Kindi Perspectives

495

The Relation of Historyof Science to Philosophyof Science in TheStructure of ScientiordfcRevolutions and Kuhnrsquoslater philosophical work

Vasso KindiUniversity of Athens

In this essay I argue that Kuhnrsquos account of science as it was articulated inThe Structure of Scientiordfc Revolutions was mainly defended on philosophicalrather than historical grounds I thus lend support to Kuhnrsquos later claimthat his model can be derived from ordfrst principles I propose a transcendentalreading of his work and I suggest that Kuhn uses historical examples asanti-essentialist Wittgensteinian ldquoremindersrdquo that expose a variegated land-scape in the development of science

T S Kuhn was a physicist who became a historian but who also wrote abook the Structure of Scientiordfc Revolutions (1970)1 that proved seminal inphilosophy of science In that book unlike what was typical in philosophyof science before the so-called historical turn there are quite a few refer-ences to particular cases from the history of science After the Structureand all the praise and controversy that it evoked Kuhn published Black-Body Theory and Quantum Discontinuity (1987) a book of history of scienceThe question that repeatedly arose was what is the relation of Kuhnrsquos phi-losophy of science to the history of science He was asked for examplewhether the Black-Body book exhibited the pattern of development thatwas elaborated in the Structure His commentators wanted to knowwhether the facts pertaining to quantum theory provided conordfrmation forhis philosophical model whether they could serve as a testing ground butalso whether his account of these particular facts was shaped by his philo-sophical model They wanted to know that is whether his philosophy was

I would like to thank Professor Theodore Arabatzis of the University of Athens and ananonymous referee for helpful comments and suggestions

1 Subsequently referred to as Structure or SSR

Perspectives on Science 2005 vol 13 no 4copy2006 by The Massachusetts Institute of Technology

496 History of Science and Philosophy of Science in Kuhnrsquos Late Works

dependent upon history but also whether in doing history he wasinordmuenced by his philosophy Kuhn vehemently denied both

[ ] I have myself resisted attempts to amalgamate history andphilosophy of science though simultaneously urging increased in-teraction between the two History done for the sake of philosophyis often scarcely history at all (Kuhn 1980 p 183)

I have said repeatedly and I will say again you cannot do historytrying to document or to explore or to apply a point of view (Kuhn2000d pp 313ndash314 emphasis in the original)

If you have a theory you want to conordfrm you can go and do historyso it conordfrms it and so forth itrsquos just not the thing to do (Kuhn2000d p 314 emphasis in the original)

Irsquom never a philosopher and a historian at the same time (Kuhn2000d p 316)

Kuhn with this response guards himself against a number of objectionsthat will be recounted later Yet independently of Kuhnrsquos reaction thequestion remains what is the relation of history of science to philosophyof science in Kuhnrsquos model Kuhn made use of history he even claimedthat the two disciplines history and philosophy can do and should inter-act and fertilize each other (Kuhn 2000d pp 315-316) If the relation be-tween them is not that of theory to evidence what is it Irsquoll confine myselfto the exegetical point aiming at a coherent account that would compriseboth what Kuhn does and what Kuhn says he does I will try in particularto reconcile Kuhnrsquos historical orientation in the Structure with his laterclaim that his model of science can be derived from first principles2 I willbegin by illustrating the wide range of different opinions on the issue inthe recent and not so recent secondary literature on Kuhn and I will thenproceed to propose my own account

I How the relation between history and philosophy of science in the caseof Kuhnrsquos work is perceived in the recent literatureIn some of the recent literature one finds the following claims as regardsthe relation between history and philosophy of science in Kuhnrsquos work

1 Michael Friedman in his Dynamics of Reason repeatedly characterizesKuhnrsquos work as the ldquotheory of the nature and character of scientiordfc revolu-tionsrdquo (Friedman 2001 pp 19 41ndash44 119) Friedman argues in favor ofrelativized constitutive a priori principles in the sciences and he states that

2 See Kuhn 2000b p 95 Kuhn 2000c

Perspectives on Science 497

history of science paradigmatically illustrated in Kuhnrsquos work providesthe conordfrmation needed He explicitly calls Kuhnrsquos contribution our ldquobestcurrent historiography of sciencerdquo (Friedman 2001 pp 43 44 51 52)and he is not referring to Kuhnrsquos purely historical research He is alludingto Kuhnrsquos non-cumulative model of science involving paradigm changeAlthough Friedman draws comparisons and highlights differences be-tween Kuhnrsquos work and the philosophies of Carnap or Quine he clearlyconsiders the former to be historiographical in nature

2 Alexander Bird in his book Thomas Kuhn says ldquoKuhnrsquos The Structure ofScientiordfc Revolutions is not primarily a philosophical text Rather it is awork in what I call lsquotheoretical historyrsquordquo (Bird 2000 p viii) By ldquotheoreti-cal historyrdquo Bird understands an empirical investigation into the historyof science which by being theoretical requires as he says a deep engage-ment with philosophy (Bird 2000 p 29) Bird does not explain how phi-losophy turns an empirical investigation into a theoretical project but ad-vances the view that Kuhn acts as a scientist3 He claims that Kuhninvestigates the pattern of historical development and provides both a de-scription and an explanation Scientiordfc change is the puzzle to be solvedand after observing a regularity in nature Kuhn presents a hypothesis toexplain it (Bird 2000 pp 67 141) However the description offered byKuhn which distinguishes between successive phases of normal and revo-lutionary science is according to Bird inaccurate and so the explanationwhich involves the ldquotheory of paradigmsrdquo is if not redundant and mis-taken rather ldquoinadequate to the factsrdquo (Bird 2000 p 49) Bird believesthat Kuhn took the wrong turn4 when he supposedly abandoned his em-pirical studiesmdashwhich are considered to be not only historical but alsopsychologicalmdashin favor of a more explicitly Kantian critical standpointthat draws conclusions from ordfrst principles

3 Steve Fuller in his Thomas Kuhn A Philosophical History for Our Times(2000) advances a convoluted reading On the one hand he credits Kuhn

3 One may infer from Birdrsquos suggestion that even in the sciences theoretical endeav-ors require the assistance of philosophy However given Birdrsquos description of Kuhnrsquos proj-ect (ie as scientiordfc in the inductivist Baconian manner) one cannot see how philosophyintervenes The only allusion to philosophy by Bird is when he credits Kuhn with unstatedconservative philosophical presuppositions empiricist positivist and Cartesian which al-legedly did not allow him to break with the tradition

4 It seems that the wrong turn is not solely the turn to philosophy but the turn awayfrom a particular philosophy associated with externalist epistemology and externalist andcausal theories of reference (Bird 2000 pp 278ndash279) See also Bird 2002 SurprisinglySharrock and Read note in their book pace Bird ldquoKuhnrsquos apparently deepening commit-ment to lsquoscientiordfc naturalismrsquo as his career progressedrdquo (Sharrock and Read 2002 p 223n 2 also pp 202ndash203)

with a normative ideal which qualiordfes him [Kuhn] for inclusion in whatFuller takes to be the Platonist cult5 On the other he criticizes Kuhn forldquomarginalizing his prescriptivismrdquo in order to maintain modern sciencersquosstatus quo Actually it is claimed that Kuhn may have entered a ldquoFaustianbargainrdquo According to Fuller Kuhn saw that science in its present formeven if it fell short of Kuhnrsquos own normative ideal provided ldquoa stable mili-tary-industrial infrastructure and virtually the only source of legitimateauthority for an increasingly fragmented and volatile populacerdquo (Fuller2000 p 74) Kuhn then in Fullerrsquos view became ldquostrategically vaguerdquo asregards the status of his book wavering between prescription and descrip-tion Thus he managed to ldquoward off the drastic calls for the disestablish-ment of sciencerdquo (ibid) What is more in order to preserve the authorityof science Kuhn according to Fuller restricted the historical basis of hismodel to examples drawn from the period between 1620ndash1920 so thathis normative ideal is not challenged by different developments before andafter this period (Fuller 2000 p 73) This ldquohopscotching across the centu-riesrdquo serves to give the impression that it is possible ldquoto understand thescientiordfc turn of mind regardless of the time and place in which science ispracticedrdquo (Fuller 2000 p 215)

Fullerrsquos concern is mostly political (ie that Kuhn does not questionsciencersquos present status quo) and does not at all address the philosophicalproblems that would arise had Kuhn based his model on historical factsirrespective of whether they were few or many He does not consider the isought divide nor is he concerned with the problems of underdeterminationand induction His non-political objections focus on Kuhn being a coun-terfeiter of history but also an inventor of mythical constructs or on amore charitable tone of Weberian ideal types in order to sensitize us tosalient features of the object of inquiry (Fuller 2000 p 195) Fuller be-lieves for instance that Kuhn developed the concept of normal science bysuperimposing different perspectives of science from different moments inhistory in order to construct a mythical image that he then treated as ac-curate6 So in Fullerrsquos view Kuhn had undertaken a covert conservativepolitical mission which he served by being unfair to the facts and by beingvague as regards the status of his philosophy

4 Finally Wes Sharrock and Rupert Read maintain that Kuhnrsquos objec-tives are overwhelmingly philosophical (Sharrock and Read 2002


5 Besides Plato and Kuhn the Platonist cult includes according to Fuller AugusteComte Leo Strauss Max Planck and ldquoprobably all [ ] incarnations [of positivism]rdquo(Fuller 2000 p 38)

6 It is here that Fuller brings in the discussion Panofskirsquos iconography from the historyof art (Fuller 2000 p 61)

pp 106ndash109) Commenting on the ordfrst section of the SSR which seeks arole for history they write

[A] main purpose of The Structure of Scientiordfc Revolutions is to make acase as to how scientists do in fact come to replace one theory withanother This makes it sound as though SSR is one of his historicalstudies but it is not that How then is the historical stuff theldquostalking horserdquo for the philosophical how is the latter aspect dom-inant in SSR SSR differs from Kuhnrsquos properly historical studiesfor he is not here primarily concerned to detail what occurred invarious speciordfc episodes in the history of science but instead tosay how the events in such episodes should be philosophically con-strued (Sharrock and Read 2002 p 27 emphasis in the original)

In this passage Sharrock and Read claim that Kuhnrsquos interest in historicalfacts is philosophical What they mean by that is that Kuhn needed aproperly done history of science to enable him to understand scientificchange ldquoin a way that directly conflicted with the American philosophy ofscience Kuhn inherited and which he thought basically mistakenrdquo (Sharrockand Read 2002 p 6 emphasis in the original)7 They take his aim to bethe revision of a particular philosophical image of science and they pres-ent Kuhnrsquos project of reconstructing the history of science as ldquoa pretextrdquo(Sharrock and Read 2002 p 10) It is not that in their view Kuhn was notreally interested in history only that his real purpose apart from possiblymisleading appearances was philosophical The historical objective wassubjected to the philosophical According to Sharrock and Read Kuhnrsquoshistorical studies clarify Kuhnrsquos philosophical arguments and it is anldquooverreactionrdquo to say that his model can be derived wholly from first prin-ciples (Sharrock and Read 2002 p 199-200)

The two authors claim that they put forward a Wittgensteinian defla-tionary reading of Kuhn From their perspective Kuhn albeit not alwaysconsistently aimed on the one hand to dissolve philosophy of science oreven cure us from it (Sharrock and Read 2002 p 211) and on the otherwith respect to science he aimed at leaving everything as it is (Sharrockand Read 2002 p 209) So the status of Kuhnrsquos project is taken to bephilosophical in a particular sense ie in a quietist Wittgensteiniansense In their view it is not philosophical in a robust normative sense be-cause it does not meddle with scientific practice it does not aim at issuingstandards of validity and method


7 ldquoWe have portrayed Kuhn as continuously concerned largely with one issuemdashspell-ing out the meaning of properly historical studies of episodes in the history of science forthe philosophy of sciencerdquo (Sharrock and Read 2002 p 199)

Now one could say that Kuhnrsquos project need not be philosophical toleave science as it is Even if it were merely historical it would again notinterfere with what scientists do History of science even when it is prop-erly done (ie the way Kuhn suggested) is not really relevant to scienceproper So Sharrock and Read need to advance an argument in defense oftheir claim that Kuhnrsquos work is philosophical and not historical in charac-ter

To that end they observe first that Kuhnrsquos account of science isldquolargely unevidencedrdquo (Sharrock and Read 2002 p 107) that we have totake his claims ldquoon trustrdquo (ibid) that his historical examples offer ldquopre-cious littlerdquo (Sharrock and Read 2002 p 108) The import of the histori-cal cases cited is ldquoto exemplify and dramatize the progress of philosophicalrevolutions-but that is perhaps their only philosophical relevancerdquo(Sharrock and Read 2002 p 109) Then Sharrock and Read ask whetherany of the above reflect negatively on Kuhn And they answer ldquoAs a mat-ter of fact not at all Rather we take it as a strong indication that one ofour central claims is true that Kuhn is a philosopher of sciencerdquo (ibid em-phasis in the original) They end the discussion by citing Kuhnrsquos own dec-laration of his philosophical goals And they comment ldquoWhat is strikingis first the degree of self-identification as a philosopher [ ] and secondthe surprising and indeed trumpeted willingness to dispense with thedensity of actual historical examples in favor of the abstraction of lsquofirstprinciplesrsquo To say it again Kuhn is philosopher above all (Sharrock andRead 2002 p 110) So what Sharrock and Read are virtually saying isthat Kuhnrsquos ldquoessential identityrdquo is philosophical because first he fails toprovide evidence for his claims and second he declares it to be so Theyseem to argue as follows If Kuhn were making empirical generalizationshe would have provided evidence He does not provide evidence hence heis a philosopher since he also says so himself

II Early assessments of the relation of history of science to philosophyof science in Kuhnrsquos modelFrom early on the empirical basis of Kuhnrsquos model of science was taken tobe an issue of concern Israel Scheffler maintained ldquothere can be no appealto ostensibly paradigm-neutral factual evidence from history in support ofKuhnrsquos own new paradigmrdquo (Scheffler 1972 p 367)8 If Kuhn were toclaim that he says he would be guilty of self-refutation SimilarlyShapere has noted that historical facts are not open to direct inspectionand they can equally bear out opposing philosophical views concerning


8 See also Schefordmer 1967 pp 21ndash22 53 74

scientific development (Shapere 1980 p 31) He says that Kuhn got car-ried away by ldquothe logic of his notion of paradigmrdquo and that the relativismthat ensues is not the result of empirical historical research but ldquothe logi-cal outgrowth of conceptual confusionsrdquo (Shapere 1980 pp 37-38)

NR Hanson (1965) detects a ldquological imperfectionrdquo in Kuhnrsquos meth-odology He says that Kuhn wavers between putting forward a genuinehistorical thesis on the one hand and an elaborate set of definitions onthe other In the first case we have an informative and yet possibly falsethesis and in the second an unfalsifiable exposition of the meanings thatthe terms ldquoparadigmrdquo and ldquorevolutionrdquo have or are given by Kuhn

This methodological issue would seem to affect our entire concep-tion of the historiography of science We are as much desirous inbeing illuminated about the facts as in being illuminated by Kuhnrsquosdecisions concerning how he will use certain expressions We maybecome better historians from information of either kind but itwould nevertheless be a help for Professor Kuhn to make unambig-uously clear which of these two endeavors did inform his very im-portant book (Hanson 1965 p 375 emphasis in the original)

Hanson acknowledges Kuhnrsquos contribution in the historiography ofscience and asks solely for the disambiguation of the use of Kuhnrsquosexpressions

Paul Feyerabend having read a draft of The Structure of Scientiordfc Revolu-tions sent to Kuhn two letters that were subsequently published byHoyningen-Huene (1995) In these letters Feyerabend criticized Kuhn forusing a double talk ldquoevery assertion may be read in two ways as the re-port of a historical fact and as a methodological rule You thereby takeyour readers inrdquo (Hoyningen-Huene 1995 p 355) This is a ldquobewitchingway of presentationrdquo (ibid) Feyerabend says that covers up a ldquoquestion-able monolithic ideologyrdquo (Hoyningen-Huene 1995 p 367)9mdashthat of theconservative character of normal sciencemdashin the form of history Hecharges that Kuhn never states clearly that his model amounts to ldquoanidealrdquo but insinuates instead that this is what historical research teacheshim (Hoyningen-Huene 1995 p 360)10 But history according to Feyera-bend is irrelevant to methodology (Hoyningen-Huene 1995 p 366) ldquoIsrdquodoes not imply ldquooughtrdquo and Kuhn should refrain from putting for-ward mere beliefs as if they were indisputable and inescapable facts


9 See also Hoyningen-Huene 1995 p 35510 Similar remarks are made in Feyerabend (1970) Kuhn is blamed of intentionally

blending ambiguous descriptive and prescriptive claims (Feyerabend 1970 p 199)

(Hoyningen-Huene 1995 pp 355 368) Historical facts Feyerabendsays admit of alternative interpretations and Kuhn should not hide thisfrom his readers (Hoyningen-Huene 1995 p 355) He should not beafraid of their criticism and he should make his point of view explicit(Hoyningen-Huene 1995 pp 366 355)

Another different point of criticism though is that Kuhn is led by hisldquohidden predilection for monism (for one paradigm) to a false report ofhistorical eventsrdquo (Hoyningen-Huene 1995 p 367 see also p 381) Soaccording to Feyerabend Kuhn not only hides his ideology in the coversof history but he is also guilty of falsely reporting historical events

Lastly Janet Kourany develops a different line of criticism In her viewKuhn provides no historical justiordfcation for his model The historical ex-amples are scattered sketchy undocumented and sometimes they appeareven to refute Kuhnrsquos claims (Kourany 1979 pp 50 52) Statements likeCopernicus saw a star where Ptolemy had seen a planet or Lavoisier sawoxygen where Priestly saw dephlogisticated air are used by Kuhn accord-ing to Kourany as ldquoa little more than imaginative illustrations of his posi-tion rather than items of historical support for itrdquo (Kourany 1979 p 55)At least part of the justiordfcation offered by Kuhn in support of the claimthat scientiordfc development is not cumulative Kourany says ldquoseems littlebetter that an argument a priorirdquo (Kourany 1979 p 49)

III Summation of the critical arguments and outline of the accountoffered by themIn summary the positions advanced regarding the relation of history ofscience to philosophy of science in Kuhnrsquos work in both early and late sec-ondary literature can be charted out as follows

1 It is claimed on the one hand that Kuhnrsquos work is clearly historio-graphical (FriedmanBird) or empirical in general (Bird) In that context itis often claimed that the historical account given by Kuhn is partly atleast inaccurate (Bird Fuller Feyerabend)11 Kourany however maintainsthat the historical research needed for the factual basis of Kuhnrsquos model iscompletely absent (Kourany 1979 p 56) while Sharrock and Read notethat Kuhnrsquos account is largely unevidenced

2 Moving a little further up from the factual level Bird credits Kuhnwith ldquotheoretical historyrdquo which can be taken either as a purely empiricalinductive generalization inferred from inspected individual cases or in the


11 It should be noted here that although Feyerabend speaks of inaccuracies he recog-nizes that facts admit of alternative interpretations

sense alluded to by Feyerabend of a Hegelian-like philosophy of history(Hoyningen-Huene 1995 p 353)12

3 A different line of criticism ordfnds Kuhnrsquos account wavering between de-scription and prescription (Fuller Feyerabend Hanson) In particular ac-cording to Fuller and Feyerabend Kuhn disguises his ideology and propa-ganda in the covers of history

4 Finally a number of commentators (Sharrock and Read KouranyShapere) recognize the a priori or philosophical status of Kuhnrsquos model inorder either to blame him for unfounded beliefs (Kourany Shapere) or tocredit him with particular philosophical agendas (Sharrock and Read)

In what follows I will give my own account of Kuhnrsquos model whichdiffers considerably from the above I will ordfrst show that in the StructureKuhn lays out a philosophical project and does not derive his model fromhistorical evidence Then I will argue that this philosophical projectwhich draws upon but is not based on historical examples is very muchsimilar to the grammatical investigations undertaken by Ludwig Witt-genstein but also to Strawsonrsquos transcendental analysis

I will not dwell on the critical points made above for a number of rea-sons First because I agree with those of Kuhnrsquos commentators who claimthat Kuhn is not always clear or explicit as regards the descriptive or nor-mative status of his work Secondly because I believe that the charges ofan alleged political agenda on the part of Kuhn or of the intentional be-witchment of Kuhnrsquos readership are highly speculative Lastly because Iwant to give if possible a charitable account of Kuhnrsquos historically ori-ented philosophy of science Kuhn has been repeatedly described in theliterature as ldquophilosopher manqueacuterdquo (Bird 2002 p 459) as ldquo[not knowingor understanding] the philosophical heritage he was working in andagainstrdquo (Bird 2002 p 460) as ldquoin a state of blissful but perhaps forgiv-able innocencerdquo (Friedman 2001 p 19) Yet I do not want to base my as-sessment of Kuhnrsquos work on grounds that pertain to his biography I willnot right away dismiss what he says by attributing to him ignorance or el-ementary logical mistakes13 If Kuhn were to base his philosophy on his-tory for instance he would have to address the isought divide the prob-lems of underdetermination and self-refutation but also his very limited


12 Feyerabend claims that Kuhn just like Hegel takes history to be a judge the dif-ference being that Kuhn refers to the past while Hegel to the future Strangely Feyerabendputs Wittgenstein together with Hegel

13 Kuhn remarks ldquo[People treated me as though I were a foolrdquo (Kuhn 2000bp 315)

empirical base If on the other hand Kuhn were to do theoretical historyin the Hegelian manner he would have to respond to the typical criticismraised against such approaches namely unfettered speculation and teleol-ogy

Of course Kuhnrsquos limited formal philosophical training gives one rea-son to suppose that logical mistakes and problematic ramiordfcations ofsome of the views he puts forward may have passed undetected yet I wantto check whether the text itself can sustain a coherent account that alsoagrees with Kuhnrsquos own explicit remarks on the issue under consideration

IV Kuhnrsquos philosophical arguments for the non-cumulative growthof scienceIn chapter 9 of the SSR Kuhn discusses a key contention of the booknamely the view that science progresses non-cumulatively He asks why achange of paradigm should be called a revolution why the emergence of anew paradigm works destructively for the old He starts with a simileJust like political revolutions he says scientific revolutions require orpresuppose a period of crisis People do not just import revolutions Theyfirst have to experience dissatisfaction with the institutions they have inorder to proceed to change them In a similar manner in science the newparadigm emerges only after normal scientific research falls short of thescientistsrsquo expectations Nowhere does Kuhn say that historical researchshowed him as a matter of fact that crisis always precedes revolutionsReference to historical examples in this respect is made for reasons ofillustration

Then Kuhn proceeds to give arguments why the assimilation of a newsort of phenomenon or of a new scientiordfc theory requires the rejection ofthe old paradigm First he explains that it is logically possible to add newphenomena or a new theory to an old paradigm This may happen if theparadigm expands in a new domain or when the developed theory inte-grates on a higher level previously held beliefs So if logic does not pre-scribe the rejection of the old paradigm what does Kuhn claims ldquothere isincreasing reason to wonder whether it [the ideal image of science-as-cumulative] can possibly be an image of sciencerdquo (Kuhn 1970 p 96) Whatkind of reason is that Initially at least the reason seems to originate inhistory

[T]he assimilation of all new theories and of almost all new sorts ofphenomena has in fact demanded the destruction of a previous para-digm and a consequent conordmict between competing schools ofscientiordfc thought Cumulative acquisition of novelties proves to bean almost non-existent exception to the rule of scientiordfc develop-


ment The man who takes historic fact seriously must suspect thatscience does not tend toward the ideal that our image of its cumu-lativeness has suggested (Kuhn 1970 p 96 emphasis added)

After however this gesture towards history Kuhn puts forward anotherline of argument ldquo[A] second look at the ground we have already coveredmay suggest that cumulative acquisition of novelty is not only rare in factbut improbable in principlerdquo (ibid emphasis added) Letrsquos see what his argu-ments are now The ordfrst says that unanticipated novelty emerges only af-ter an anomaly is detected Anomaly is simply deviation from the nor-malcy laid out by the old paradigm When a new paradigm turns ananomaly into a normal lawful phenomenon it cannot be compatible withthe old And though Kuhn claims that logical inclusiveness however per-missible is a historical implausibility he insists ldquothe examples of discov-ery through paradigm destruction [ ] did not confront us with merehistorical accident There is no other effective way in which discoveriesmight be generatedrdquo (Kuhn 1970 p 97)

The second argument is again philosophical If we accept Kuhn saysthe prevalent conception of scientiordfc theories as it was formed by the log-ical positivists and their successors then no theory can ever be challengedAdequately interpreted all theories can be preserved as special cases ofsubsequent ones and all grievances against them can be attributed to theextravagant and ambitious claims made by the fallible human beings Ifthat is the case then science would stop

But to save theories in this way their range of application must berestricted to those phenomena and that precision of observationwith which the experimental evidence in hand already dealsCarried just a step further (and the step can scarcely be avoidedonce the ordfrst is taken) such a limitation prohibits the scientistfrom claiming to speak ldquoscientiordfcallyrdquo about any phenomenon notalready observed [ ] But the result of accepting them would bethe end of the research through which science may develop further(Kuhn 1970 p 100)

Kuhn says that if we accept the standard image of cumulative growth inscience (an image that has no room for the challenge and rejection of theo-ries) then we end up with an absurdity Not a logical absurdity but an ab-surdity nevertheless if we consider how science is practiced Kuhn charac-terizes this point a tautology (ibid) Without unrestricted commitmentto a paradigm there could be no normal science Without normal sciencethere could be no surprises anomalies crises Without crises there couldbe no extraordinary science Without all these there is no science


If positivistic restrictions on the range of a theoryrsquos legitimate ap-plicability are taken literally the mechanism that tells the scientiordfccommunity what problems may lead to fundamental change mustcease to function And when that occurs the community will inevi-tably return to something much like its pre-paradigm state a con-dition in which all members practice science but in which theirgross product scarcely resembles science at all (Kuhn 1970 p 101)

There is one ordfnal argument against cumulative growth in science againphilosophical It is the argument about meaning change which is sup-posed to show that one cannot derive Newtonian from Relativistic dy-namics as the positivists surmised

So in this crucial part of Kuhnrsquos account of science reliance on historyis only supplementary The reasons he gives for the non-cumulative devel-opment of science are mainly philosophical and not factual RobertWestman (1994 p 82) who revisits Kuhnrsquos The Copernican Revolution re-inforces that point He writes that ldquo[w]hen Kuhn argues [in the SSR] forthe radical transformative character of lsquoseeingrsquo that occurs in the after-math of a revolution he produces a ordfctive speech delivered at an undeter-mined moment by a lsquoconvertrsquo rather than the speciordfc utterance of a his-torical agentrdquo Westmanrsquos observation that Kuhn ldquoproduces ordfctive speechrdquoinstead of concrete historical evidence lends I think further support tothe claim that Kuhn did not use history to ground his philosophicalclaims

I agree then with commentators like Sharrock Read and Kouranywho recognize the philosophical aspect of Kuhnrsquos work But I disagreethat this follows from the description of his account as ldquounevidencedrdquoSharrock and Read argue invalidly that Kuhn is a philosopher because hedoes not provide evidence for his claims whereas Kourany speaks re-proachfully of a priori arguments after she showed the thin and question-able historical basis of Kuhnrsquos model Sharrock and Read go further andcredit Kuhn with a therapeutic philosophy which much like Wittgen-steinrsquos is supposed to cure philosophical misconceptions but they do notelaborate on how history features in this context Their suggestion thathistorical cases simply exemplify or dramatize philosophical abstractionscasts on Kuhn a very traditional philosophical light (any philosophy ofscience can use historical examples) does not account for the historicalbend of Kuhnrsquos approach and leaves unexplained where his so-calledphilosophical abstractions are derived from Moreover their claim cannotaccount for the fact that Kuhn insists that he does not write history forphilosophical purposes


My view is that Kuhn is engaged from the start in a philosophical en-terprise because his target is philosophical He aims at bringing down anideal image of science that may have been drafted by the philosophers butheld nearly everybody captive This ideal image is attacked with philo-sophical arguments as it was shown above but also with the help of thehistory of science as it will be shown below The question of course ishow exactly Kuhn viewed and used history and how important it wasgiven that later in his work he contends that ldquomany of the most centralconclusions we drew from the historical record can be derived insteadfrom first principlesrdquo (Kuhn 2000c p 112)

V Kuhnrsquos later account of the relation between history and philosophy ofscienceThe sentence cited immediately above is taken from a longer passage inwhich Kuhn describes clearly how he views his project retrospectively

[M]y generation of philosopherhistorians saw ourselves as buildinga philosophy on observations of actual scientiordfc behavior Lookingback now I think that that image of what we were up to is mis-leading Given what I shall call the historical perspective one canreach many of the central conclusions we drew with scarcely aglance at the historical record itself The historical perspective wasof course initially foreign to all of us The questions which led usto examine the historical record were products of a philosophicaltradition that took science as a body of knowledge and asked whatrational warrant there was for taking one or another of its compo-nent beliefs to be true Only gradually as a by-product of our studyof historical ldquofactsrdquo did we learn to replace that static image with adynamic one an image that made science an ever-developing enter-prise or practice And it is taking longer still to realize that withthat perspective achieved many of the most central conclusions wedrew from the historical record can be derived instead from ordfrstprinciples Approaching them in that way reduces their apparentcontingency making harder to dismiss as a product of muckrackinginvestigation by those hostile to science (Kuhn 2000c pp 111ndash112)

In this passage Kuhn maintains that he and other philosophershistoriansof his generation were initially under the impression that they were ad-vancing a new philosophy of science based on the observation of actualscientiordfc practice past and present Old philosophers like the logicalpositivists were dealing with an ideal image of science completely de-tached from the experience of those who were involved in the scientiordfc en-


terprise whereas Kuhn and his contemporaries thought that they usheredin the actual life of science14 Later Kuhn proceeds to state that the studyof historical facts wasnrsquot at all necessary It may have helped geneticallybut logically it wasnrsquot needed at all They could reach the same conclu-sions if they just considered science from a historical perspective whichrequires us to look at things as developing over time Given this perspec-tive the static image of science would be immediately replaced by a dy-namic one All the rest ie the crucial parts of the model would then fol-low There would be no ordfxed Archimedean platform to judge therationality of individual beliefs but a moving historically situated rea-soned comparative evaluation of change of belief no resort to facts thatare prior to the beliefs they are supposed to supply evidence for and noconvergence to an ultimate truth

Kuhnrsquos contention is that the consideration of any developmental pro-cess would yield these characteristics They are not peculiar to science butnecessary features of any evolutionary practice (see Kuhn 2000c pp 116119) The only thing required in order to ordfnd them is to ldquoapproach sci-ence as a historian mustrdquo ie by ldquopick[ing] up a process already underwayrdquo (Kuhn 2000b p 95) In that sense Kuhn says he derives his conclu-sions from ordfrst principles ldquoIrsquove reached that position [ie that facts arenot prior to conclusions drawn from them] from principles that must gov-ern all developmental processes without that is needing to call upon ac-tual examples of scientiordfc behaviorrdquo (Kuhn 2000c p 115) It is also fromordfrst principles that Kuhn derives ldquospeciationrdquo (the branching out of dis-tinct scientiordfc specialties after some revolutionary change) andincommensurability both of which he sees as the necessary prerequisitesfor the advancement of knowledge and the authority of science

With much reluctance I have increasingly come to feel that thisprocess of specialization with its consequent limitation on commu-nication and community is inescapable a consequence of ordfrst prin-ciples Specialization and the narrowing of the range of expertisenow look to me like the necessary price of increasingly powerfulcognitive tools (Kuhn 2000b p 98)

Lexical diversity and the principled limit it imposes on communicationmay be the isolating mechanism required for the development of knowl-edge (Kuhn 2000b pp 98ndash99)


14 In (Kuhn 2000b p 95) Kuhn says ldquoI and most of my coworkers thought historyfunctioned as a source of empirical evidence That evidence we found in historical casestudies which forced us to pay close attention to science as it really was Now I think weoveremphasized the empirical aspect of our enterprise (an evolutionary epistemology neednot be a naturalized one)rdquo

I am increasingly persuaded that the limited range of possible part-ners for fruitful intercourse is the essential precondition for what isknown as progress in both biological development and the develop-ment of knowledge [ ] [I]ncommensurability properly under-stood could reveal the source of the cognitive bite and authority ofthe sciences (Kuhn 2000b p 99)

In a later article however Kuhn allows for some albeit minimal con-tribution of historical observation regarding speciation He says that un-like the thesis that facts are not prior to the conclusions drawn from themspeciation ldquois not a necessary or an a priori characteristic [of a historicalperspective] but must be suggested by observations The observations in-volved [ ] require in any case no more than a glancerdquo (Kuhn 2000cp 116)

Since recourse to history is limited to ldquono more than a glancerdquo I do notthink that there is signiordfcant change of position between the two articlesKuhn needs history only to provide him with the historical perspectiveOnce this perspective is suggested all the rest follow ldquoWhat has for meemerged as essential is not so much the details of historical cases as theperspective or the ideology that attention to historical cases brings withitrdquo (Kuhn 2000b p 95)

VI Assessment of Kuhnrsquos later accountKuhn clearly lays emphasis on first principles to avoid the problems hewould face had he given prominence to the empirical aspect of his workThis move puts him ldquosafelyrdquo on the philosophersrsquo side He avoids contin-gency and all the criticism mentioned above (underdetermination self-refutation limited empirical basis) He also manages to preserve impor-tant elements of the philosophersrsquo standard image of science by emphasiz-ing the dynamic appraisal of change of belief rather than belief tout courtsome continuity across revolutions communication among scientists rea-soned evaluation of incremental change of belief based on the ever-presentvalues of science (Kuhn 2000c pp 112-119)

The cost of this double move is that exclusive reliance on a priori princi-ples (with history entering only by a glance) not only raises new concernsmainly concern over the justiordfcation of such an account but it does notexplain the difference between science and other developmental processesWhat is more the alignment of Kuhnrsquos model with that of the traditionalphilosophers deprives it of its notorious radical pronouncements (disconti-nuity in the development of science incommensurability conversion ex-perience etc) If the change of belief is incremental (even if adjustmentsare required) if the rationality of the comparative evaluation is accounted


for by invoking typical standards (shared neutral observations commoneven if equivocal values) then the development of science over timeseems more continuous than the SSR had us believe Kuhn of course an-ticipating objections like this one does not fail to observe that emphasison the appraisal of change of belief rather than appraisal of belief tout courthas radical philosophical repercussions namely that the Archimedeanplatform with higher criteria of rationality ldquois gone beyond recallrdquo (Kuhn2000c p 115) that the evaluation of beliefs is not against an independentworld but only comparative that there is no ultimate truth to be reachedno unique method of science I do not want to underestimate the revolu-tionary character of these suggestions Yet I believe that Kuhn under theordferce and relentless criticism that he received from the philosophers wastoo eager to play down some of his most radical ideas in order to accom-modate in his model observations that seemed to his critics preposterousto have been overlooked and omitted It was repeatedly pointed out tohim for instance that scientists do communicate even in periods of cri-ses that their communication is reasoned that established scientiordfc theo-ries are not completely overthrown and replaced Obviously Kuhn waswell aware of such facts and did not want to be seen as disregarding or go-ing against them So by shifting emphasis on the appraisal of change of be-lief he modiordfed his model to accommodate them In that shape the radi-cal philosophical implications of his work are integrated in a moreplausible account and can become more easily acceptable The problemhowever is that as I said before Kuhn in this retrospective reappraisaldoes not really make room for history in his model and also takes awaysome of the bite that it had What is more he conordmates two levels of dis-cussion

On the historical empirical level it is indeed true as many of Kuhnrsquoscritics have pointed out that there are several indications of continuity orreasoned debate across paradigmatic change15 It is often the same individ-uals that change allegiance and certainly one cannot say that these indi-viduals do not understand their previous self (as the incommensurabilitythesis has been taken to imply) As Kuhn himself admits ldquocommunication


15 See for instance what Abner Shimony says in Klein Shimony and Pinch (1979p 436) ldquoOn the whole the intellectual processes of the few physicists immersed in black-body research seems to me to have been wonderfully rationalrdquo Toulmin (1972 pp 103ndash105) makes a similar point in relation to both the Copernican revolution and the transitionfrom Newtonian to Einsteinian physics Daniel Garber (2001) on the other hand whileinsisting that adherents of competing paradigms remain intelligible to each other claimsthat rational argumentation breaks down when transition is under way But in disagree-ment with Kuhn he does not attribute this to incommensurability and conceptual gapsbetween the rival paradigms but to more general cultural factors

goes on however imperfectly metaphor serving as a partial bridge acrossthe divide between an old literal usage and a new one To speak as I re-peatedly have of a communityrsquos undergoing a gestalt switch is to com-press an extended process of change into an instant leaving no room forthe microprocesses by which the change is achievedrdquo (Kuhn 2000a p 88)Also many beliefs stay intact and are not revised after a revolution Infact the closer one studies the historical circumstances of a period thesmaller the changes will seem the more continuity will be found Evenconceptually the very notion of change requires something that remainsunchangeable be that the scientists themselves the scientiordfc communitythe scientiordfc practice the scientiordfc enterprise at large If there is no un-derlying identity one would not speak of change but of substitution ofone alien entity by another In that case we would not even speak of thehistory of some entity eg of a discipline However radical the change af-ter a revolution be it political or scientiordfc there need to be found somecontinuity something that remains stable in order to speak of change andrevolution in the ordfrst place

On the philosophical level however where the philosophers work atsome distance from what empirically goes on the issue of continuity inscientiordfc development is raised differently The continuity of which tradi-tional philosophers spoke and traditional historians of science recordedwas not an empirical ordfnding16 It was a consequent of the ideal image ofscience which presupposed a common atemporal method for the sciencesin all times and places It was also a demand of the particular theory ofmeaning incorporated in the ideal image This image required that mean-ing seeps through from the level of experience up to the more complextheoretical abstractions If now the only root of meaning is observationcaptured in intersubjectively avowed protocol sentences and carried for-ward by correspondence rules then continuity in science is warranted bysameness of meaning If any theoretical construct is latched on to theworld by neutral observation statements in order to acquire meaning thenthere is always a common core whatever the change Continuity in theminds of the philosophers concerned propositions and words and hadnothing to do with shared elements of the actual scientiordfc practice17


16 Historians of science who took their cue from the philosophers did not discover con-tinuity because they observed the facts from close proximity rather the opposite was thecase Their research and its results were shaped by the philosophical tradition (see Kuhn2000c p 111)

17 It shouldnrsquot be forgotten that the term ldquosciencerdquo for the philosophers who mouldedthe ldquoreceived viewrdquo signiordfed solely scientiordfc theories which were taken to be systems ofpropositions The so-called ldquoexternal factorsrdquo of science were not supposed to be of interestto philosophy

When now Kuhnrsquos critics and Kuhn himself indicate and acknowl-edge historical and sociological evidence of continuity drawn from the ac-tual practice of scientists they concede to the pressure still exercised bythe traditional ideal image of science They are trying to preserve or sal-vage some non-Archimedean but nonetheless shared terra ordfrma In myview though the facts that are cited and the points made by both sidesare not directly relevant to the philosophical issue addressed Kuhnrsquos tar-get was the philosophersrsquo image of science which rested on a particulartheory of meaning Once this theory is challenged the radical implica-tions of the critique (discontinuity incommensurability) follow no matterhow much evidence of continuity the historians and sociologists accumu-late

I am not saying that the philosophersrsquo account ought to be independentof or without regard to the facts I am saying that the supposition of conti-nuity and of a common scientiordfc method was not derived from facts butwas rather a postulate a philosophical requirement imposed upon themWhen historical research was freed from the streamlining that this monis-tic methodology imposed it illustrated how diverse the practices of scien-tists have been It has then provided philosophers like Kuhn with a reasonto challenge the ubiquitous presence and validity of the standard image Itis not that Kuhn inferred from facts incommensurability and radicalbreaks Rather his historical research helped him see the philosophicalcharacter and presuppositions of the previous historical works

Kuhnrsquos claims about discontinuity are not theoretical constructionsbased on empirical evidence but follow from seeing the scientiordfc enter-prise differently from the received view ie as a developmental processand not as an axiomatically arranged set of sentences When his critics citeall the evidence of continuity and communication they ordfght I believe astraw man As stated before Kuhn could not have been unaware of themundane facts of reasoned exchange between scientists advancing andsupporting different theories and hypotheses When he talked of incom-mensurability and resort to persuasion when he used the metaphors ofconversion and living in different worlds he could not have implied thatif historians looked at revolutionary periods they would ordfnd scientists in-dulging in typically ldquoirrationalrdquo behavior like for instance preaching orengaging in idiosyncratic and irrelevant monologues scientists would stillbe seen addressing the scientiordfc community in the typical manner ie bypresenting arguments and articulating explanations Nor would have heimplied that the evidence of difordfculties in communication would neces-sarily take the form of explicit remarks stating mutual incomprehensionand unintelligibility His point rather I take it was that historians andphilosophers taking for granted the standard image of science tend to


overlook or do not attentively look for nuanced changes in use and mean-ing of terms that may indicate that divergent directions of developmentare under way18

On the mundane non-philosophical level then scientists engage in thecivil and reasoned behavior we normally associate with them And they docommunicate because they share the wider practice of science which pro-vides the conditions for making sense of each other On the philosophicallevel however where the speciordfc articulation of how terms mean and howconcepts function matters Kuhnrsquos contention is that once the standardimage is challenged there are no more common meanings secured bycommon observations Then continuity and communication are threat-ened And once the changes become deep lack of communication followsif one remains within the philosophical perspective Kuhnrsquos pronounce-ments that pertain to incommensurability are provocative and radical onlyagainst the background of the received view They do not question the ra-tionality of ordinary affairs

Admittedly Kuhn is not always clear or consistent as to what exactlyhe is doing whether that is he is advancing empirical or philosophicalobservations But I am offering an interpretation that I believe explainshow Kuhnrsquos claims are radical on the one hand (on the philosophical level)and yet close to the scientistsrsquo experience on the other Kuhn never wantedto be seen as advocating theses that would distance him from the scientiordfccommunity and the view that science is a rational enterprise

VI In what sense is Kuhnrsquos model necessary and a prioriIn the previous section I have presented Kuhnrsquos later view that he can de-rive his model from first principles with history entering only by aglance I also contested the relevance of offering empirical evidence tocounter Kuhnrsquos claims about science and I argued that his project is philo-sophical But I have not still explained what kind of a philosophical proj-ect that is I believe that such an explanation is needed if Kuhnrsquos accountespecially the later one is not to be dismissed as speculative armchair phi-losophy It needs to be shown how Kuhnrsquos model is not completely arbi-trary and how despite its derivation from first principles it leaves someroom for history

I contend that his project is a transcendental one offering the condi-tions of possibility of science He maintains that science as a practice de-


18 ldquoI donrsquot think that the people who were doing history by and large saw everythingin it that I was seeing in it They were not coming back asking lsquoWhat does this do to thenotion of truth what does it do to the notion of progressrsquo or if they did they were ordfndingit too easy to ordfnd answers that seemed to me superordfcialrdquo (Kuhn 2000d pp 311ndash12)

pends logically on following rules which in the case of science are set byconcrete exemplars of scientific achievement This practice has a dogmaticcharacter and establishes the normalcy and normativity required The restof Kuhnrsquos model is shaped accordingly Science develops by solving thepuzzles provided by the exemplars and paradigms Emerging anomalies(ie deviations from normalcy) if they are to be made ldquolawfulrdquo may forcea change of rules in which case crises and revolutions follow If nowmeaning is given by the rules of normal science then when revolutionsoccur there is change of meaning which may be quite radical yielding in-commensurable results Revolution is not inferred from historical observa-tion but introduced as a concept to account for radical change19

Admittedly transcendental arguments mainly because of Kant havebeen associated with concepts more fundamental than science-like experi-ence or knowledge- and they have been taken to serve a particular anti-skeptical purpose They are construed as showing that skeptical doubtsthemselves or a claim a skeptic would not challenge (for instance that wehave experience) presuppose the truths about the world that the skeptictakes to be questionable I do not claim that Kuhn is engaged in a similarproject He is not mounting an anti-skeptical rebuttal When I say thathis analysis is transcendental I am trying to appropriate two features dis-tinctive of transcendental arguments which are also of interest to Kuhnnecessity and a priority

Transcendental arguments bring out necessary connections betweenconcepts Typically they have the form ldquoone thing (X) is a necessary con-dition for the possibility of something else (Y) so that (it is said) the lattercannot obtain without the formerrdquo (Stern 1999 p 3) Stroud has arguedthat such a description ldquowould make all valid deductions transcendentalrdquo(Stroud 1999 p 158) and suggested that what is peculiar about transcen-dental arguments is both their beginnings and their end They start frompsychological facts (that we think and experience in certain ways) and theyaim to prove the necessity of some non-psychological facts (how things inthe world must be)20 The latter are shown to be the conditions of possi-bility of the former Stroud (1968) has questioned however whether this


19 That is why it is immaterial to wonder as some critics do how many revolutionsthere are in the history of science Kuhn did not count paradigms or revolutions he didnot say how grand or how sweeping they might be

20 Quassim Cassam (1999) spoke also of ldquoself-directed transcendental argumentsrdquo asopposed to the so-called ldquoworld directedrdquo The ldquoworld directedrdquo argue from how thought isto how the world must be The ldquoself-directedrdquo which were employed by Kant in the Prole-gomena to Any Future Metaphysics in the context of his ldquoanalyticrdquo method do not address theskeptic and argue from certain cognitive achievements to how the cognitive faculties of theknowing subject are

project can ever succeed without the baggage of idealism and in his laterwritings (Stroud 1999) sided with Strawson who accepting Stroudrsquos criti-cism construed transcendental arguments as tracing connections withinour conceptual scheme and not across the bridge from how we think tohow the world must be21

Kuhn can be seen as undertaking a similar analysis Although he doesnot address the skeptic although he does not start from indispensable psy-chological facts we can say by stretching the use of transcendentalmethod that Kuhn is interested in investigating the connections betweenthe concepts that comprise the phenomenon we call science These con-nections are not to be construed as simply causal but as logical and in thatsense necessary and a priori22 His whole model can then be taken as an ar-ticulation of these kinds of connections

But is Kuhnrsquos model just a set of interdefined terms and the connec-tions between them merely analytic necessities This possibility cannot beruled out even in the case of the Kantian arguments23 Nor can it be ruledout in the case of Kuhn despite the fact that it does not seem contradictoryto say for instance that science does not rely on dogma But even if thenecessities that Kuhnrsquos model indicates are analytic it can be maintainedthat there is still something to be gained by employing the transcendentalstrategy As Stroud observes (2000 p 233) even in case the necessities are


21 Strawson (1985 pordm23) embracing the perspective of the naturalist philosopheropts for a weakened version of transcendental arguments in the context of his ldquodescriptivemetaphysicsrdquo He claims that transcendental arguments could be used to investigate theconnections between ldquothe major structural elements of our conceptual scheme-to exhibitit not as a rigidly deductive system but as a coherent whole whose parts are mutually sup-portive and mutually dependent interlocking in an intelligible wayrdquo

22 Kuhn does not seem to distinguish between ldquonecessaryrdquo (a metaphysical concept)and ldquoa priorirdquo (an epistemological concept) He does not expand on them but it seems thathe is following Kant in assuming that ldquoif we have a proposition which in being thought isthought as necessary it is an a priori judgmentrdquo (Kant 1933 B3) Kripke thought the twoterms are not coextensive and claimed that there exist necessary a posteriori and contingent apriori truths (1980 p 38)

23 Strawson according to Stroud advanced in The Bounds of Sense a particular interpre-tation of the Kantian transcendental strategy in order to avoid the threat ldquothat what Kantestablishes are at most lsquoanalyticrsquo or lsquodeordfnitionalrsquo necessitiesrdquo (Stroud 2000 p 234)Strawson maintained that the necessities sought by Kant were not those between conceptsor meanings but between conceptual and experiential capacities (Stroud 2000 p 235) Butsubsequently Stroud notes ldquoit is just possible I suppose to see even that enterprise asyielding at best only lsquoanalyticrsquo or lsquodeordfnitionalrsquo necessities They would express what is lsquoco-vertly containedrsquo not in the concept of say lsquoexperiencersquo or the concept of lsquosubject of expe-riencesrsquo but in the concept lsquopossesses the concept of experiencersquo or the concept lsquopossessesthe concept of a subject of experiencersquo lsquoAnalysisrsquo might reveal that the quite different con-cept lsquothinks of the world as containing objective particularsrsquo is contained in one or both ofthese conceptsrdquo (Stroud 2000 p 239)

construed as analytic it does not follow that the connections can be imme-diately seen by simply ldquogazing insiderdquo the concepts It is an operationmuch more complex and rewarding than that

Still I would say that Kuhn aimed for different necessities not analyticbut synthetic and without the Kantian idealist baggage By giving theconditions of the possibility of science Kuhn was not trafficking simplyamong concepts nor was he engaged in linguistic analysis He was sayingsomething about the world something that he had picked up from his ex-perience as a scientist and a historian namely that we cannot have scienceas we know it unless we have dogmatic training paradigms and rulesScience would not be possible if any of these things did not obtainNow if this is how we construe the Kuhnian project we preserve some atleast of its transcendental character but then Kuhn has to face the earlyStroudian challenge to transcendental arguments ie that their validityhangs on verification So if the necessities are intended as synthetic a pri-ori the proof is in verification if the necessities are analytic one is con-fined within a particular conceptual scheme24

The first option that of construing the Kuhnian necessities regardingscience as synthetic a priori would bring back the problems related to em-pirical confirmation that Kuhn was moving away from As for the otherhorn of the dilemma I believe that Kuhn would have welcomed thehistorization of what science presupposes and requires25 As he had wel-


24 Strawson (1985 p 26) talking about the Kantian project feared that if the tran-scendental investigation is conordfned within a particular conceptual scheme it would behistoricized and consequently weakened and relativized It would mean that metaphysicswould become an essentially historical study much in the spirit of Collingwood whosought the ldquoabsolute presuppositionsrdquo of each historical epoch He also stressed that ldquoif westick to the actual behavior of words then what we will discover will not be sufordfcientlygeneral or sufordfciently far reaching to satisfy our urge for full metaphysical understandingrdquo(Strawson 1970 p 319) Stroud also observed that ldquothe lsquohistoricalrsquo conception of meta-physics could endorse what Kant would call the lsquoanalyticrsquo character of the necessities it dis-coversrdquo (Stroud 2000 p 233)

25 Fuller is of a different view He claims (2000 pp 73 195 215) that Kuhn has of-fered a description of a mythical image of science which he created by disregarding thegreat differences in scientiordfc practice in the course of history Kuhn did that according toFuller in his effort to carry out a Platonic mission which called for the legitimation of thecontemporary scientiordfc-industrial-military status quo Fuller thinks that a transcendentalstrategy lends support to an ahistorical understanding of science ldquothe transcendental argu-ment [ ] attempts to convert an impoverished imaginationmdashspeciordfcally our inabilityto envisage what the world would be like if progress and rationality turned out to be com-plete mythsmdashinto a guarantee that our faith in these myths is well-placedrdquo (Fuller 2000p 29) Strawson had already made the point ldquoThe transcendental arguer is always exposedto the charge that even if he cannot conceive of alternative ways in which conditions of thepossibility of a certain kind of experience or exercise of conceptual capacity might befulordflled this inability may simply be due to lack of imagination on his partmdasha lack whichmakes him prone to mistake sufordfcient for necessary conditionsrdquo (Strawson 1985 p 23)

comed the historization of scientific knowledge claiming that he is aKantian with moveable categories (Kuhn 2000d p 264 also Kuhn2000b p 104) he would not I think object to the possibility of imagin-ing a differently shaped practice of science And if he is credited with sucha view his model would give us then the historicized conditions of thepossibility of science But then again the problem that his model carriesonly analytic necessities re-emerges The worry is that the sentences stat-ing these connections would be true solely in virtue of the meanings of theterms involved which would imply that they are simply a matter of stipu-lation

This fear however can be alleviated It can be maintained that analyticnecessities are not detached completely from the world Quassim Cassam(2000 p 60) in a paper where he reconsiders the distinction between em-piricism and rationalism as regards a priori knowledge refers to QuineBoghossian and Peacocke and defends the view that ldquono sentence is truebut reality makes it so [ ] It makes no sense to suppose that linguisticmeaning alone can generate truthrdquo Even analytically necessary proposi-tions Cassam says have factual content and they are true by virtue of howthe world is26 Then following a reading of Lockersquos Essay by James Tully27

Cassam proceeds to distinguish between natural and conventional reality inorder to explain how analytic necessities can be both conventional andldquoabout the worldrdquo They respond to reality but a reality that has been setup They are answerable to social facts (Cassam 2000 p 59)28 This doesnot make them merely arbitrary In a similar vein Wittgenstein has thisto say regarding the ldquoarbritrarinessrdquo of our color system a system that pre-cludes as a matter of conceptual truth the existence of a color intermediatebetween red and green

We have a color system as we have a number system

Do the systems reside in our nature or in the nature of things Howare we to put itmdashNot in the nature of numbers or colors (Z 357)


26 According to Cassam (2000 p 60) who cites Boghossian the difference with syn-thetic statements lies in that in the case of analytic statements grasping the meanings ofthe terms involved sufordfces for justiordfed belief in their truth

27 According to Tullyrsquos (1980) reading of Locke our ideas of substances are inadequatebecause they do not represent natural substances accurately whereas our ideas of modes areadequate since they (our ideas of modes) do not copy but deordfne their objects (Cassam2000 pp 58ndash9) It follows that the necessity of analytic propositions involving adequateideas of modes becomes conventional

28 John Searle (1995) distinguishes between institutional or social facts on the onehand and ldquobruterdquo facts on the other and argues that the former are no less real than the lat-ter

Then is there something arbitrary about this system Yes and noIt is akin both to what is arbitrary and to what is non-arbitrary(Z 358)

According to Wittgenstein our systems with their conceptual and logicaltruths may owe a lot to what we contribute but they are not whimsicalconstructions of the mind29 They are constituted and constrained by anetwork of practices and by the concrete and very real implications thatthese practices have Cassam commenting also on the necessity of theproposition ldquoNothing can be red and green all over at the same timerdquonotes that ldquothere does not seem to be a straightforward answer to the ques-tion whether it is nature or convention to which [this proposition] owesits necessary truthrdquo (Cassam 2000 p 59) He then approvingly quotesD F Pears ldquoperhaps the emphasis on either side is a mistake perhaps theculprit is neither convention alone nor nature alonerdquo (ibid)

The point of the above digression into Wittgenstein and Cassam is toindicate that even if the necessities one deals with are analytic it can bemaintained that they are accountable to reality and not merely arbitraryReturning to Kuhn I would claim that he undertakes a transcendentalanalysis of science and establishes a priori necessities which even if re-garded as analytic are not devoid of factual content and are not capri-ciously stipulated An obvious objection however to this line of thoughtwould be the following if the only necessities Kuhn establishes are ana-lytic irrespectively of how they are interpreted why not call what he doesconceptual rather than transcendental analysis I prefer transcendental toconceptual analysis for the following reasons (a) Conceptual analysis isusually taken to be a merely conventional linguistic analysis MichaelDevitt and Kim Sterelny (1999 pp 282-3) for instance write ldquoSince allthe concepts that the analysts are interested in are ones for which we havewords (on the language-of-thought hypothesis the concepts are mentalwords synonymous with the public words that express them) theirmethod is hardly distinguishable from the ordinary language philoso-phersrsquo investigation of the use and misuse of wordsrdquo Devitt and Sterelnycall the method of conceptual analysis ldquoarmchair thought experimentrdquo andthey quote G J Warnock who says that philosophy as conceptual analysisis ldquothe study of the concepts that we employ and not of the facts phenom-ena cases or events to which those concepts might be or are appliedrdquo(Devitt and Sterelny 1999 p 282) I do not think that Kuhn would en-dorse this kind of philosophy He was very much interested in the factsand cases of science and could not settle for armchair reflection As I will


29 For the sense in which grammar is arbitrary and non-arbitrary see Forster (2004)

argue his historical studies were very significant in shaping his philo-sophical position (b) Conceptual analysis is all too often understood as adefinitional inquiry into the essence of things very much like the activityof the ancient Greeks who sought to define beauty justice knowledgevirtue etc Kuhn however was not an essentialist philosopher He wasvery much opposed to trying to define science independently of time andspace (c) I believe that transcendental analysis unlike conceptual analysiswhich oscillates between essentialism and conventionalism fits betterKuhnrsquos project30 It gives the conditions of possibility of science (iedogmatic training and exposure to exemplars)31 it accounts for the neces-sity and a priority that Kuhn bestows upon his model and finally it is inline with Kuhnrsquos own understanding of himself as a Kantian with move-able categories

VII A role for historyIf now Kuhnrsquos project is taken to be transcendental what role does his-tory of science play in it How can something empirical be accommodatedin an a priori investigation One thing to consider is the following pointmade by Kripke ldquoSomething may belong in the realm of such statementsthat can be known a priori but still may be known by particular people onthe basis of experience [ ] So lsquocan be known a priorilsquo doesnrsquot meanlsquomust be known a priorirsquordquo (Kripke 1980 p 35 emphasis in the original)Kuhn himself is a case in point As he said he was led as a matter of his-torical fact to the formulation of his model on the basis of his studies inthe history of science but he came to realize that he could have derived itfrom first principles (Kuhn 2000c pp 111-2) But is then the contribu-tion of history dispensable Was it only an accident that historical researchled to the articulation of Kuhnrsquos account of science

Kuhn has said that what emerged as essential for him was the perspec-tive or the ideology that attention to historical cases brings with it (Kuhn2000b p 95) This perspective offered Kuhn a dynamic conception of sci-ence-as opposed to the static one that dominated the so-called receivedview-which yielded in turn the rest of his model So what Kuhn says is


30 Robert Hanna in the entry ldquoConceptual analysisrdquo in the The Routledge Encyclopedia ofPhilosophy (1998) comparing transcendental arguments to conceptual analysis capturesnicely what is distinctive in Kuhnrsquos work ldquoTranscendental arguments extend the scope ofconceptual analysis from the mere deordfnitional or logical exploration of conceptual contents(something also called lsquophilosophical grammarrsquo) towards insights into ordfrst principles ex-pressing the lsquoconceptual geographyrsquo of the common sense worldrdquo

31 Science is Y and dogmatic training and exemplars are X in Sternrsquos description oftranscendental arguments (ldquoone thing (X) is a necessary condition for the possibility ofsomething else (Y) so that (it is said) the latter cannot obtain without the formerrdquo)

that his a priori investigation begins after the historical perspective is as-sumed In that sense the contribution of history seems only to be prepara-tory It gets us acquainted with the concepts and facts involved in theanalysis that will follow But what concept of science does Kuhnrsquos histori-cal research bequeath to us what kind of facts does it report

I believe that Kuhnrsquos historical studies offer much more than just a dy-namic conception of science One would get such a conception by justconsidering that science is not merely a set of atemporal propositions com-prising scientific theories but a social activity that develops over timeOne would not even need to conduct historical research to get such anidea Once science was perceived as undergoing change in time the dy-namic conception would immediately become available But this is notenough to yield the Kuhnian model The dynamic conception by itself canaccount for several of the concerns that Kuhn had in his later writings(evaluating change of belief instead of belief tout court preserving ele-ments of continuity despite change) but not for all It does not entail rad-ical differences across revolutions it is compatible with cumulativegrowth of scientific knowledge

Kuhn (2000b p 98) drew parallels between biological and cognitiveprocesses to account for revolutionary change He said that speciation inbiological evolution is the analogue of incommensurability in the evolu-tion of knowledge But this move is simply analogical It comes post festumto illustrate a point already assumed Where then does Kuhn get the ideaof discontinuity in the development of science I believe that I have shownin section VI above that Kuhn has mainly philosophical arguments to de-fend this thesis But how does history fit in

Kuhn described a role for history in the very first sentence of his bookHe summarized his project succinctly

History if viewed as a repository for more than anecdote or chro-nology could produce a decisive transformation of the image of sci-ence by which we are now possessed (Kuhn 1970 p 1)

In this sentence Kuhn contends that we are possessed by an image of sci-ence-which he also calls ideal-and suggests a use of history to transform itHe rejects both anecdotal history ie history that narrates the strikingachievements of science and the compilation of chronicles that assigncredit for inventions and discoveries Both these types of history are in theservice of the ideal image They take for granted the view of science thatthis image provides and they exemplify and enhance it Kuhnrsquos alternativeuse of history aims to combat the ideal image But how does it do it if itdoes not provide evidence to refute a hypothesis

I believe that one can get a better idea of Kuhnrsquos use of the history of


science if one compares it to what Wittgenstein did in relation to lan-guage Wittgenstein in his Philosophical Investigations ([1951] 2000) at-tacked an essentialist idea of meaning by bringing forward the multiplic-ity of ways language is and can be used He cited real and fictitiousexamples of application in order to show that meaning is not an entity at-tached to words but a matter of practice that may vary widely from lan-guage game to language game He said that he assembled reminders tobring into relief the great diversity that characterizes the employment oflanguage Wittgensteinrsquos reference to particular examples of language useis not a recourse to empirical facts in an effort to ground or refute philo-sophical pronouncements Wittgenstein does not say ldquoThis is how we or-dinarily use language So philosophical usage is wrongrdquo Had this argu-ment been successful it would have discredited not only philosophicalusage but any non-ordinary use of language (eg scientific or literary)and what is more any new application of words32 Wittgenstein attacks aphilosophical idea of meaning by questioning its presuppositions whichconcentrate on our ldquocraving for generalityrdquo on our abstracting from theconcrete case He unmasks this strategy of assimilation by considering themultiplicity of language use Yet he is not interested in the actuality ofthe cases he gives as examples nor in how probable they are His investiga-tion ldquois directed [ ] towards the possibility of phenomenardquo (PI 90) Hesays ldquoIrsquom just citing what is possible and am therefore giving grammati-cal examplesrdquo (Wittgenstein 1993 p 187)

Similarly I would say Kuhn attacks an essentialist idea of science Hedoes it philosophically but he also summons concrete examples from thehistory of science to illustrate the different routes science has taken Heeven constructs ldquofictiverdquo examples as R Westman has pointed out (seeSection IV above) His aim is to loosen the grip of the ideal image whichinsisted on an assumed uniformity of scientific method and possessed usjust like Wittgenstein was aiming to loosen the grip of the picture of lan-guage which insisted on an assumed general form of proposition and heldus captive (PI 115) Wittgenstein reminds us of facts of our natural history(PI 415) Kuhn reminds us of facts of ldquoscientific historyrdquo The former areopen to view to whoever succeeds in overcoming their familiarity InKuhnrsquos case facts had to be dug out they had to be freed from thestraightjacket that had put them in the service of the official image of sci-ence Once they were out they functioned as objects of comparison Justlike language games ldquoare meant to throw light on the facts of our lan-guage by way not only of similarities but also of dissimilaritiesrdquo (PI 130)Kuhnrsquos historical cases are supposed to show (not in the sense of prove)


32 I have discussed this more extensively in Kindi (1998)

how varied things have been and can be in the future The different scien-tific traditions that Kuhn describes are not presented as instances of an al-ternative a priori idea of what science is nor are they collected to be usedas (feeble) evidence in support of a new theory about science Put side byside they map out a mosaic of possibilities Instead of the monolithic uni-formity of the ideal image Kuhn draws our attention to the diversity ofpractices set up by different exemplars and rules The result of this processis an ldquoopen conceptrdquo of science characterized not by delimiting necessaryand sufficient conditions but by a complicated network of similarities anddissimilarities33

VIII Is Kuhnrsquos historiographical work consistent with thepresent readingCertainly an overall assessment of Kuhnrsquos historiographical work in rela-tion to the present reading falls outside the scope of the present paperwhich is concerned to account for the relation between history and philos-ophy of science in the SSR and Kuhnrsquos later philosophical writings Yet Iwould like to consider brieordmy if there is any prima-facie evidence inKuhnrsquos historical projects that would go against the reading advanced Itwould indeed be quite strangemdashlet alone unaccommodating for thepresent interpretationmdashif Kuhn did and claimed he did one thing whenusing history in his philosophy and went in the opposite direction whenengaged in historiography

What would this opposite direction be If Kuhnmdashindependently of hisrepeated explicit denialsmdashdid in fact when doing history use it to deriveground or illustrate his philosophical model then the reading outlinedabove would have very limited and disputable value It would apply pos-sibly only to Kuhnrsquos philosophical writings while Kuhn himself wouldappear torn between the two ordfelds But the proposed reading would beenhanced if Kuhn in his historical work did not engage in the effort ofusing history to construct and justify his philosophical model In that re-spect I offer the following considerations First in Kuhnrsquos majorhistoriographical writings one early before the StructuremdashThe CopernicanRevolutionmdashand one late after the StructuremdashBlack-Body Theory and Quan-tum DiscontinuitymdashStructurersquos concepts are conspicuously absent which isan indication that Kuhnrsquos historical accounts were carried out independ-ently of his philosophical concerns He says so himself ldquoI do my best for


33 Kuhn says that the aim of the Structure is ldquoa sketch of the quite different concept ofscience that can emerge from the historical record of the research activity itselfrdquo (Kuhn1970 p 1) Compare what Wittgenstein says about ldquofamily resemblancerdquo terms likeldquogamesrdquo (PI 65ndash71)

urgent reasons not to think in these [philosophical] terms when I do his-tory and I avoid the corresponding vocabulary when presenting my re-sultsrdquo (Kuhn 1987 p 363) Commentators also point to this fact AbnerShimony writes about the Black-Body book ldquoKuhn deserves credit for ana-lyzing the discovery of quantum discontinuity without any apparent con-trol by the conceptions of scientiordfc change expressed in [the SSR] The in-terpretations which he offers in his new history whatever their strengthsand weaknesses seem to be the product of immersion in the scientiordfctexts and of reordmection upon scientiordfc problemsrdquo (Klein Shimony andPinch 1979 p 437) The same claim is made by Trevor Pinch

It is as though The Structure of Scientiordfc Revolutions had never beenwritten The familiar notions of paradigm normal science exem-plars problem solving anomalies crises extraordinary science andincommensurability are nowhere to be found In short virtually allthe terminology associated with Kuhnrsquos earlier work has beenpurged from the present one Most of the concepts have vanished aswell or else remain in such a heavily veiled form that their sig-niordfcance is unclear (Klein Shimony and Pinch 1979 p 438)

In The Copernican Revolution Kuhn uses terms such as conceptual scheme con-ceptual disparities (Kuhn 1957 p 229) alternative cosmologies which do notappear in the Structure but which admittedly can be compared to Struc-turersquos paradigm incommensurability of concepts and talk of different worlds re-spectively However despite these analogies34 one can ordfnd claims in thesame book that do not conform to Kuhnrsquos philosophical schema Kuhnwrites for instance in The Copernican Revolution ldquoEach new theory preservesa hard core of the knowledge provided by its predecessor and adds to itrdquo(Kuhn 1957 p 3 emphasis added) This is very different from what issaid in the Structure Robert Westman also notes ldquoKuhn produces no con-version storiesrdquo in The Copernican Revolution (1994 p 95) and he addsldquoKuhn had shifted to a more radical notion of revolution [by the time hepublished SSR] [ ] CRrsquos overall image of science by contrast resem-bles the positivist and pragmatist conventionalism of Henri BergsonHenri Poincareacute and Ernst Mach and especially some of the moreconventionalist passages in Pierre Duhemrsquos The Aim and Structure of Physi-cal Theoryrdquo (Westman 1994 p 84) Trevor Pinch makes a similar claimabout incongruities between the Structure and Black-Body Theory (Klein


34 Another similarity between the Structure and The Copernican Revolution lies in theconsideration of the wider social and intellectual context and of course in the use of theterm ldquorevolutionrdquo As regards the latter however Toulmin insists that Kuhn used the termdifferently in the two texts (1972 pp 107ndash12)

Shimony and Pinch 1979 p 439) ldquoThere is also a discernible shift of em-phasis in Kuhnrsquos current work [Black-Body Theory] Science is portrayed asa process much less susceptible to human or even social inordmuence nature isordfrmly in the driverrsquos seatrdquo Finally Martin Klein in his essay on the samebook again notes that Kuhnrsquos historical work cannot be used to clarify hisphilosophical categories

Since the creation of quantum physics would be called a scientificrevolution on any account of that term we might well anticipatethat Kuhnrsquos new book [Black-Body Theory and Quantum Discontinu-ity] would be a kind of scholarly descendant of Karl Marxrsquos Eigh-teenth Brumaire of Louis Bonaparte This would be a book in whichthe theorist of revolutionary change analyzes a particular revolutionin terms of his general categories using them to improve our his-torical understanding of the events in question and at the sametime clarifying their meanings But Kuhn has not written anotherEighteenth Brumaire (Klein Shimony and Pinch 1979 p 430)

The second point to consider then is that Kuhnrsquos historical work farfrom being the ground or the illustration of his philosophical contentionsincludes claims that can and have been interpreted as even going againstclaims in the Structure

Kuhn himself did not ordfnd the two accounts (historical and philosophi-cal) as divergent as the above critics have taken them to be ldquoOften I donot know for some time after my historical work is completed the respectsin which it does and does not ordft Structure Nevertheless when I do lookback I have generally been well satisordfed by the extent to which my narra-tive ordft the developmental schema that Structure providesrdquo (Kuhn 1987p 363) In this remark Kuhn restates that he conducts his historical andphilosophical research independently of each other but recognizes thatretrospectively his historical narrative turns out not to have underminedormdashmore positivelymdasheven to be in general accord with the developmen-tal schema he proposed He further claims that speciordfc episodes in the his-tory he recounted may be taken again retrospectively to illustrate con-cepts of the Structure35 If we bracket the thought that historical data canbe made to conform to any philosophical model including obviouslyKuhnrsquos own how can we explain the ordft that Kuhn sees Is it a happy coin-cidence or did he ordfddle with the facts This is an important issue to con-sider because if Kuhn induced the harmony between the two accounts he


35 According to Kuhn Boltzmannrsquos probabilistic derivation of the entropy of a gas isan illustration of the concept of paradigm in the sense of concrete model (Kuhn 1987p 363)

did not really work independently which means that he may have used hishistorical research to produce and support his philosophical model And ifthis is correct then the reading I have proposed is not borne out at leastby Kuhnrsquos historical work

I think Kuhn was right (and sincere) when he said that he did notamalgamate the two tasks I have offered evidence above (that the vocabu-lary of the Structure is missing in Kuhnrsquos historical work and that there areincongruities between the two domains) which gives us reason to believethat Kuhn worked independently in the two ordfelds The same evidence in-dicates that if there is a match it is not a perfect one Kuhn himself wascautious enough to write that he has only ldquogenerally been well satisordfed bythe extent to which [his] narrative ordft the developmental schema that Struc-ture providesrdquo (emphasis added) Still how did this general alignmentcome about

The reading I have proposed offers a perspective in which the overallagreement between Kuhnrsquos historical and philosophical accounts can beexplained but not as a matter of happy coincidence or as a match betweenempirical data and theory Kuhnrsquos accounts run more or less parallel toeach other because in both Kuhn entertained similar concerns Even whendoing history his aim was to tear apart the seamless image of sciencewhich when cast on facts mufordmed all diversity and difference He aimedat ridding the minds of historians from this distorting mould in order toadvance a different understanding of science ldquo[H]istorical study couldyield a new sort of understanding of the structure and function ofscientiordfc researchrdquo he wrote in The Copernican Revolution (1957 p ix)36

while in his ldquoldquoAfterword Revisiting Planckrdquo appended to the Black-BodyTheory he claims that history ldquocan inordmuence views about the nature ofknowledge and about the procedures to be employed in its pursuitrdquo (1987p 370) His objective in writing the Black-Body book was he explicitlysaid later ldquoa fundamental reinterpretation of Planckrsquos thought and of thestages in its gradual transformationrdquo (1987 p 349) He reinterpreted thestandard account of this period which ldquo[matched closely] a still cherishedview of the nature of science and its developmentrdquo (1987 p 370) Evenwhen Kuhn writes that this cherished view ought to be recognized as amyth (ibid) I do not think that his remark ought to be taken in the strictsense that he has proven this view false in the manner of disconordfrming atheory by data Rather given his other just cited expressed comments


36 Compare also Kuhn (1957 p 4) ldquoHistorical analysis may not answer questionslike these [What is a scientiordfc theory On what should it be based on to command our re-spect What is its function its use What is its staying power] but it can illuminate themand give them meaningrdquo

(eg that history can inordmuence our views) I take it that the alternative nar-rative that he offered showed how facts are when freed from the burden ofthe ideal image ldquo[M]y primary object was just to get the facts straightrdquo(ibid) This is an aim internal to historymdashit is constitutive of the disci-pline itselfmdashwhich can simultaneously serve Kuhnrsquos philosophicalpurpose37 of pointing to the diversity of scientiordfc development Gettingthe facts straight calls attention to the speciordfcs of the cases under exami-nation and requires sensitivity for detail It thus highlights complexityand difference instead of assimilation and conformity to philosophicalschemes

ldquoI was enough of a historian to know that the agreement did not existamong the people who were [concerned]rdquo (Kuhn 2000d p 296) Kuhnmakes this remark in his autobiographical interview trying to explain howthe concept of paradigm as model emerged when he was writing Structureand in particular the chapter on normal science The received view ap-proach to scientiordfc theoriesmdashwith which he says he was workingmdashrequired agreement among scientists which would appear in the axiom-atization of the theories in the form of axioms or deordfnitions (ibid)Knowing as a historian that this vast agreement wasnrsquot there he came upwith the concept of paradigm as model Here one might take it thatKuhn is relying on historical data to draft his philosophical categoriesBut that would be I think a mistake In the same interview he mentionsanother source that might have inordmuenced him to get to the idea of para-digm namely Polanyirsquos tacit knowledge (ibid) It is again I think thehistorianrsquos sensitivity for difference that helped him together with otherconsiderations to shrug off a particular philosophical image that squeezeddiversity into rigid preconceived moulds

So is Kuhnrsquos historiographical work consistent with the reading ad-vanced in the present paper I think it is What are this readingrsquos relevantcontentions First that Kuhnrsquos philosophical model does not rely on his-tory I have shown this in the bigger part of the paper by concentrating onKuhnrsquos philosophical writings and mainly on the Structure In the presentsection I have offered evidence to show that Kuhnrsquos history is not used toground or illustrate his model This does not mean that Kuhnrsquos historicalwork bears no relevance to his philosophy it shares a goal with it namelythat of undermining the standard image of scientiordfc development This isachieved by adherence to historyrsquos constitutive principle ie ldquoGetting thefacts straightrdquo which by itself contributes to the proliferation of diversityand difference So the second relevant contention of the reading I put for-


37 Compare Kuhnrsquos remark that he was ldquo[a] physicist turned historian for philosophi-cal purposesrdquo (Kuhn 2000d pp 320ndash1)

ward namely that the role of history in Kuhnrsquos philosophical model isthat of illustrating diversity ordfnds implicit support here I am saying ldquoim-plicit supportrdquo because there is no explicit description and statement ofsuch a task My reading offers a way to incorporate historical research in anotherwise a priori model and Kuhnrsquos historical work not only does notpreclude such a reading but by being autonomous and concentrating onthe details of facts contributes to the achievement of Kuhnrsquos philosophicalgoal

ConclusionKuhnrsquos was a philosophical project aiming to undermine a particularessentialist image of science a project that he carried out with the help ofthe history of science His attack comes on two fronts on the logical and apriori and on the historical On the logical he gives an account of sciencethat focuses on the practice that sustains it (giving its conditions of possi-bility) and on the historical he lays emphasis on the differences ratherthan the similarities that historical record illustrates This is where theKantian and the Wittgensteinian aspects of Kuhnrsquos work merge A tran-scendental investigation into the conditions of possibility of science is metwith the observation of historical facts which once the mould cast by thestandard image is broken disclose a variegated landscape rather than astrictly and rigorously bounded lot Kuhnrsquos historical research is not sim-ply preparatory on the way to an a priori analysis but really vital in show-ing concretely how distorting the ideal image was It is true that Kuhncould have defended the thesis of diversity based on the claim that mean-ing is not given by some invariant extra-linguistic entity but by rules thatgive rise to concrete practice Since these rules may change the differencesbetween the corresponding practices may turn out to be radical Yet hadKuhn confined his investigation to this general and abstract level hewould not have succeeded in shaking our deep commitment to the viewthat science is an exception that it is not subject to radical change Even ifhis presentation of historical facts turns out to be wrong he has managedto show concretely that things could have been otherwise And that was aliberating move for both history and philosophy of science

ReferencesBird Alexander 2000 Thomas Kuhn Chesham Bucks Acumenmdashmdashmdash 2002 ldquoKuhnrsquos Wrong Turningrdquo Studies in History and Philosophy

of Science 33(3) 443ndash463Cassam Quassim 1999 ldquoSelf-Directed Transcendental Argumentsrdquo

Pp 155ndash172 in Transcendental Arguments Edited by Robert Stern Ox-ford Clarendon Press


mdashmdashmdash 2000 ldquoRationalism Empiricism and the A priorirdquo Pp 43ndash64in New Essays on the A priori Edited by Paul Boghossian and Christo-pher Peacocke Oxford Clarendon Press

Edwards Charles ed 1998 The Routledge Encyclopedia of Philosophy Lon-don Routledge

Feyerabend Paul 1970 ldquoConsolations for the Specialistrdquo Pp 197ndash230 inCriticism and the Growth of Knowledge Edited by Imre Lakatos and AlanMusgrave Cambridge Cambridge University Press

Forster Michael N 2004 Wittgenstein on the Arbitrariness of GrammarPrinceton Princeton University Press

Friedman Michael 2001 Dynamics of Reason Stanford CSLI Publica-tions

Fuller Steven 2000 Thomas Kuhn A Philosophical History for Our TimesChicago The University of Chicago Press

Garber Daniel 2001 ldquoDescartes and the Scientiordfc Revolution SomeKuhnian Reordmectionsrdquo Perspectives on Science 9(4) 405ndash422

Hanson Norwood Russell 1965 ldquoA Note on Kuhnrsquos Methodrdquo Dialogue4 371ndash375

Hoyningen-Huene Paul 1995 ldquoTwo Letters of Paul Feyerabend toThomas S Kuhn on a Draft of The Structure of Scientiordfc RevolutionsrdquoStudies in History and Philosophy of Science 26(3) 353ndash387

Kant Immanuel 1933 Critique of Pure Reason Translated by N KempSmith London Macmillan

Kindi Vassiliki 1998 ldquoIs Wittgensteinrsquos Resort to Ordinary Language anAppeal to Empirical Factsrdquo Metaphilosophy 29(4) 298ndash305

Klein Marion J Abner Shimony and Trevor J Pinch 1979 ldquoParadigmLost A Review Symposiumrdquo Isis 70(3) 429ndash440

Kripke Saul 1980 Naming and Necessity Cambridge MA Harvard Uni-versity Press

Kuhn Thomas S 1957 The Copernican Revolution Cambridge Mass Har-vard University Press

mdashmdashmdash (1962) 1970 The Structure of Scientiordfc Revolutions Chicago TheUniversity of Chicago Press

mdashmdashmdash 1980 ldquoThe Halt and the Blind Philosophy and History ofSciencerdquo British Journal of the Philosophy of Science 31 181ndash192

mdashmdashmdash 1982 ldquoCommensurability Comparability CommunicabilityrdquoPhilosophy of Science Association 2 1ndash19

mdashmdashmdash (1978) 1987 Black- Body Theory and the Quantum Discontinuity1894ndash1912 Chicago The University of Chicago Press

mdashmdashmdash 2000a ldquoPossible Worlds in History of Sciencerdquo Pp 58ndash89 inThe Road since Structure Edited by James Conant and John HaugelandChicago The University of Chicago Press


mdashmdashmdash 2000b ldquoThe Road since Structurerdquo Pp 90ndash104 in The Road sinceStructure Edited by James Conant and John Haugeland Chicago TheUniversity of Chicago Press

mdashmdashmdash 2000c ldquoThe Trouble with the Historical Philosophy of SciencerdquoPp 105ndash120 in The Road since Structure Edited by James Conant andJohn Haugeland Chicago The University of Chicago Press

mdashmdashmdash 2000d ldquoA Discussion with Thomas Kuhnrdquo Pp 255ndash323 in TheRoad since Structure Edited by James Conant and John Haugeland Chi-cago The University of Chicago Press

Kourany Janet A 1979 ldquoThe Nonhistorical Basis of Kuhnrsquos Theory ofSciencerdquo Nature and System 1 46ndash59

Schefordmer Israel 1967 Science and Subjectivity Indianapolis Bobbs-MerillCompany

mdashmdashmdash 1972 ldquoVision and Revolution A Postscript on Kuhnrdquo Philosophyof Science Association 39 366ndash374

Searle John 1995 The Construction of Social Reality New York The FreePress

Shapere Dudley (1964) 1980 ldquoThe Structure of Scientiordfc RevolutionsrdquoPp 27ndash39 in Paradigms and Revolutions Edited by Gary Gutting NotreDame The University of Notre Dame Press

Sharrock Wes and Rupert Read 2002 Kuhn Philosopher of Scientiordfc Revo-lutions Cambridge Polity Press

Stern Robert ed 1999 Transcendental Arguments Oxford ClarendonPress

Strawson Peter F 1964 Individuals London Routledgemdashmdashmdash 1966 The Bounds of Sense London Methuenmdashmdashmdash 1970 ldquoAnalysis Science and Metaphysicsrdquo Pp 312ndash320 in The

Linguistic Turn Edited by Richard Rorty Chicago The University ofChicago Press

mdashmdashmdash 1985 Skepticism and Naturalism Some Varieties New York Co-lumbia University Press

Stroud Barry 1968 ldquoTranscendental Argumentsrdquo Journal of Philosophy65 241ndash256

mdashmdashmdash 1999 ldquoThe Goal of Transcendental Argumentsrdquo Pp 155ndash172 inTranscendental Arguments Edited by Robert Stern Oxford ClarendonPress

mdashmdashmdash 2000 ldquoThe Synthetic A priori in Strawsonrsquos KantianismrdquoPp 224ndash243 in Understanding Human Knowledge Oxford Oxford Uni-versity Press

Toulmin Stephen 1972 Human Understanding Princeton Princeton Uni-versity Press


Tully James 1980 A Discourse on Property John Locke and His AdversariesCambridge Cambridge University Press

Westman Robert S 1994 ldquoTwo Cultures or One A Second Look atKuhnrsquos The Copernican Revolutionrdquo Isis 85 79ndash115

Wittgenstein Ludwig 1970 Zettel Translated by G E M AnscombeBerkeley University of California Press

mdashmdashmdash 1993 ldquoPhilosophyrdquo Pp 161ndash199 in Philosophical OccasionsEdited by James Klagge and Alfred Nordmann Indianapolis Hackett

mdashmdashmdash (1951) 2000 Philosophical Investigations Translated by G E MAnscombe Oxford Blackwell



dependent upon history but also whether in doing history he wasinordmuenced by his philosophy Kuhn vehemently denied both

[ ] I have myself resisted attempts to amalgamate history andphilosophy of science though simultaneously urging increased in-teraction between the two History done for the sake of philosophyis often scarcely history at all (Kuhn 1980 p 183)

I have said repeatedly and I will say again you cannot do historytrying to document or to explore or to apply a point of view (Kuhn2000d pp 313ndash314 emphasis in the original)

If you have a theory you want to conordfrm you can go and do historyso it conordfrms it and so forth itrsquos just not the thing to do (Kuhn2000d p 314 emphasis in the original)

Irsquom never a philosopher and a historian at the same time (Kuhn2000d p 316)

Kuhn with this response guards himself against a number of objectionsthat will be recounted later Yet independently of Kuhnrsquos reaction thequestion remains what is the relation of history of science to philosophyof science in Kuhnrsquos model Kuhn made use of history he even claimedthat the two disciplines history and philosophy can do and should inter-act and fertilize each other (Kuhn 2000d pp 315-316) If the relation be-tween them is not that of theory to evidence what is it Irsquoll confine myselfto the exegetical point aiming at a coherent account that would compriseboth what Kuhn does and what Kuhn says he does I will try in particularto reconcile Kuhnrsquos historical orientation in the Structure with his laterclaim that his model of science can be derived from first principles2 I willbegin by illustrating the wide range of different opinions on the issue inthe recent and not so recent secondary literature on Kuhn and I will thenproceed to propose my own account

I How the relation between history and philosophy of science in the caseof Kuhnrsquos work is perceived in the recent literatureIn some of the recent literature one finds the following claims as regardsthe relation between history and philosophy of science in Kuhnrsquos work

1 Michael Friedman in his Dynamics of Reason repeatedly characterizesKuhnrsquos work as the ldquotheory of the nature and character of scientiordfc revolu-tionsrdquo (Friedman 2001 pp 19 41ndash44 119) Friedman argues in favor ofrelativized constitutive a priori principles in the sciences and he states that

2 See Kuhn 2000b p 95 Kuhn 2000c




























































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































































Documents

Kindi Perspectives