CROTCHETS & QUIDDITIES

Ludwik Fleck and the Art-of-FactKENNETH WEISS

In science we routinely assume thatwhatever theory we may have aboutthem, facts themselves are things thatare—out there in the world to be ob-served. However, it has become ratherstandard for historians and philoso-phers to assert instead that “facts” arehuman constructs understandableonly in a particular historical or soci-etal context. They note that scientiststypically argue over very different the-ories to interpret the same data. We inturn resist this deconstruction by out-siders who themselves don’t have toface the struggles of understandingNature. Nonetheless, there may havebeen more than idle poetry in Keats’assertion that “truth is beauty andthat is all ye need to know.” Facts maybe more determined by the theoreticallenses through which we view theworld than we like to think.

TRUTH AND PROGRESS INSCIENCE: “THE BONDS OF

HISTORY CAN NEVER BE CUT”

Thomas Kuhn’s famous book TheStructure of Scientific Revolutions1

had a transforming effect across boththe academic and popular culturelandscapes. It gave us the satisfyingterm “paradigm shift” for the episodic

bursts of change by which we so often,perhaps so vainly, characterize ourown research or field. Kuhn arguedthat a scientific revolution occurswhen a new explanation for the avail-able data becomes accepted by thebody of scientists, a sociocultural phe-nomenon very different from the pre-vailing notion of science as a fact-driven process. The notion has beenapplied, but not without discussion, toall fields of endeavor including an-thropology.2,3

A new theory may be more accuratein some ways, or account for morefacts, but Kuhn’s notion of a paradigmshift was not about contests betweendifferent ways to apply an existingtheory, like that of evolution, to theavailable facts, such as fights betweencladists and phenetic systematists, orout-of-Africa versus regional continu-ity models of evolution. Rather, a par-adigm shift is a gestalt change inwhich a new interpretation of thebody of existing facts replaces, and isincommensurable with, an existingtheory. Evolution contrasts with cre-ationism in this respect, for example.

Most scientists strongly assert—be-lieve—that such changes constituteprogress in the sense of major stepscloser to understanding the truth thatwe assume is out there to be found.However, it is perhaps not sufficientlyappreciated that scientific theories arealmost always inconsistent even withsome of the established facts of theirtime, and this was true of the victorsin classical Kuhnian “revolutions,” in-cluding even the archetypal Coperni-can one. Whether this is progress to-ward truth is harder to answer thanwe may think. Kuhn drew an analogy

with biological evolution: scienceworks by contingently sorting throughavailable facts or ideas in the contextof their own time, but cannot really beteleologically heading for truth, be-cause we have no way to know whereto find it.

Kuhn was influenced by a little-known 1935 book he had stumbledacross, Genesis and Development of aScientific Fact, by a Polish physiciannamed Ludwik Fleck.4 Fleck had antic-ipated (or shaped) many of Kuhn’sideas. Though rather obscure today,Fleck has been called the founder of thephilosophy of modern medicine. Hisconcern was not so much with revolu-tions but with the way that even thesupposed facts of science are driven bycontext. He stresses that “concepts arenot spontaneously created but are de-termined by their ‘ancestors’.” Fleck’sterm comparable to Kuhn’s “paradigm”was thought collective (we might say“school of thought”) and his objectivewas to understand how an unorches-trated community of scientists estab-lishes what are considered facts arounda theory that necessarily grows out ofhistorical roots. What counts as factvaries with time and context. Fleck lik-ens science to troops on the march, asmall vanguard followed by a mainbody. New observations provide somecorrective, but which of the vanguardsthe main troops follow is unpredictable,and affected—or determined—by so-ciopolitical and cultural factors.

This challenges our cherished my-thology that we are doing “original”research. Most of us are not as origi-nal nor independent as we may fancy.Without a herdlike ideological coher-ence, science as the public enterprisewe know today might not be possible,since any individual can only see or doso much. But the herd itself defineswhat progress is, and the more who

Kenneth Weiss is Evan Pugh Professor ofAnthropology and Genetics at Penn StateUniversity.

Evolutionary Anthropology 12:168–172 (2003)DOI 10.1002/evan.10118Published online in Wiley InterScience(www.interscience.wiley.com).

Truth may be beauty, but more in the eye of the beholder than we generallyacknowledge. Does it matter?

168 Evolutionary Anthropology

participate the stronger and more re-sistant to displacement (creative orfree thinking) it becomes, and the pre-vailing thought collective becomes athought constraint. “Whole eras” canbe ruled by a thought constraint. Her-etics are not well tolerated; we don’tburn them at the stake but the Ly-senko era showed that we may sendthem to freeze to death in laborcamps, and today we freeze them outof research funds. As with any ideol-ogy, defendants of the thought collec-tive argue that they defend it becauseit’s true.

Ludwik Fleck worked in the pre-World War II era in his native Poland.Poland was intimately involved inconflicts in that region, including theCommunist revolution in neighboringRussia. Social, historical, evolution-ary, and contextual analysis of humanaffairs was part of the intellectual lifeof that era in Europe. Fleck’s choice ofthe term “thought collective” seems toreflect that context, and he states arather classically Hegelian or Marxistassessment of conflict in science, asbeing between opposed theoreticalviews that resolve into a new consen-sus. Fleck says that “after the changein thought style, the earlier problem isno longer completely comprehensi-ble.” This contrasts a bit with Kuhn’sidea, because a problem is not identi-cal to an explanatory framework. Butto Fleck the two are inextricably re-lated.

ON BECOMING FACT: THEWASSERMANN TEST FOR

SYPHILIS

The predominant medical problemof the time was infectious disease, andFleck exemplified his ideas by show-ing how concepts of syphilis, growingfrom earlier notions of causation, ledto the development of the Wasser-mann test to detect the disease. Syph-ilis had long been viewed primarily inmoral dimensions, a “carnal scourge.”In part a legacy of the age-old humeraltheory, diseases of generalized orvague symptomatology were assumedto be due to fouling of the blood. Thisnotion grew to the point that the early20th Century demanded a “blood test”to detect the disease. Infectious mi-croorganisms were known by then,and Pasteur had recently organized(and orchestrated) the scientific andbureaucratic machinery in Francearound the concept of infectious dis-ease causation.5 European nationsscrambled not to be left scientificallybehind in this endeavor. However,symptoms, specificity, and sensitivityof known reactions and available testsmade both the disease and its diagno-sis problematic. Symptoms of syphiliswere not coincident with the visual(microscopically detectable) presenceof the spirochete bacterium Trepo-nema pallidum, though people wereconvinced that was the cause. The dis-ease took different, often indefinitelydelayed, courses in different individu-

als, and other diseases such as leprosyalso generated a positive test reactionsas well as similar symptoms.

Under these conditions, whether itwas a “fact” that someone had syphilisor not was not purely objective but de-pended on the assumptions about itsparticular cause. Many investigators—the “thought collective”—worked, ifonly in a loosely coordinated way, tofind a syphilis blood test. In 1906, Au-gust von Wassermann and colleaguesdeveloped the famous reaction thatbore his name. On the assumption thatthe test directly detected bacterial anti-gens, a positive Wassermann reactionwas used to define syphilis, separating itby this single criterion from other dis-eases with similar symptoms.

Despite these elements of whatseemed like “truth,” the inferencesfrom experiments leading to the Was-sermann test were incorrect or purelyempirical in essential ways. The theo-retical understanding of the reactionwas in fact wrong. A positive reactionwas interpreted as proving spirochetecausation because the theory was thatantigen-antibody reactions were spe-cific. The method to detect the pres-ence of antigen produced by the spi-rochete is known as complementfixation (Figure 2). In the first step,syphilitic tissue used as a source ofantigen was mixed with blood fromthe person being tested. In an infectedperson, the antigen binds to antibody,and complement (a component of theimmune system) is fixed to the anti-gen-antibody complex. A properlyprepared reagent, like red blood cells(rbc’s), is then added. This causes noreaction in samples from infected per-sons that at this state no longer con-tain free complement, but the un-bound complement in the blood of anuninfected person binds to the rbc’s,producing an easily detected hemo-lytic reaction.

Much of Fleck’s book is an accountof the many ways this process was aprisoner of its history and thethought-collective’s erroneous choiceof assumptions: “It is only after thechoice has been made that the associ-ations produced by it are seen as nec-essary.” I can’t give the details here,but over time, a legion of investigatorshad to tinker with many of the condi-tions to make the test even reasonably

Figure 1. A. Thomas Kuhn (http://www.malaspina.com/site/person_733.asp); B. Ludwik Fleck(reprinted with permission, from Polish Philosophy web site: http://www.fmag.unict.it/PolPhil/Fleck/Fleck.html).

CROTCHETS & QUIDDITIES Evolutionary Anthropology 169

reliable—at least for clear-cut or late-stage cases. They had to filter the“facts” empirically and selectivelythrough their presumptive lens be-cause, among other things, it was laterfound that uninfected tissue used as asource of antigen could also generatea positive reaction. We now know thata constituent of healthy tissue, cardi-olipin, mimics the antigenic proper-ties of T. pallidum. Other conditionsalso induce antibodies to cardiolipin,including the other treponemal organ-isms, and why T. pallidium does so isstill unclear. Nonetheless, the drivefor a blood test paid off. The reason-ing by which a positive reaction wasrelated to syphilis was wrong. But thetest became fact.

We can consider this in light of late19th Century science, with its belief inatomic units of ultimate, universalcausation. Biomedical science foundits causal units in infectious organ-isms. The contemporary view was re-

flected in G.B. Shaw’s introduction tohis 1911 play The Doctor’s Dilemma:6

“The whole art of healing could besummed up in the formula: Find themicrobe and kill it. . . . When therewas no bacillus it was assumed that,since no disease could exist without abacillus, it was simply eluding obser-vation.” These views were so en-trenched that Shaw mused that it wasnot clear whether vaccination pro-grams are forced by doctors onto thepublic or vice versa.

The causal metaphor of the timewas one of combat between an indi-vidual against an invasion of microor-ganisms. However, many microorgan-isms live commensally with us, andthe highly variable relationship be-tween the spirochete and symptomsshows that the metaphor was imper-fect. The parasite and antibodies arefound in the blood, but syphilis is notreally a disease “of” the blood the wayanemia or leukemia are. The patho-

gen can be present without the symp-toms, or vice versa, so that definingsyphilis as “the disease caused by[Treponema] pallida” reflects a thought-collective’s preconceptions about cau-sation as much as it does the reality.History constrains thought so that atany time most of us can’t think in anyother way. Since Fleck’s time we havedeveloped more precise ways includ-ing direct detection of treponemalgene sequences to detect the presenceof what we have defined as the cause.But the etiological concept is not theonly possible, and may not be the best,definition of diseases. They could al-ternatively be classified by the treat-ment they respond to, virulence, theorgans they affect, or their symptoms,and there are often multiple causalfactors. But that is not how we thinkin science in our present era.

A MODERN RERUN?

There are striking parallels in to-day’s AIDS epidemic. AIDS, like syph-ilis, is viewed as a disease of the blood(because of immune deficiency), andhas even been seen as another “carnalscourge” of moral dimensions. Onecan have symptoms of AIDS withoutHIV and vice versa. The standard HIVtest is an indirect one for antibody tothe pathogen, and does not detect per-fectly nor immediately after infection.Many host factors like nutrition andthe burden of other disease affect thecourse of the disease, which can beindefinitely delayed. As with syphilis,the symptoms are not unique to HIVas the “cause”; there are many ways toget pneumonia, but pneumonia insomeone HIV-positive is defined asAIDS. There is even a cottage industrythat claims the huge momentum ofthe AIDS “thought collective” hasbeen misdirected by an unjustifiedproclamation in the 1980s that HIV isa proper etiological entity. Are theyflat-earthers or could the thought col-lective actually be wrong? Or if truthis not clearly an attainable goal, is“wrong” even the wrong concept?

PARADIGM ORPROPAGANDA?

Fleck describes how a thought col-lective maintains itself by variousways of inculcation. Academic de-

Figure 2. Complement fixation test. Left, positive test; right, negative test. Moving top down,antigen (diamonds) introduced to blood sample binds to antibody (Y) which capturescomplement (dark ovals) in infected (left). Complement remains free in uninfecteds (right).When red cells (rbc’s, grey ovals) are added, they are bound by the free complement inuninfecteds, and lysed. In uninfected, no complement is free to lyse the rbc’s. Schematic.

170 Evolutionary Anthropology CROTCHETS & QUIDDITIES

grees are required for induction intothe guild. Popular science ensures itsdominance in the public arena, as canbe seen today on Nova and in theTimes, through oversimplified, emo-tive vividness, personal narrative, andgossipy controversies to lionize fig-ureheads for the prevailing view (e.g.,Wassermann, Pasteur) and promoteenthusiasm for the accepted theory.Isolated facts are fitted gropingly tothe theory in professional journals,and stereotypical digests are thenbuilt into textbooks, handbooks, orlaboratory manuals that inculcatenew students. The latter are often outof date even when published.7

Fleck goes to some lengths to showhow scientific illustrations are ideo-grams of the accepted view. Here hechooses examples of anthropologicalinterest. He compares anatomicaldrawings from medieval times to thepresent (for a visual tour of this fasci-nating history see www.nlm.nih.gov/exhibition/dreamanatomy/da_intro.html). The earlier drawings were func-tionally schematic. For example (Fig-ure 3), 12th Century figures show thethorax as a figurative rib “cage,” whilein 1543 Vesalius drew a death-meta-phor that obscured anatomic details.Today we feel it is “scientific” to knowthose details, but Fleck notes that this

requires much advance preparationand decision-making (e.g., dissectingaway all material but the skeleton,leaving the cartilage but not intercos-tal muscle, membranes or ligaments,identifying muscle attachments) thatmakes the result a product of our cur-rent view of what the skeleton “is.”

BUT TOASTERS WORK!

Fleck and many analysts sinceKuhn have stressed the ways histori-cal context leads a view to becomeaccepted, and that the accepted viewis not the only possible one. The incul-cation process converts observations

Figure 3. Ideogrammatic schema rather than detail in earlier scientific anatomical drawings: what is the rib cage? A,B: medieval ideas; B:Vesalius’ Hamlet and the deeper meaning of death; C: a modern (19th Century) mechanical view showing all the “scientific” details(Sources: A, B, D,4 C7).

CROTCHETS & QUIDDITIES Evolutionary Anthropology 171

into “facts” that define what is “true”because the thought collective agreesabout them. This leads historians, bi-ographers, and philosophers to pro-claim the fallibility of science, that sci-entists are but arbitrary products oftheir times, or even that the onlyTruth we approach is that of our ownvested interest.

There is some merit to these allega-tions, but scientists do not cling to theirbelief system only out of ideology orcaprice. Science has grown to predom-inance because, as one of my studentsput it, “toasters work.” But does thismean that we understand the real na-ture of metals (bimetallic thermostats),polymers (plastic knobs), and electrons(heating elements)? Or is a toaster sim-ply well within the empirical predictivepower of our experience? Similarly, theWassermann test has successfully diag-nosed exposure in many thousands ofinstances, and nobody questions a rela-tionship between the pathogen andsymptoms. So are we really being mis-led in any important way by our collec-tive ideologies? Is it of more than mildhistorical interest why or that we viewthe pathogen as “the cause” of AIDS orsyphilis?

We still seek point causes today, butthey are a different kind of scourge:bad behavior or bad genes. The mid-dle 20th Century proclaimed the con-quest of infectious disease, and ourbiomedical fixation moved dramati-cally, first to environmental, and thento genetic causation, the idol of to-day’s biological thought collective. Toparaphrase Shaw, when there is nogene we assume that since no diseasecan exist without such a gene, it mustsimply be eluding observation. In thesense that theory and fact are insepa-rable, the assumed gene is consideredfact—we even name the gene for thetrait—when it has not yet been iden-tified (genes “for” dyslexia, homosex-uality, and diabetes are examples).

Every issue described here in regardto infectious causation also applies tocurrent genetic view of disease causa-tion. We develop DNA (blood) tests todetect mutations in clear-cut cases,but genetic causation is far less clearin most cases, because a person canhave the symptoms without the alleleor the allele without the symptoms,and the symptoms are highly variable

and sometimes indefinitely delayed.There are countless examples, but ourdetermination to find genes is un-checked.8 Ironically, in having ourcollective heads turned the geneticway, we may be overlooking elusiveinfectious pathogens responsible formany chronic diseases after all.9,10

Evolutionary theory may be espe-cially vulnerable to these kinds of ideo-logical problems, because inferencesthat deal with the unobservable pasthave to be so highly indirect. The waywe apply genetics to adaptive argu-ments, for example, rests heavily ontheory.7 This applies to anthropologicalinterests, and in my next installment I’lllook at another anatomical exampleraised by Fleck: the structure of the hu-man brain, in the subtle lights andshadows of evolutionary theory.

In regard to progress, scientistswhen pressed often claim that allwe’re trying to do is to develop betterpredictive approximations. But deepdown, we act as if we believe, andmany assert that we’re approachingthe real truth. Whether or not that’san illusion, it is a powerful motivator.But there are often alternative expla-nations. Wassermann and his contem-poraries “tuned” their work to melo-dies to which they could resonate.Exceptions were overlooked, ignored,or rationalized, and experiments de-signed that left little room for alterna-tive interpretation. The same kind ofphenomenon as been observed in eth-nology and archeology, such as whatthe “facts” are regarding sophisticatedpre-Columbian Amazonian cultureand technology,11 or whether behav-ioral evolution explains why (or that?)the Yanomami are “the fierce people.”Kuhn notes that scientists struggle toaccommodate “anomalies” that are“reinterpreted to make them con-form.” Almost all theories “containsome element of wishful thinking bytheir scientific proponents.” That’s notwhat engineers do: they build in em-pirical safety factors so they need notworry about the ultimate theory, be-cause they really are just trying tomake toasters that work.

The point is not that a given view ofthe world is superficial or wrong, butthat it is somewhat arbitrary and con-textual and neither necessarily moretrue than alternative views, nor a step

on a path to the presumed absolutetruth. But the chosen view of the worldcan have impact beyond the mere intel-lectual imprisonment of ideas. For ev-ery path taken there is a path not taken.For example, putting our effort errone-ously into genetics when environmentor infection are the more importantcause of disease, or misapplying evolu-tionary theory to justify racial profiling,would be detrimental to knowledge andour own society.

In practice, organized science maydepend upon thought-collectives—theory, paradigms, or even ideology.But if these are cultural artifacts, andthe best we humans can actuallyachieve is an approximation with pre-dictive value but unknown relevanceto truth, then the art-of-fact may bewhat counts most in science.

NOTES

I welcome comments on this column:kenweiss@psu.edu. I have a feedbackpage at www.anthro.psu.edu/rsrch/weiss_lab/index.html. I thank AnneBuchanan, Malia Fullerton, NancyTuana, and John Fleagle for criticallyreading this manuscript, and Dom Eg-gart on toasters.

REFERENCES

Many things discussed here can beprofitably explored by web searching.

1 Kuhn T. 1962. The structure of scientific revo-lutions. Chicago: University of Chicago Press.2 Chamberlain JG, Hartwig WC. 1999. ThomasKuhn and paleoanthropology. Evol Anthropol 8:42–44.3 Cartmill M. 1999. Revolution, evolution, andKuhn: a response to Chamberlain and Hartwig.Evol Anthropol 8:45–47.4 Fleck L. 1935. Genesis and development of ascientific fact (1976 reprint, with editorial com-ment by Kuhn and others). Chicago: Universityof Chicago Press.5 Latour B. 2000. The pasteurization of France.Cambridge, Mass: Harvard Press.6 Shaw GB. 1911. The doctor’s dilemma (1980reprint). New York: Penguin.7 Weiss KM. 2002. Come to me my melancholicbaby! Evol Anthropol 12:3–6.8 Weiss KM, Buchanan AV. 2003. Evolution byphenotype. Perspectives in Biology and Medicinein press.9 Ewald P. 1994. Evolution of infectious disease.New York: Oxford Press.10 Ewald P. 2000. Plague time: how stealth in-fections cause cancers, heart disease, and otherdeadly ailments. New York: Free Press.11 Raffles H. 2002. In Amazonia: a natural his-tory. Princeton: Princeton University Press.

© 2003 Wiley-Liss, Inc.

172 Evolutionary Anthropology CROTCHETS & QUIDDITIES