure 5. Patent drawing of the Klett-Summerson Photoelectric Colorimeter. (US Patent No.2193437 (1940)).



35/68


to use more cumbersome methods solong as they did not collude in the sci-entists wishful thinking? The medicaldoctor recommending the use of colo-rimetry to the busy physician despitethe additional time required does t inhere.25 For him, making physiologicalmeasurements was more important thannaming a disease. On the other hand,even when commercial photo-electriccolorimeters were available, those re-quiring comparison by eye were stillwidely used. Of course, the methods

differ in degree, and not in kind butstill, the main criteria here seemed to be ease-of-use and statistical accuracy.

Thinking of Latourianimmutablemobiles certainly does provide an explanation for why the CIE system wasmore desirable than a zoo of dif cult-to-reproduce color charts.26 But why

then were the latter still so persistent?The ability of the OSA committee to produce tables and graphs gave itmore power in the war of standards

still it was not entirely successful.However , Bruno Latours observa-tion that the requirements put onknowledge are utterly different if one wants to use it to settle a localdispute or to participate in theexten-

sion of a network far away certainlyapplies here.27 Scientific systemswere not intrinsically and automati-cally universal, their universality hadto be achieved painstakingly. Colors, patients, flour and lamps all had to be papered over to fit into the net-work. However, in many ways these25 (Stone 1923)26 (Latour 1988, 1-40)27 (Latour 1987)

concepts do not go beyond beingde- scriptive only, at least in the sensethat they do not provide us with un-derlyingmotivations . 28

Certainly this period of the historyof colorimetry cannot be describedas a linear progression on a straight

path towards objective truth it isa much more complex subject. Be-yond this realization though, whichtheory of the history of science doesit support? Some of them aim moreat presenting the mentality of scien-

tists and their motives. Others focuson describing how scientific theoriesare extended into different aspects of our world. Each point of view suc-cessfully illuminates some aspectsof our story, and certainly makes itmore interesting. It also suggests adirection of further investigation. But

only if we were to look at our storyin a certain way, and decide to givemore weight to some of its elementsthan to others (be they individuals,scientific apparatus, communities or publications), could we really makea decision.

In this way, we are in a situationnot too different from the membersof the various committees describedhere. They had to choose a standardobserver and exclude certain aspects of color vision from their consideration tocome to a conclusion. But unlike themwe are not forced to do so. After all,not adopting only a single method for 28 See (Shapin 1988, 533-550) for an interestingcriticism of these ideas. On the other hand it may be argued that in the Latourian view, speaking of motivations (in a way that they are often used inthe sociology of scienti c knowledge) is not verymeaningful.


36/68

Translation of CIE 1931 resolutionson colorimetry. In 2007. Colorimetry :Understanding the CIE system., ed. J. Schanda,1-8. Hoboken, N.J.; Chichester: Wiley ; JohnWiley [distributor].

Klett-summerson photoelectric colorimeter.general directions: Detailed analytical procedures1950. . New York: KlettManufacturing.

Bello, Leonard J., and Harold S. Ginsberg. 1967.Inhibition of host protein synthesis in type 5adenovirus-infected cells. J.Virol. 1, (5): 843-50.

Daston, Lorraine, and Peter Galison. 2007.Objectivity. New York; Cambridge, Mass.: ZoneBooks; Distributed by the MIT Press.

Gage, John. 1999. Color and meaning : Art,

science, and symbolism. Berkeley ;Los Angeles:University of California Press.

Guild, J. 1934. The instrumental side of colorimetry. Journal of Scienti c Instruments11, (3): 69-78.

Jaksch, Rudolf v., 1855-1947. 1899. Clinicaldiagnosis : The bacteriological, chemical, andmicroscopical evidence of disease. 4th speciallyrev. and enl. by the author from the 3rd English

of the translation by James Cagney. ed. London:C. Grif n.

Jasanoff, Sheila. 1998. The eye of everyman:Witnessing DNA in the simpson trial. SocialStudies of Science 28, (5-6): 713-40.

Jennings, John Ellis. 1896. Color-vision andcolor-blindness. A practical manual for railroadsurgeons. Philadelphia,: F. A. Davis.

Johnston, Sean F. 1996. The construction of colorimetry by committee. Science in Context9, (04): 387.

Johnston, Sean, 1956-. 2001. A history of light and colour measurement : Science in the

Bibliography

looking at the history of science willnot threaten the ef ciency of our econ -omy.

shadows. Bristol ;Philadelphia, PA: Institute of Physics Publishing.

Judd, Deane Brewster, 1900-1972. 1950.Colorimetry. Vol. 478. Washington: U.S. GovtPrint Off.

Katz, David, 1884-1953. 1911. Dieerscheinungsweisen der farben und ihre beein ussung durch die individuelle erfahrung.Vol. I. Abt.; Ergnzungsband 7. Leipzig: J.A.Barth.

Katz, David, 1884-1953., MacLeod, RobertBrodie,1907-1972, tr., and Fox, CharlesWarren,1904- joint tr. 1935. The world of colour.London: K. Paul, Trench, Trubner.

Kuehni, Rolf G., and Andreas Schwarz. 2008.Color ordered : A survey of color order systemsfrom antiquity to the present. Oxford ;New York:Oxford University Press.

Kuhn, Thomas S. 1996. The structure of scienti crevolutions. 3rd ed. Chicago, IL: University of Chicago Press.

Latour, Bruno. 1988. Visualisation andcognition: Thinking with eyes and hands. In , ed.

H. Kuklick. Vol. 6, 1-40Jai Press. . 1987. Science in action : How tofollow scientists and engineers through society.Cambridge, Mass.: Harvard University Press.

Leitz, E., I. 1929. Leitz microscopes :Biological (binocular and monocular) polarising(petrographical) metallographic, universal

binocular of wide eld type, special models.magni ers, microtomes, colorimeters, projection

apparatus, accessories, etc. New York: E. Leitz,Inc.

Myers, Victor Caryl, 1883-. 1924. Practicalchemical analysis of blood : A book designed asa brief survey of this subject for physicians andlaboratory workers. 2nd rev. ed. St. Louis: C. V.Mosby.

Niedere, Reinhard. 1998. DieErscheinungsweisen der Farben und ihre

stetigen bergangsformen. Habilitationsschrift.,Universitt Kiel.

Optical Society of America, Committee onColorimetry. 1953. The science of color. In ,100-152. New York,: Crowell.



37/68


Shapin, Steven. 1988. Following scientistsaround. Social Studies of Science 18, (3) (August1): 533-50.

Sherman, Paul D. 1981. Colour vision in thenineteenth century : The young-helmholtz-maxwell theory. Bristol :Philadelphia, PA: A.Hilger ;Distributed by Heyden.

Stone, Willard John, 1877-1943. 1923. Bloodchemistry; colorimetric methods for the general practitioner, with clinical comments and dietarysuggestions. New York,: Hoeber.

Troland, L. T. 1922. Report of Committee onColorimetry for 1920--21. J.Opt.Soc.Am. 6, (6):527-91.

Waldron, C., and F. Lacroute. 1975. Effect of growth rate on the amounts of ribosomal andtransfer ribonucleic acids in yeast. J.Bacteriol.122, (3): 855-65.

Westland, A. F., and F. E. Beamish. 1955.Colorimetric determination of iridium with p-nitrosodimethylaniline. Analytical Chemistry27, (11) (nov): 1776-8.

Wood, Francis Carter, 1869-. 1905. Chemical andmicroscopical diagnosis, by francis carter wood... with one hundred and eighty-eight illustrationsin the text and nine colored plates. New York andLondon,: D. Appleton and company.

Wright, W. D. 2007. Paper from the golden jubilee book: The historical and experimental background to the 1931 CIE system of colorimetry. In Colorimetry : Understanding theCIE system., ed. J. Schanda, 9-22. Hoboken, N.J.;Chichester: Wiley ; John Wiley [distributor].

. 1944. The measurement of colour,.London,: A. Hilger ltd.

Photograph by Grace Kim


38/68

t echnology and F act

Jay Forrester and Magnetic Core Memory

This collection of engines and mechanisms disconcerts theFrench reader who was looking for a coordinated sequenceof hypotheses on the constitution of matter and a hypotheti-cal explanation of this constitution.

Pierre Duhem, The Aim and Structure of Physical Theory

Helen Keefe

Magnetic core memory is a scien-ti c instrument, but not in thesame way as a telescope or a barom-eter. From the early 1950s until its re- placement with the silicon chip in the1970s, core memory was the mecha-nism used by computers to store infor-mation entered by users into its system.As an instrument then, core memorydiffers from telescopes, barometers,and other similar devices because its primary purpose is to be a technologyrather than a tool for uncovering prop-erties about the universe. Still, thereis an inherent link between these twocategories of instruments that is rootedin the dual-nature of science itself, withits both epistemic and practical aims.A technology such as magnetic corememory can only run successfully if itis based on facts that accurately re ect

physical reality. Indeed, it would seemthat the most solidi ed facts are thosethat have crucial human technologiesconstantly depending on them. To this

effect then, technologies contribute tothe side of science that discovers facts,even though this is not their explicitaim.

The story of Jay Forrester a com - puter engineer turned business ana-lyst from MIT and the magneticcore memory he made is an apt casestudy through which to examine therelationship between technologies andscienti c theories. In addition to corememory, Forrester is known for hav-ing developed a theory called systemdynamics, a method for making mod-els of complex systems based on en-gineering principles applied to humanorganizations. In this way, Forrester incorporated knowledge gleaned fromhis work as a computer engineer into alarge-scale theory claiming to explainthe actual workings of a human social

system. Yet the legitimacy of theoriz-ing according to this paradigm whicheffectively regards human interactionsas a machine-like system is question -



39/68

Technology and Fact 32

able for many philosophers of science,including notably the physicist andhistorian of science Pierre Duhem. Acritical look then at the progression of Forresters dynamic career and scien-ti c contributions will illuminate howtechnology and theory interact, dem-onstrating how these distinct domainsare closely connected. At the sametime, this connection does not meanthat technologies can always be used toderive theories, since reality often goes beyond what the technology is capableof describing.

Technology Based on Fact: Magnetic Core Memory

In a time when nearly every collegestudent has his own personal computer,it is hard to imagine the world prior tothe current electronic age. Stranger still is the realization of how discon-

nected modern computers are fromtheir ancestral predecessors, whoseoriginal purpose was quite differentfrom the present application. Whencomputers were rst invented, they

were large, specialized calculating de-vices designed to meet wartime needs;and magnetic core memory was a ma- jor improvement on the quick-access of a computers stored information. The

rst successful invention of practicalcore memory happened while Forrester worked on the US Navys Whirlwind

Project, which was to build an aircraftsimulator to train bomber ight crews but which ultimately became a projectto make a new digital computer.1 Acrucial problem for computer engineersat the time was in the area of memoryand information access. Forrester ad-dressed this issue by determining how

to control magnetization through theintersection of two wires, developinga process called coincident currentaddresing. He found that the ferriterings arranged on the wires would ro-tate consistently in either of two ways,depending on which binary number passed through the system.2 One number caused an individual ring to rotatein a clockwise direction while the other caused a counter-clockwise rotation.

1 (Redmond and Smith, 1980) (Alderman andRichards, 2007)2 (Redmond and Smith, 1980)This summary rather oversimpli es the amountof work that Forrester and his team of graduatestudents particularly Bill Papian put into de -veloping the technology of core memory. It alsoleaves out the historical contexti.e. the pressuresthe team faced in meeting their deadline and de-livering positively on their agreement with thegovernment. For a more complete description,see Redmond and Smith, 180-185.

Figure 1. A magnetic core memory plane.This plane, made out of ferrite iron ringsstrung on wires and held together by awooden frame, would have been assembledin stacks within a rather gigantic computer.Courtesy of the Harvard Collection of His-torical Scienti c Instruments.


40/68

By means of these two different physi-cal effects, the core memory processedthe logical commands of a program andsaved information that had been enteredinto the computer. Indeed, this savingfunctionality made possible by hold -ing the positions of the rings madecore memory practically advantageousin yet another way, since it could main-tain stored information even when itssupply of electricity was turned off.

A closer look at Forresters work onmagnetic core memory reveals a rela-

tionship between scienti c theory andtechnology that characterizes the tra- jectory of this invention. Apparently,Forrester knew previously about the potential of magnetic cores for stor-ing data but could not determine howto make it practical.3 In other words,there was a theory derived from es-

tablished electromagnetic principlesthat dictated how such a technologycould work were it to be manufac-tured.4 With the pressure of a gov-ernment deadline on his back andstill lacking a memory base withrapid-access, Forrester investedtime into relearning the particular electromagnetic principles relevantto core memory, which was then justa prospect. According to this theory,the magnetic rings should serve the3 (Redmond and Smith, 1980)4 It would be misleading not to mention thatthere was a previous patent for core memory inexistence prior to Forresters, as well as two morethat happened simultaneously with his patent.(See Edwin D. Reilley, Milestones in Computer Science and Information Technology, WestportCT: Greenwood Press, 2003, 164). Yet Forrester is still credited with the invention, given that hismodel was the practical one that was rst used ina digital computer.

dual-function of being both writableand readable, allowing for both theinput and output of information.5 Asthe authors of Project Whirlwind:The History of a Pioneer Computer describe it, Forresters endeavor

was simply [to convert] an attrac-tive theoretical principle to reliable practice.6 In this way then, thetechnology Forrester built was bornof theory according to the usual, in-tuitive way that is, knowledge aboutthe way matter behaves (theory) be-came the basis for the design of the

practical craft (technology). Fromanother perspective though, the suc-cessful creation of magnetic core5 (Redmond and Smith, 1980)6 (Redmond and Smith, 1980)

Figure 2. Jay Forrester holding a plane ofmagnetic core memory. Forresters work asa computer engineer became the inspirationfor the system dynamics he later intro-duced to the realm of business and socialsciences. Courtesy of the Computer HistoryMuseum, Mountain Range, California.



41/68

Technology and Fact 34

memory itself pointed to the veracityof the electromagnetic principles thatinspired its design what was merely potential in the abstract became actual-ized in a concrete way, the core memory being a material proof for the theorysaccuracy. In this sense, making a tech-nology functions as a kind of experi-ment, even though the two operationshave markedly distinct aims. Theories,in a way, bene t when the technolo -gies they inspire prove victorious.

Fact Based on Technology: System Dynamics

As the example of magneticcore memory shows, theories andtechnologies are crucially relat-ed. Technologies need theoreticalknowledge for both their very exis-

tence and also their proper function-ing; in the same vein, a technologycan act like an experiment in provid-ing confirmation of the theory fromwhich they arise. But what happenswhen this order is reversed, and man-made technologies instead becomethe foundation of knowledge rather

than theories? Given the previousobservations as to the closeness of these two scientific branches, sucha question does not yield an obvi-ous answer. The next episode in theForrester story offers an example of this new kind of paradigm, and thekind of results that follow when tech-nology and theory switch places.

Forresters career as a scientist isuniquely marked by its interdisci- plinary nature. In 1956, he ended hiswork in computer engineering to join

the faculty of the MIT Sloan Schoolof Management, where he becameinvolved in business and social sci-ences.7 In the same year, he beganworking on system dynamics, amethod for producing models show-ing how complex systems work, andmeant to be applied indiscriminatelyto human organizations and physicalsystems alike.8 Grounded in engi-neering knowledge, system dynamicsis a theory inspired by technology, asit treats the reality of human social

interactions as though it were a ma-chine.9 Anticipating criticisms of thisclaim about human behavior, Forrester states:

People are reluctant to believe physical sys-tems and human systems are of the same kind.Although social systems are more complex than physical systems, they belong to the same classof high-order, nonlinear, feedback systems as do physical systems.10

According to Forrester then, hu-

man interactions resemble a ma-chine; like technologies, they can becontrolled or made to function in a particular way. Forrester then en-dorses a deterministic view of human beings, which denies that humans actfreely and thus explains why their in-teractions can be managed like a ma-chine. In this way, Forresters sys-7 (Forrester, 1992)8 (Forrester, 1958)9 I am here omitting a discussion of how 17th and18th century thinkers used mechanical explana-tions to describe the workings of the world, al-though I do touch on it a bit later. Though rele-vant, this important consideration is not the focusof this paper, and would complicate the issue inways that are unnecessary.10 (Forrester, 1958)


42/68

tem dynamics is a theory followingfrom a technological inspiration.

Duhem: A Critique

The theory of system dynamics isdominated by the use of models pro- posing how interactions happen. Usingmodels to demonstrate theories how-ever is not an innovation on the partof Forrester; especially during the 19th century, many scientists have employedmodels as a way to both understand nat-ural phenomena and communicate their

ndings to non-scientists. But scienti ctheories that rely too heavily on mod-els alone frustrate Pierre Duhem, a late19th and early 20th century physicist and philosopher, who wrote about the limitsof such an approach in his well-knownwork The Aim and Structure of Physical Theory .11 Duhem nds that a sciencecompletely committed to the purelymechanical explanation of physical phe-nomena, can be neither very accurateno

the matter of fact 2

Documents