The difficulties in identifying and conceptualizing scientificrevolutions involve many of the most challenging issues inepistemology, methodology, ontology, philosophy of language, and evenvalue theory. With revolution we immediately confront the problem ofdeep, possibly noncumulative, conceptual and practical change, now inmodern science itself, a locus that Enlightenment thinkers would havefound surprising. And since revolution is typically driven by newresults, or by a conceptual-cum-social reorganization of old ones,often highly unexpected, we also confront the hard problem ofunderstanding creative innovation. Third, major revolutions supposedlychange the normative landscape of research by altering the goals andmethodological standards of the enterprise, so we face also thedifficult problem of relating descriptive claims to normative claimsand practices, and changes in the former to changes in the latter.
Comparing the world of business innovation and economic theoryprovides a perspective on the difficulty of these problems, for boththe sciences and the industrial technologies change rapidly andsometimes deeply (in the aforementioned ways), thanks to what might betermed “innovation push”—both the pressure toinnovate (to find and solve new problems, thereby creating newdesigns) and the pressure to accommodate innovation (see, e.g.,Christensen 1997; Christensen and Raynor, 2003; Arthur 2009). In amarket economy, as in science, there is a premium on change driven byinnovation. Yet most economists have treated innovations as exogenousfactors—as accidental, economically contingent events that comein from outside the economic system to work their effects. It issurprising that only recently has innovation become a central topic ofeconomic theorists. Decades ago, the Austrian-American economistJoseph Schumpeter characterized economic innovation as
the process of industrial mutation—if I may use that biologicalterm—that incessantly revolutionizes the economic structurefrom within, incessantly destroying the old one, incessantlycreating a new one. This process of Creative Destruction is theessential fact about capitalism. [1942, chap. VII; Schumpeter’semphasis]
Unfortunately, economists largely ignored this sort of claim (madealso by a few others) until the recent development of economic growththeory (e.g., Robert Solow, Paul Romer, and W. Brian Arthur: seeBeinhocker 2006 and Warsh 2006). The result was an inability ofeconomic models to account for economic innovation endogenously and,thereby, to gain an adequate understanding of the generation ofeconomic wealth.
The parallel observation holds for philosophy of science. Here, too,the leading philosophers of science until the 1960s—the logicalempiricists and the Popperians—rejected innovation as alegitimate topic, even though it is the primary intellectual driver ofscientific change and producer of the wealth of skilled knowledge thatresults. The general idea is that the so-called context of discovery,the context of creatively constructing new theories, experimentaldesigns, etc., is only of historical and psychological interest, notepistemological interest, and that the latter resides in the epistemicstatus of the “final products” of investigation. On thisview, convincing confirmation or refutation of a claim enablesscientists to render an epistemic judgment that detaches it from itshistorical context. This judgment is based on the logical relations oftheories and evidence rather than on history or psychology. Accordingto this traditional view, there exists a logic of justification butnot a logic of discovery. The distinction has nineteenth-centuryantecedents (Laudan 1980). Cohen and Nagel (1934) contended that totake historical path into account as part of the epistemic assessmentwas to confuse historical questions with logical questions and therebyto commit what they called a “genetic fallacy.” However,the context of discovery / context of justification distinction (orfamily of distinctions) is often attributed to Reichenbach (1938).(See the entry on Reichenbach. For recent discussion see Schickore and Steinle, 2006.)
Today there are entire academic industries devoted to various aspectsof the topic of scientific revolutions, whether political orscientific, yet we have no adequate general theory or model ofrevolutions in either sphere. This article will focus on ThomasKuhn’s conception of scientific revolutions, which relies partlyon analogies to political revolution and to religious conversion.Kuhn’s is by far the most discussed account of scientificrevolutions and did much to reshape the field of philosophy ofscience, given his controversial claims about incommensurability,rationality, objectivity, progress, and realism. For a general accountof Kuhn’s work, see the entry on Kuhn. See also Hoyningen-Huene (1993), and Bird (2001).
What history lies behind the terms ‘revolution’ and‘scientific revolution’? The answer is an intriguing mixof accounts of physical phenomena, political fortunes, and conceptionsof chance, fate, and history. Originally a term applying to rotatingwheels and including the revolution of the celestial bodies (as inCopernicus’ title: De Revolutionibus Orbium Coelestium)and, more metaphorically, the wheel of fortune,‘revolution’ was eventually transferred to the politicalrealm. The term later returned to science at the metalevel, todescribe developments within science itself (e.g., “theCopernican Revolution”). Christopher Hill, historian ofseventeenth-century Britain and of the so-called English Revolution inparticular, writes:
Conventional wisdom has it that the word ‘revolution’acquired its modern political meaning only after 1688. Previously ithad been an astronomical and astrological term limited to therevolution of the heavens, or to any complete circular motion. [Hill1990, 82]
Hill himself dates the shift to the governmental realm somewhatearlier, pointing out that the notion of overturning was alsopresent in groups of reformers who aspired to return human society toan earlier, ideal state: overturning as returning.This conception of revolution as overturning was compatible with acyclical view of history as a continuous process.
It was in the socio-political sphere that talk of revolution as asuccessful uprising and overturning became common. In this sense, arevolution is a successful revolt, ‘revolution’ being anachievement or product term, whereas ‘to revolt’ is aprocess verb. The fully modern conception of revolution as involving abreak from the past—an abrupt, humanly-made overturning ratherthan a natural overturning—depended on the linear, progressiveconception of history that perhaps originated in the ItalianRenaissance, gained strength during the Protestant Reformation and thetwo later English revolutions, and became practically dogma among thechampions of the scientific Enlightenment. The violent EnglishRevolution of the 1640s gave political revolution a bad name, whereasthe Glorious Revolution of 1688, a bloodless, negotiated compromise,reversed this reputation.
When did the term ‘revolution’ become a descriptor ofspecifically scientific developments? In the most thorough treatmentof the history of the concept of scientific revolution, I. B. Cohen(1985) notes that the French word revolution was being usedin early eighteenth-century France to mark significant developments.By mid-century it was pretty clear that Clairaut, D’Alembert,Diderot and others sometimes applied the term to scientificdevelopments, including Newton’s achievement but also toDescartes’ rejection of Aristotelian philosophy. Cohen fails tonote that Émilie Du Châtelet preceded them, in herInstitutions de Physique of 1740, where she distinguishedscientific from political revolutions (Châtelet and Zinnser2009, p. 118). However, the definition of revolution in theEncyclopédie of the French philosophes wasstill political. Toward the end of the century, Condorcet could speakof Lavoisier as having brought about a revolution in chemistry; and,indeed, Lavoisier and his associates also applied the term to theirwork, as did Cuvier to his. Meanwhile, of course, Kant, in TheCritique of Pure Reason (first edition 1781), spoke of his“Copernican Revolution” in philosophy. In fact, Cohen(1985) and Ian Hacking (2012) credit Kant with originating the idea ofa scientific revolution, although Kant had read Du Châtelet.Interestingly, for Kant (1798) political revolutions are, by nature,unlawful, whereas Locke, in his social contract theory, had permittedthem under special circumstances.
It was during the twentieth century that talk of scientificrevolutions slowly gained currency. One can find scientists using theterm occasionally. For example, young Einstein, in a letter to hisfriend Habicht, describes his new paper on light quanta as “veryrevolutionary” (Klein 1963, 59). The idea of radical breaks wasforeign to such historians of science as Pierre Duhem and GeorgeSarton, but Alexandre Koyré, in ÉtudesGaliléennes (1939), rejected inductivist history,interpreting the work of Galileo as a sort of Platonic intellectualtransformation. (See Zambelli (2016) for a revealing account ofKoyré’s own background.) In The Origins of ModernScience: 1300–1800 (1949 and later editions), widely usedas a course text, Herbert Butterfield, a political historian workingmainly from secondary sources, wrote a compact summary of theScientific Revolution, one that emphasized the importance ofconceptual transformation rather than the infusion of new empiricalinformation. The anti-whiggism that he had advocated in his TheWhig Interpretation of History (1931) became a major constrainton the new historiography of science, especially in the Anglophoneworld. In Origins, Butterfield applied the revolution labelnot only to the Scientific Revolution and to several of its componentsbut also to “The Postponed Revolution in Chemistry” (achapter title), as if it were a delayed component of the ScientificRevolution. His history ended there. A revolution for Butterfield is amajor event that founds a scientific field. Taken together, theserevolutions founded modern science. As the title of his book suggests,he was concerned with origins, not with what comes after the founding.In the Introduction he famously (or notoriously) stated that theScientific Revolution
outshines everything since the rise of Christianity and reduces theRenaissance and Reformation to the rank of mere episodes, mereinternal displacements, within the system of medieval Christendom.
For Butterfield, the Scientific Revolution was a watershed event onthe scale of total human history, an event that, somewhat ironicallyand somewhat like Christianity according to its believers, enabled thesciences, to some degree, to escape from history and thereby to becomeexceptional among human endeavors. Subsequently, A. Rupert Hall, afull-fledged historian of science who worked from primary sources,published The Scientific Revolution: 1500–1800 (Hall1954). Soon many other scholars spoke of the ScientificRevolution, the achievements of the period from Copernicus to Newton,including such luminaries as Kepler, Galileo, Bacon, Descartes,Huygens, Boyle, and Leibniz.
Then Thomas Kuhn and Paul Feyerabend challenged received views ofscience and made talk of revolutionary breaks and incommensurabilitycentral to the emerging new field of history and philosophy ofscience. They asserted that major conceptual changes lay in the futureof mature, modern sciences as well as in their past. Kuhn (1962, ch.IX) contended that there will be no end to scientific revolutions aslong as systematic scientific investigation continues, for they are anecessary vehicle of ongoing scientific progress–necessary tobreak out of dated conceptual frameworks. In other words, there areboth founding revolutions, in something like Butterfield’s senseof threshold events to maturity, and a never-ending series of laterrevolutions within an ongoing field, no matter how mature it is.However, soon after Structure, Kuhn had second thoughts andeventually abandoned the Butterfield conception of revolution, on theground that even his so-called preparadigm schools had their paradigms(Kuhn 1974, 460, note 4; details below). So multiple Kuhnian paradigmsin long-term competition now became possible.
The Scientific Revolution was the topic around which the field ofhistory of science itself came to maturity. Kuhn’spopularization of the idea that even the mature natural sciencesundergo deep conceptual change stimulated much general intellectualinterest in the history of science during the 1960s and 1970s. Therevolution frame of reference was also a boon to historiographicalnarrative itself (see Cohen 1985 and Nickles 2006). And by challengingthe received, quasi-foundational, Enlightenment conception of science,history of science and related philosophies of science gained greatcultural significance for a time.
In recent decades, however, many historians have contested even theclaim that there was a single, coherent development appropriatelycalled “the Scientific Revolution.” Steven Shapin (1996,1) captured the unease in his opening sentence: “There was nosuch thing as the Scientific Revolution, and this is a book aboutit.” Everyone agrees that a series of rapid developments ofvarious kinds took place during the period in question, but theoperative word here is ‘various’. One difficulty is thatno one has succeeded in capturing a 150-year (or more) period of workin an insightful, widely accepted characterization that embraces theimportant changes in theory, method, practices, instrumentation,social organization, and social status ranging over such a widevariety of projects. The very attempt has come to seem reductionist.Older styles of historical writing were characterized by grandnarratives such as “the mechanization of the worldpicture” (Dijksterhuis 1961; original, Dutch edition, 1950) andhumanity’s passage from subjective superstition to objectivityand mathematical precision (Gillispie 1960). Philosophically orientedwriters attempted to find unity and progress in terms of the discoveryof a new, special scientific method. Today even most philosophers ofscience dismiss the claim that there exists a powerful, general,scientific method, the discovery of which explains the ScientificRevolution and the success of modern science. Quite the contrary:effective scientific methods are themselves the product of painstakingwork at the frontier—scientific results methodized—and arehence typically laden with the technical content of the specialty inquestion. There is no content-neutral, thereby general and timelessmethod that magically explains how those results were achieved(Schuster and Yeo 1986, Nickles 2009).
Continuity theorists such as Pierre Duhem (1914), John Herman Randall(1940), A. C. Crombie (1959, 1994), and more recent historians such asPeter Dear (2001) have pointed out a second major difficulty inspeaking of “the Scientific Revolution.” It is hard tolocate the sharp break from medieval and Renaissance practices thatdiscontinuity historians from Koyré to Kuhn have celebrated.When examined closely in their own cultural context, all the supposedrevolutionaries are found to have had one foot in the old traditionsand to have relied heavily on the work of predecessors. In this vein,J. M. Keynes famously remarked that Newton was “the last of themagicians,” not the first of the age of reason (Keynes 1947).Still, most historians and philosophers would agree that the rate ofchange of scientific development increased notably during this period.Hence, Shapin, despite his professional reservations, could stillwrite an instructive, synthetic book about the Scientific Revolution.The most thorough appraisal of historiographical treatments of theScientific Revolution is H. Floris Cohen’s (1994).
The Scientific Revolution supposedly encompassed all of science ornatural philosophy, as it then existed, with major socialimplications, as opposed to more recent talk of revolutions withinparticular technical fields. Have there been other multidisciplinaryrevolutions? Some have claimed the existence of a “secondscientific revolution” in the institutional structure of thesciences in the decades around 1800, especially in France, others(including Kuhn 1977a, ch. 3) of a multidisciplinary revolution in the“Baconian sciences” (chemistry, electricity, magnetism,heat, etc.) during roughly the same time period. Enrico Bellone 1980),Kuhn, and others Kuhn have focused on the tremendous increase inmathematical abstraction and sophistication during the early-to-midnineteenth century that essentially created what we know asmathematical physics. Still others have claimed that there was ageneral revolution in the sciences in the decades around 1900. (Seealso Cohen 1985, chap. 6, for discussion of these claims.)
For many historians, ‘the Scientific Revolution’ nowdescribes a topic area rather than a clearly demarcated event. Theyfind it safer to divide the Scientific Revolution into several moretopic- and project-specific developments. However, in their unusuallycomprehensive history of science textbook, Peter Bowler and Iwan Morus(2005) query of practically every major development they discusswhether or not it was a genuine revolution at all, at least by Kuhnianstandards. More recently, David Wootton’s (2015) is arevisionist account that returns to a more heroic understanding of theScientific Revolution.
Commitment to the existence of deep scientific change does not, forall experts, equate to a commitment to the existence of revolutions inKuhn’s sense. Consider the historically–orientedphilosopher Stephen Toulmin (1953, 1961, 1972), who wrote of“ideals of natural order,” principles so basic that theyare normally taken for granted during an epoch but that are subject toeventual historical change. Such was the change from the Aristotelianto the Newtonian conception of inertia. Yet Toulmin remained criticalof revolution talk. Although the three influential college coursetexts that he co-authored with June Goodfield recounted the majorchanges that resulted in the development of several modern sciences(Toulmin and Goodfield 1961, 1962, 1965), these authors could write,already about the so-called Copernican Revolution:
We must now look past the half-truths of this caricature, to whatCopernicus attempted and what he in fact achieved. For in science, asin politics, the term ‘revolution’—with itsimplication that a whole elaborate structure is torn down andreconstructed overnight—can be extremely misleading. In thedevelopment of science, as we shall see, thorough-going revolutionsare just about out of the question. [1961, 164]
The Toulmin and Goodfield quotation invites us to ask, When did talkof scientific revolutions enter philosophy of science in a significantway? And the answer seems to be: there is a sprinkling of uses of theterm ‘scientific revolution’ and its cognates prior toKuhn, but these were ordinary expressions that did not yet have thestatus of a technical term.
Given the prominence of the topic today, it is surprising that we donot find the term in Philipp Frank’s account of the positivistdiscussion group in Vienna in the early twentieth century. However,Frank (1957) does speak of their perception of a “crisis”in modern physics caused by the undermining of classical mechanics byspecial relativity and quantum mechanics, and it was common to speakof this or that worldview or world picture (Weltanschauung,Weltbild), e.g., the electromagnetic vs. the Einsteinian vs.the mechanical picture. Nor do we find talk of scientific revolutionsin the later Vienna Circle, even after the diaspora following the riseof Hitler. The technical term does not appear in Karl Popper’sLogik der Forschung (1934) nor in his 1959 English expansionof that work as The Logic of Scientific Discovery, at leastnot important enough to be indexed. Hans Reichenbach (1951) speaksrather casually of the revolutions in physics. The technical term isnot in Ernest Nagel’s The Structure of Science (1961).Nor is it in Stephen Pepper’s World Hypotheses (1942).It plays no significant role in N. R. Hanson’s Patterns inDiscovery (1958), despite its talk of the theory-ladenness ofobservation and perceptual Gestalt switches. Meanwhile, there were, ofcourse, a few widely-read works in the background that spoke of majorontological changes associated with the rise of modern science,especially E. A. Burtt’s Metaphysical Foundations of ModernPhysical Science (1924). Burtt’s book influencedKoyré, who, in turn, influenced Kuhn.
In his retrospective autobiographical lecture at Cambridge in 1953,Popper did refer to the dramatic political and intellectual events ofhis youth as revolutionary:
[T]he air was full of revolutionary slogans and ideas, and new andoften wild theories. Among the theories which interested meEinstein’s theory of relativity was no doubt by far the mostimportant. The others were Marx’s theory of history,Freud’s psycho-analysis, and Alfred Adler’s so-called‘individual psychology’. [Popper 1957]
And during the 1960s and 1970s, Popper indicated that, according tohis “critical approach” to science and philosophy, allscience should be revolutionary—revolution in permanence. Butthis was a tame conception of revolution compared to Kuhn’s,given Popper’s two logical criteria for a progressive newtheory: (a) it must logically conflict with its predecessor andoverthrow it; yet (b) “a new theory, however revolutionary, mustalways be able to explain fully the success of its predecessor”(Popper 1975). As we shall see, Kuhn’s model of revolutionrejects both these constraints (depending on how one interprets hisincommensurability claim) as well as the idea of progress towardfinal, big, theoretical truths about the universe. Kuhn dismissedPopper’s notion of revolution in perpetuity as a contradictionin terms, on the ground that a revolution is something that overthrowsa long and well–established order, in violation of the rules ofthat order. Kuhn (1970) also vehemently rejected Popper’sdoctrine of falsification, which implied that a theory could berejected in isolation, without anything to replace it. According toPopper, at any time there may be several competing theories beingproposed and subsequently refuted by failed empiricaltests—rather like several balloons being launched, over time,and then being shot down, one by one. Popper’s view thus facesthe difficulty, among others, of explaining the long-term coherencethat historians find in scientific research.
Beginning in the 1960s, several philosophers and historians addressedthis difficulty by proposing the existence of larger units (thantheories) of and for analysis. Kuhn’s paradigms, ImreLakatos’s research programmes, Larry Laudan’s researchtraditions (Lakatos 1970, Laudan 1977), and the widespread use ofterms such as ‘conceptual scheme’, ‘conceptualframework’, ‘worldview’, and Weltanschauung(Suppe 1974) instanced this felt need for larger-sized units amongAnglo-American writers, as had Toulmin’s old concept of idealsof natural order. These stable formations correspondingly raised theeventual prospect of larger-scale instabilities, for an abrupt changein such a formation would surely be more dramatic, more revolutionary,than a Popperian theory change. However, none of the other writersendorsed Kuhn’s radical conception of scientific revolution.Meanwhile, Michel Foucault 1963, 1966, 1969, 1975), working in aFrench tradition, was positing the existence of “discursiveformations” or epistemes, sets of deep-structuralcultural rules that define the limits of discourse during a period.Section 5 returns to this theme.
I. B. Cohen (1985, chap. 2) lays down four historical tests, fournecessary conditions, for the correct attribution of a revolution.First, the scientists involved in the development must perceivethemselves as revolutionaries, and relevant contemporaries must agreethat a revolution is underway. Second, documentary histories mustcount it as a revolution. Third, later historians and philosophersmust agree with this attribution and, fourth, so must later scientistsworking in that field or its successors. By including both reportsfrom the time of the alleged revolution and later historiographicaljudgments, Cohen excludes people who claimed in their day to berevolutionaries but who had insufficient impact on the field tosustain the judgment of history. He also guards against whiggish, posthoc attributions of revolution to people who had no idea that theywere revolutionaries. His own four examples of big scientificrevolutions all have an institutional dimension: The ScientificRevolution featured the rise of scientific societies and journals, thesecond was the aforementioned revolution in measurement from roughly1800 to 1850 (which Kuhn, too, called “the second scientificrevolution”; 1977, 220). Third is the rise of universitygraduate research toward the end of that century. Fourth is thepost-World War II explosion in government funding of science and itsinstitutions.
Cohen sets the bar high. Given Copernicus’ own conservatism andthe fact that few people paid attention to his work for half acentury, the Copernican achievement was not a revolution byCohen’s lights. Or if there was a revolution, should it not beattributed to Kepler, Galileo, and Descartes? This thought furtherproblematizes the notion of revolution, for science studies experts aswell as scientists themselves know that scientific and technologicalinnovation can be extremely nonlinear in the sense that a seeminglysmall, rather ordinary development may eventually open up an entirenew domain of research problems or a powerful new approach. ConsiderPlanck’s semi-classical derivation of the empirical blackbodyradiation law in 1900, which, under successively deeper theoreticalderivations by himself and (mainly) others over the next two and ahalf decades, became a pillar of the revolutionary quantum theory. AsKuhn (1978) shows, despite the flood of later attributions to Planck,it is surprisingly difficult, on historical and philosophical grounds,to justify the claim that he either was, or saw himself as, arevolutionary in 1900 and for many years thereafter. (Kuhn 2000boffers a short summary.) Augustine Brannigan (1981) and Robert Olby(1985) defend similar claims about Mendel’s alleged discovery ofMendelian inheritance.
These examples suggest that Cohen’s account of scientificrevolution (and Kuhn’s) is tied too closely to the idea ofpolitical revolution in placing so much weight on the intentions ofthe generators. In the last analysis, many would agree, revolution,like speciation in biology, is a retrospective judgment, a judgment ofeventual consequences, not something that is always directlyobservable as such in its initial phases, e.g., in the stated theintentions of its authors. On the other hand, a counterintuitiveimplication of this consequentialist view of revolutions is that therecan be revolution without revolt (assuming that revolt is a deliberatecourse of action), revolutionary work without authors, so to speak, orat least revolutionary in eventual meaning despite the authors’intentions. Then why not just speak of evolution rather thanrevolution in such cases? For, as we know by analogy from evolutionarybiology, in the long run evolution can be equally transformative, evenmoreso (see below).
A related point is that, insofar as revolutions are highly nonlinear,it is difficult to ascribe to them any particular reason or cause;for, as indicated, the triggering events can be quite ordinary work,work that unexpectedly opens up new vistas for exploration. A smallcause may have an enormous effect. To be sure, the state of therelevant scientific system must be such that the events do function astriggers, but we need not expect that such a system always be readilyidentifiable as one in crisis in Kuhn’s sense. Rather, thehighly nonlinear revolutionary developments can be regarded asstatistical fluctuations out of a “noisy” background ofordinary work. At any rate, on this view it is a mistake to think thatexplaining revolutions requires locating a momentous breakthrough(Nickles 2012a and b).
What of the common requirement that revolutions be rapid, event-like,unlike the century-and-a half-long Scientific Revolution? Brad Wray(2011, 42f) answers that there is no reason that a revolution need bean abrupt event. What is important is how thoroughgoing the change isand that it be change at the community level rather than a Gestaltswitch experienced by certain individuals. (After the originalpublication of Structure, Kuhn acknowledged his confusion inattributing Gestalt switches to the community as a whole as well as toindividuals.) On Wray’s view, evolution and revolution are notnecessarily opposed categories. And with this understanding, theToulmin and Goodfield comment quoted above becomes compatible withrevolutionary transformation, which, not surprisingly, takes time tobecome thoroughgoing. Meanwhile, the Butterfield quotation suggeststhat what counts as a striking change is a matter of historical scale.By our lights today, 150 years is a long time; but, against the longsweep of total human history, a change of the magnitude of theScientific Revolution was quite rapid. Perhaps today’s rapidpace of scientific and technological innovation makes us impatientwith slower-scaled developments in the past. And it is surely the casethe some of the slow, large-scale transformations now underway arescarcely visible to us.
Finally, what of Butterfield’s criterion of broader socialimpacts? Kuhn retained this criterion in The CopernicanRevolution, but revolutions increasingly become changes inspecialist communities in his later work, since those communitiesinsulate themselves from the larger society. In the chapter on theinvisibility of revolutions in Structure, Kuhn tells us thata tiny subspecialty can undergo a revolution that looks like acumulative change even to neighboring fields of the same scientificdiscipline. In this respect Kuhn remained an internalist.
Although virtually no one in the science studies fields acceptsKuhn’s model in Structure as correct in detail, therehas been a revival of interest in his views since his death and, morerecently, in connection with the fiftieth anniversary in 2012 of thebook’s original publication. Some examples are: the fiftiethanniversary edition of Structure itself, including a valuableintroduction by Ian Hacking; Kuhn (2000a), a collection that recordsthe later evolution of Kuhn’s thought; Sankey (1997); Bitbol(1997); Fuller (2000); Bird (2001); Friedman (2001); Andersen (2001);Sharrock and Read (2002); Nickles (2003a); González (2004);Soler et al. (2008); Agazzi (2008); Gattei (2008); Torres (2010); Wray(2011), Kindi and Arabatzis (2012), De Langhe (2013), Marcum (2015),and Richards and Daston (2016). Kuhn on revolutions has helped toshape many symposia on scientific realism and related matters, forexample, Soler (2008) on contingency in the historical development ofscience and Rowbottom and Bueno (2011) on Bas van Fraassen’s(2002) treatment of stance, voluntarism, and the viability ofempiricism. Since Kuhn’s work is discussed in some detail inother contributions to this Encyclopedia (see, especially,“Kuhn, Thomas”, and “The Incommensurability ofScientific Theories”), a brief account will suffice here. For adetailed reading guide to Structure, consult Preston(2008).
According to Kuhn in Structure, a loosely characterized groupof activities, often consisting of competing schools, becomes a maturescience when a few concrete problem solutions provide models for whatgood research is (or can be) in that domain. These exemplaryproblems-cum-solutions become the basis of a “paradigm”that defines what it is to do “normal science.” As itsname suggests, normal science is the default state of a mature scienceand of the community of researchers who constitute it. The paradigminforms investigators what their domain of the world is like andpractically guarantees that all legitimate problems can be solved inits terms. Normal science is convergent rather than divergent: itactively discourages revolutionary initiatives and essentially novel(unexpected) discoveries, for these threaten the paradigm. However,normal research is so detailed and focused that it is bound to turn upanomalous experimental and theoretical results, some of which willlong resist the best attempts to resolve them. Given the historicalcontingencies involved in the formation of guiding paradigms as wellas the fallibility of all investigators, it would be incrediblyimprobable for everything to end up working perfectly. According toKuhn, anomalies are therefore to be expected. Historically, allparadigms and theory complexes face anomalies at all times. If andwhen persistent efforts by the best researchers fail to resolve theanomalies, the community begins to lose confidence in the paradigm anda crisis period ensues in which serious alternatives can now beentertained. If one of these alternatives shows sufficient promise toattract a dominant group of leading researchers away from the oldparadigm, a paradigm shift or paradigm change occurs—and that isa Kuhnian revolution. The radicals accomplish this by replacing theformer set of routine problems and problem-solving techniques(exemplars) by a new set of exemplars, making the old practices seemdefective, or at least old fashioned.
The new paradigm overturns the old by displacing it as no longer acompetent guide to future research. In the famous (ornotorious)chapter X of Structure, Kuhn claims that the changeis typically so radical that the two paradigms cannot be comparedagainst the same goals and methodological standards and values.Moreover, the accompanying meaning shift of key terms, such as‘simultaneous’, ‘mass’, and‘force’ in physics, leads to communication breakdown. Ineffect, scientists on different sides of a paradigm debate “livein different worlds.” Kuhn speaks of scientists experiencing akind of gestalt switch or religious conversion experience. The heatedrhetoric of debate and the resulting social reorganization, he says,resemble those of a political revolution. “Like the choicebetween political institutions, that between competing paradigmsproves to be a choice between incompatible modes of communitylife” (1970, 94). The comparison of scientific with politicalrevolutions should not surprise, given the entangled history of theterm ‘revolution’, but claiming such close similarityenraged philosophical and cultural critics of Kuhn.
The typical paradigm change does not involve a large infusion of newempirical results, Kuhn tells us (chs. IX and X). Rather, it is aconceptual reorganization of otherwise familiar materials, as in therelativity revolution. A paradigm change typically changes goals,standards, linguistic meaning, key scientific practices, the way boththe technical content and the relevant specialist community areorganized, and the way scientists perceive the world. (For the oftenneglected practices dimension in Kuhn’s account, see Rouse,2003.) Nor can we retain the old, linear, cumulative conception ofscientific progress characteristic of Enlightenment thinking; for,Kuhn insists, attempts to to show that the new paradigm contains theold, either logically or in some limit or under some approximation,will be guilty of a fallacy of equivocation. The meaning changereflects the radical change in the assumed ontology of the world. Asecond Kuhnian objection to cumulative progress is what has come to becalled “Kuhn loss” (see Post 1971, 229, n. 38). Rarelydoes the new paradigm solve all of the problems that its predecessorapparently solved. So even in this sense the new paradigm failscompletely to enclose the old. The consequence, according to Kuhn, isthat attempts to defend continuous, cumulative scientific progress bymeans of theory reduction or even a correspondence relationship (e.g.,a limiting relationship) between a theory and its predecessor mustfail. Revolutions produce discontinuities.
Given all these changes, Kuhn claimed that the two competing paradigmsare “incommensurable”, a technical term that he repeatedlyattempted to clarify. Traditional appeals to empirical results andlogical argument are insufficient to resolve the debate. For detailsof the incommensurability debate, see the entry “TheIncommensurability of Scientific Theories.” as well asHoyningen-Huene and Sankey (2001) as a sample of the large literatureon incommensurability.
Naturally, many thinkers of a logical empiricist or Popperian bent, orsimply of an Enlightenment persuasion, were shocked by these claimsand responded with a barrage of criticism—as if Kuhn hadcommitted a kind of sacrilege by defiling the only human institutionthat could be trusted to provide the objective truth about the world.Today there is fairly wide agreement that some of Kuhn’s claimsno longer look so radical. Meanwhile, Kuhn himself was equally shockedby the vehemence of the attacks and (to his mind) the willfuldistortion of his views (see, e.g., Lakatos and Musgrave 1970). Inlater papers and talks, he both clarified his views and softened someof his more radical claims. Critics reacted to the radical views ofPaul Feyerabend (1962, 1975) in a somewhat similar manner. (Fordetails, see the entry “Feyerabend, Paul.”)
Given that cyclic theories of history have, for the most part, longgiven way to linear, progressive accounts, readers may be surprised atKuhn critic, physicist Stephen Weinberg’s comment thatKuhn’s overall model is still, in a sense, cyclic (Weinberg2001). In fact, Kuhn himself had already recognized this. After thefounding paradigm in Kuhn’s account in Structure, wehave normal science under a paradigm, then crisis, then revolution,then a new paradigm—a development that brings back a new periodof normal science. At this abstract level of description, the model isindeed cyclic, but of course the new paradigm heads the science inquestion in a new direction rather than returning it to a previousstate. Other commentators, including Marxists, have regardedKuhn’s mechanism as dialectical, as illustrated by thesuccession of self-undermining developments in the theory of light,from a Newtonian particle theory to a wave theory to a new kind ofwave-particle duality. (For the dialectical interpretation seeespecially Krajewski 1977 and Nowak 1980 on the idealizationalapproach to science, as originated by Karl Marx.)
Somewhat ironically, Kuhn’s attempt to revolutionize theepistemology of science has had a wider socio-cultural impact thanmany scientific revolutions themselves. While some of Kuhn’sdoctrines step into the postmodern era, he still had a foot in theEnlightenment, which helps to explain his dismay at the criticalreaction to his work and to radical developments in the new-wavesociology of science of the 1970s and ‘80s. For, unlike manypostmodernists (some of whom make use of his work), Kuhn retained ascientific exceptionalism. He did not doubt that the sciences havebeen uniquely successful since the Scientific Revolution. For him,unlike for many of his critics, revolutions in his radical sense weregreat epistemological leaps forward rather than deep scientificfailures. On the science policy front, he intended his work to helppreserve the integrity of this socially valuable enterprise. It is onscience policy issues that Steve Fuller is most critical of Kuhn(Fuller 2000).
The general problem presented by Kuhn’s critique of traditionalphilosophy of science is that, although the various sciences have beensuccessful, we do not understand how they have accomplishedthis or even how to characterize this success. Enlightenment-styleexplanations have failed. For example, Kuhn and Feyerabend (1975),preceded by Popper, were among the first philosophers to expose thebankruptcy of the claim that it was the discovery of a specialscientific method that explains that success, a view that is stillwidely taught in secondary schools today. And that conclusion (onethat cheered those postmodernists who regard scientific progress as anillusion) left Kuhn and the science studies profession with theproblem of how science really does work. To explain how and why it hadbeen so successful became an urgent problem for him—again, aproblem largely rejected as bogus by many science studies scholarsother than philosophers.
Another of Kuhn’s declared tasks in Structure was tosolve the problem of social order for mature science, that is, howcohesive modern science (especially normal science) is possible(Barnes 1982, 2003). Yet another was to bring scientific discoveryback into philosophical discussion by endogenizing it in his model,while denying the existence of a logic of discovery. Whereasthe logical empiricists and Popper had excluded discovery issues fromphilosophy of science in favor of theory of confirmation orcorroboration, Kuhn was critical of confirmation theory and supportiveof historical and philosophical work on discovery. He argued thatdiscoveries are temporally and cognitively structured and that theyare an essential component of an epistemology of science. In Kuhniannormal science the problems are so well structured and the solutionsso nearly guaranteed in terms of the resources of the paradigm thatthe problems reduce to puzzles (Nickles 2003b). Kuhn kept things undercontrol there by denying that normal scientists seek essentialinnovation, for, as indicated above, major, unexpected discoveriesthreaten the extant paradigm and hence threaten crisis and revolution.So, even in normal science, Kuhn had to admit that major discoveriesare unexpected challenges to the reigning paradigm. They areanomalous, even exogenous in the sense that they come as shocks fromoutside the normal system.
But this is the working scientists’ point of view. As noted,normal science is bound to turn up difficulties that resistresolution, at least some of which are sooner or later recognized bythe community. In Kuhn’s own view, as a historian andphilosopher standing high above the fray, it is deliberate, systematicnormal research that will most readily sow the seeds of revolution andhence of rapid scientific progress. According to the oldmusicians’ joke, the fastest way to Carnegie Hall is slowpractice. For Kuhn the fastest way to revolutionary innovation isintensely detailed normal science.
When it comes to revolution on Kuhn’s account, the social orderbreaks down dramatically. And here his strategy of taming creativenormal research so as to make room for articulated discovery (thereduction of research problems to puzzles) also breaks down. Kuhn hadto acknowledge that he had no idea how the scientists in extraordinaryresearch contexts manage to come up with brilliant new ideas andtechniques. This failure exacerbated his problem of explaining whatsort of continuity underlies the revolutionary break that enables usto identify the event as a revolution within an ongoing field ofinquiry. As he later wrote:
Even those who have followed me this far will want to know how avalue-based enterprise of the sort I have described can develop as ascience does, repeatedly producing powerful new techniques forprediction and control. To that question, unfortunately, I have noanswer at all…. [1977b, 332]
Kuhn’s work on scientific revolutions raises difficult questionsabout whether science progresses and, if so, in what that progressconsists. Kuhn asks (p. 60), “Why is progress a perquisitereserved almost exclusively for the activities we call science”and not for art, political theory, or philosophy? Early critics tookhim to deny scientific progress, because he rejected the traditionalcorrespondence theory of truth and the related idea of cumulativeprogress toward a representational truth waiting out there for scienceto find it. For Kuhn the internalist, the technical goals of scienceare endogenously generated and change over time, rapidly duringrevolutions. Yet, somewhat paradoxically, Kuhn regarded revolutions asthe most progressive components of his model of science.Unfortunately, he was not able to articulate fully in what thatprogress consists, given the issues of truth, incommensurability andKuhn loss, a problem that those who reject convergent scientificrealism still face. However, problem-solving know-how and success,including predictive precision, are major components of his answer.“[T]he unit of scientific achievement is the solvedproblem....” (p. 169). In a retreat from his most radicalstatements, Kuhn responded to critics by saying that we do possess ageneral set of scientific values that enables us, usually prettyeasily, to order scientific works in historical time according to thedegree in which they realize these values. A new paradigm, he says,must always treat successfully a serious anomaly left by the old oneas well as opening up new questions for fruitful investigation.
Kuhn’s emphasis on scientific practices, relative to thephilosophical state of play in the 1960s, takes up some of the slackleft by the rejection of strong realism. His emphasis on skilledpractice may have been influenced by Michael Polanyi’sPersonal Knowledge (1958), with its “tacitknowing” component, although Kuhn denied that he foundPolanyi’s account appealing (see, e.g., Baltas et al., 2000, pp.296f).
If there have been so many revolutions, then why did the world have towait for Kuhn to see them? Because, he said, they are largelyinvisible. For, after a revolution, the winners rewrite the history ofscience to make it look as if the present paradigm is the brilliantbut rational sequel to previous work. The implication is that onlysomeone of Kuhn’s historical sensitivity could be expected tonotice this. (Skeptical critics reply that Kuhn invented the problemfor which he had a solution.) Indeed, in his large book on the historyof the early quantum theory (Kuhn 1978), he moved the origin of thequantum theory revolution forward five years, from Planck in 1900 toEinstein and Ehrenfest in 1905. Revisionist historiography by whiggishscientists, he claimed, had smoothed out the actual history bycrediting Planck with a solution that he actually rejected at the timeto a problem that he did not then have—and by diminishing thetruly radical contribution of Einstein. Kuhn’s move again raisesthe question whether the authors of a revolution must knowingly breakfrom the received research tradition.
At the end of Structure, Kuhn drew an analogy between thedevelopment of science and evolutionary biology. This was surprising,since ‘evolution’ is commonly employed as a contrast termto ‘revolution’. Kuhn’s main point was thatevolution ramifies rather than progressing toward a final goal, yetits degree of specialization through speciation can be regarded as asort of progress, a progress from a historically existingbenchmark rather than a progress toward a preordained,speculative goal. So specialization is an indicator of progress. Asfor revolutions, they correspond to macromutations.
The process described in Section XII as the resolution of revolutionsis the selection by conflict within the scientific community of thefittest way to practice future science. The net result of a sequenceof such revolutionary selections, separated by periods of normalresearch, is the wonderfully adapted set of instruments we call modernscientific knowledge. Successive stages in that developmental processare marked by an increase in articulation and specialization. And theentire process may have occurred, as we now suppose biologicalevolution did, without benefit of a set goal… . [1970, 172f]
At the time, it was striking that Kuhn compared revolutionarytransitions, rather than normal scientific developments, withevolutionary change. It seems clear that he did not considerrevolution and evolution to be mutually incompatible. But keep in mindthat, for him, normal science represents periods of stasis, whereasrevolutions are short, highly creative periods that more closelyresemble the exploration by random trial and error (p. 87) that weassociate with biological evolution. Examined on a minute time scale,however, normal science arguably also involves a (more constrained)variation and selection process, as scientific practitioners searchfor ways to articulate the paradigm. So, given the inevitability ofevolution under such conditions, Kuhn’s treatment of normalscience would appear to be too static. The important point here isthat, in Kuhn’s view revolution and evolution are compatiblewhen considered on the correct time scales (see Wray 2011 andKuukkanen 2012). Examined from afar, revolutions are simply the morenoteworthy episodes in the evolution of the sciences. Examined upclose, they (like discoveries in general for Kuhn) have a detailedstructure that is evolutionary, even something as revolutionary as thequantum theory (Kuhn 1978).
But how, then, the reader is entitled to ask, can Kuhn accommodate thesharp discontinuities that he advertised in chapter X of the book? Wecannot equate revolutions simply with speciation in Darwin’s ownsense, given that Darwin’s favorite mechanism of speciation wasanagenesis, not cladogenesis—just the long-term gradualevolution within a single line rather than the splitting of lines.Interestingly, the later Kuhn will opt for cladogenesis.
As many commentators have pointed out, the theory of punctuatedequilibrium of Niles Eldredge and Stephen Jay Gould (1992) raises thequestion of evolution versus revolution, now precisely in thebiological (paleontological) context. In fact, Gould and Eldredge werethemselves influenced by Kuhn, whom Gould once described as “adear friend”; but they denied that they had deliberatelyfashioned themselves as Kuhnian revolutionaries (Gould 1997). Gouldand Eldredge end their later review article on punctuated equilibriumby remarking:
[C]ontemporary science has massively substituted notions ofindeterminacy, historical contingency, chaos and punctuation forprevious convictions about gradual, progressive, predictabledeterminism. These transitions have occurred in field after field;Kuhn’s celebrated notion of scientific revolutions is, forexample, a punctuation theory for the history of scientific ideas.[1993, 227]
Here, as with Kuhn, those who resist Gould’s attempt to sound sorevolutionary as to be contrary to Darwin’s phyletic gradualismnote that it is only on a geological timescale that such developmentsas the Cambrian explosion appear to be episodic. When examined on thetimescale of the biological generations of the life forms in question,the development is evolutionary—more rapid evolution than duringother periods, to be sure, but still evolutionary.
Stuart Kauffman (1993) and Brian Goodwin (1994) defendedreorganization in the form of self-organization as the primarymacro-biological mechanism, with evolutionary adaptation adding onlythe finishing touches. Gould and Richard Lewontin had raised thispossibility in their famous paper of 1979, “The Spandrels of SanMarco and the Panglossian Paradigm.” Applied to the developmentof science, this view implies that revolutions determine the overallshape, while ordinary scientific work applies the adaptivemicroevolution. Meanwhile, Michael Ruse (1989) defended the view thatthe Darwinian paradigm (with its emphasis on function and adaptation)and the punctuated equilibrium paradigm (with its emphasis on Germanicideas of form and internal constraints) are complementary.
David Hull ended his book, Science as a Process (1988), withthe remark that the book can be regarded as an attempt to fulfill bothKuhn’s and Toulmin’s ambitions to provide a evolutionaryaccount of scientific development. Hull’s is the mostthoroughgoing attempt to date to provide an evolutionary account ofscientific practice, at least in a specific field. However, nothinglike Kuhnian paradigms, normal science, and revolutions are to befound in Hull’s book, and this was deliberate on his part.Whereas Kuhn originally said that paradigms correspond one-to-one withscientific communities, Hull rejected Kuhn’s idea of ascientific community as too loose. A scientific community, he said,does not consist of people who merely happen to agree on certainthings (anymore than the members of a species are individuals whohappen to share a set of traits). Mere consensus is not enough.Rather, communities are tightly causally linked in the right sorts ofways, just as species are. There is no community of biologists or evenof evolutionary biologists but only a patchwork of cliques. It ishere, locally, that the seeds of innovation are sown, most of whichare weeded out in a selective process by the larger group ofspecialists. Hull’s is a story of the socio-cultural evolutionof science without revolution.
In “The Road Since Structure,” his 1990 PresidentialAddress to the Philosophy of Science Association, Kuhn reported on abook in progress, a project that would remain unfinished at his death.(See also Kuhn 1993.) In this and other fragments of that work, hedevelops the biological metaphor broached at the end ofStructure. He retains his old parallel to biologicalevolution, that science progresses or evolves away from its previousforms rather than toward a final truth about the world; but he nowextends the biological analogy by regarding scientific specialtiesthemselves as akin to biological species that carve out research andteaching niches for themselves. In the process, he significantlymodifies his conception of scientific revolutions and attendant claimsconcerning crises and incommensurable breaks. No longer do we hear ofrevolutions as paradigm change, certainly not in the sense of largeparadigms. In fact, Kuhn prefers to speak of “developmentalepisodes” instead of revolutions. However, he does retainsomething of his original idea of small paradigms, the concreteproblem solutions that he had termed “exemplars” in the“Postscript-1969” to Structure. Most revolutions,he tells us, are not major discontinuities in which a successor theoryoverturns and replaces its predecessor. Rather, they are like(allopatric) biological speciation, in which a group of organismsbecomes reproductively isolated from the main population.
[R]evolutions, which produce new divisions between fields inscientific development, are much like episodes of speciation inbiological evolution. The biological parallel to revolutionary changeis not mutation, as I thought for many years, but speciation. And theproblems presented by speciation (e.g., the difficulty in identifyingan episode of speciation until some time after it has occurred, andthe impossibility even then, of dating the time of its occurrence) arevery similar to those presented by revolutionary change and by theemergence and individuation of new scientific specialties.…
Though I greet the thought with mixed feelings, I am increasinglypersuaded that the limited range of possible partners for fruitfulintercourse is the essential precondition for what is known asprogress in both biological development and the development ofknowledge. When I suggested earlier that incommensurability, properlyunderstood, could reveal the source of the cognitive bite andauthority of the sciences, its role as an isolating mechanism wasprerequisite to the topic I had principally in mind. … [Kuhn2000, 98–99]
In short, specialization is speciation, a scientific progressheightens communication breakdown. Experts doing similar kinds ofresearch come to realize that their use of key taxonomic terms nolonger jibes with mainline usage, in the sense that what Kuhn calls“the no overlap principle” is violated: the group is usinga taxonomic hierarchy for crucial kind terms and the associatedcategories that is incompatible with that of the establishedtradition. The group splits off and forms a distinct specialty withits own professional journals, conferences, etc., while leaving therest of the field largely intact. The incommensurability is now alocal, community-licensed, taxonomic one that creates something of abarrier to communication with neighboring specialties. One thinks, forexample, of the way different biological specialties employ thespecies concept itself, and the concept of gene. This linguisticsensitivity as a group identifier permits the kind of fullness ofcommunication, both linguistic and practical, within the group thatKuhn had stressed already in Structure and thus permits thegroup to progress more rapidly. This multiplication of specialties isthe key to Kuhn’s new conception of scientific progress. Tworecent books that directly engage these issues are Andersen et al.(2006) and Wray (2011), the first from a cognitive science point ofview, the second with emphasis on Kuhn’s social epistemology.See also Nersessian (2003, 2008) and Kuukkanen (2012).
Another striking fact about Kuhn’s last project is the demotionto “a historical perspective” of the history of science asa detailed source of data or phenomena upon which philosophers ofscience should draw. While Structure was already a curiousmix of an inductive, history-based, bottom-up approach to modelingscientific development and a more formal, top-down approach based onKuhn’s Kantianism (see below), he now concludes that his coreposition follows from first principles alone.
With much reluctance I have increasingly come to feel that thisprocess of specialization, with its consequent limitation oncommunication and community, is inescapable, a consequence of firstprinciples. Specialization and the narrowing of the range of expertisenow look to me like the necessary price of increasingly powerfulcognitive tools. [2000, p. 98]
Thus Kuhn’s new emphasis is on synchronic revolutions, in whicha field splits into subfields, rather than on the diachronicreplacement of one paradigm complex by another that his early accountfeatured. His conception of a science is therefore less monolithic. Avibrant field such as evolutionary biology can tolerate severaldistinct species concepts at the same time, a fact that contributesrather than detracts from its vibrancy. The overall result is a lesstightly integrated, less dogmatic conception of normal science underan overarching paradigm, a view that has implications also for thenecessity and size of future revolutions. For no longer need anesoteric discrepancy get the leverage to trigger a crisis thateventuates in the replacement of an entire, tightly integrated system.In this respect, Kuhn’s conception of physics becomes somewhatcloser to Godfrey-Smith’s characterization of progress inbiology as “a deluge” rather than a full-scale Kuhnianrevolution (see below and Godfrey-Smith 2007). Given that progress inbiological evolution is better regarded as the remarkableproliferation of intricate, useful design rather than movement towarda goal, the explicit parallels that Kuhn draws to biological evolutionsuggest that he is moving toward the same conception of scientificprogress as some see in biological evolution—as theproliferation of adaptive design. We may know more about his finalposition once more of the book manuscript, left incomplete at hisdeath, is published.
We pass now from the smaller-scale revolutions of Kuhn’s laterwork to large-scale movements that, in several cases, exceed thebounds of the early Kuhnian paradigms and the revolutionarytransformations linking them. Other thinkers have gone even furtherthan Kuhn, by positing the existence of cognitive formations that areboth broader and deeper than his. One prominent line of thought hereis the neo-Kantian one up through Reichenbach and Carnap, discussedand further developed by Michael Friedman (2001, 2003). Another, notentirely distinct, idea is that of a thought style or discursiveformation found variously in such writers as Ludwik Fleck (1935),Alistair Crombie (1994), Michel Foucault (1969), and Ian Hacking(1990, 2002, 2012). Especially in the accounts of Foucault andHacking, new conceptual spaces are constructed and may“crystallize” (Hacking) rapidly. Once they becomecanonical, they seem to be such obvious frameworks for making true orfalse claims that the corresponding categories of thought and actionappear to be given as part of the nature of things, aswritten in the language of nature, so to speak, when they are in facta product of the cultural conditioning of our socio-cognitive systems.In the limit we project our deeply ingrained cultural categories notonly onto our world as we encounter it but also onto all(historically) conceivable worlds. The historical change in question,once called to our attention, seems revolutionary—in a mannerthat is both broader and deeper than the transition to a new paradigmwithin a particular scientific specialty. Once again, the magnitude ofthe change is practically invisible to all but the most sensitivearcheologist of knowledge. Feyerabend was alive to this perspective inhis work on Galileo. But, unlike his treatment of the CopernicanRevolution (Kuhn 1957), Kuhn’s revolutions in Structureand beyond are more limited in scope, typically occurring more or lesswholly within a single discipline. Nor is it obvious that theemergence of a new thought style must overturn a distinct predecessor.Of course, we should not regard social constructionist /deconstructionist projects (whether or not deliberately designed) as,automatically, ongoing, enlightened processes that“unfreeze” the stones inherited from the past; for it wasthese very processes that created essentialist constructions in thefirst place. The claim is that our constructions today are nodifferent.
The four main characteristics of Hacking’s “bigrevolutions” are that they lead to whole new interdisciplinarycomplexes, replete with new social institutions and correspondingsocial changes, and they alter that period’s overall take on, or“feel” for, the universe. For critical discussion ofHacking on styles of reasoning, see Kusch (2010) and Scortino (2016).For more on Hacking, see section 5.3 below.
Kuhn several times described himself as “a Kantian with moveablecategories.” Hoyningen-Huene (1993) provides a broadly Kantianinterpretation of Kuhn (endorsed by Kuhn himself), as does Friedman(2001, 2003, 2010). Given the historical approach ofStructure, other commentators have likened Kuhn to Hegelinstead of Kant. And given the early Kuhn’s view that scientificreason is manifested more clearly in historical change as well as innormal scientific practices than in symbolic logical structures,Kuhn’s early theory of scientific change can be termed verybroadly Hegelian. Also fitting the broadly Hegelian frame isKuhn’s internalist account of normal scientific research assowing the seeds of its own destruction through unintended innovation,resulting eventually in a kind of dialectical conflict that drives theenterprise forward. However, there is no Hegelian Spirit lurkingwithin Kuhn’s model, nor a permanent logic of science as itsreplacement as there was for Lakatos (Lakatos 1970;Hoyningen-Hühne 1993; Bird 2008; Worrall 2003, 2008; Nickles2009).
Kuhn disliked being compared to Hegel, whose work he found obscure andcharacterized by a non-naturalistic philosophy of history, but it isworth commenting further on the partial resemblance. Both Kant andHegel rejected naïve empiricism, according to which all humanknowledge arises, somehow, from the accumulative aggregation ofindividual sensations or perceptions. Kant argued that we needtranscendental structures such as a system of processing rules inorder to organize sensory input into something coherent andintelligible, e.g., as physical objects interacting causally in spaceand time. These were Kant’s forms of intuition (in space andtime) and categories (substance, causality, etc.). They represent thehuman mind’s contribution to knowledge. (In this regard Kant canbe regarded as a forerunner of cognitive psychology.) Thus, ourexperience of the world is shaped to fit a priori forms, and this iswhere the Kantian version of idealism enters the picture: the world ofhuman experience is not the world of ultimate reality (Kant’sunknowable, noumenal world of things-in-themselves); rather, it is aworld shaped by our own cognitive structures.
Hegel, one of the founders of the deep conception of historical changebroadly characteristic of nineteenth-century German scholarship,proceeded to historicize Kant’s innovation, in effect byhistoricizing Kant’s categories. They are not inborn, permanent,and universal; on the contrary, they are socio-historically acquiredor lost and hence differ from one historical epoch to another. Peopleliving in different epochs cognize the world differently. It istempting to read the Kuhn of Structure as furtherrelativizing and localizing Hegel to specific scientific domains andtheir paradigms. Kuhn’s model provides endogenous, dynamicmechanisms of (inadvertent) scientific innovations that sow the seedsof the paradigm’s eventual destruction in a vaguely Hegeliansort of dialectical process. It thus becomes possible to experience achange in categorical scheme within one’s own lifetime—thevictory of the new paradigm being the small-scale scientificcounterpart of Hegel watching Napoleon march in victory through thestreets of Jena! Thus it is tempting to regard Kuhnian revolutions asHegelian revolutions writ small.
Nonetheless, in terms of historical genealogy, Kuhn is better alignedwith the Kantian tradition, especially the neo-Kantian relativizationof Kant. Interestingly, some logical empiricists (especiallyReichenbach) were influenced by the neo-Kantianism of the GermanMarburg School of philosophy to develop a historically relativized butconstitutive a priori (see below and Friedman 2001.)
There is a long historical gap between Kant/Hegel and Kuhn, and thisspace is not empty. Many other thinkers, especially those in thevarious nineteenth-century idealist traditions, were Kantian orHegelian or neo-Hegelian or neo-Kantian opponents of empiricistpositions that they considered naïve, such as that of John StuartMill. The neo-Kantian label applies even to prominent logicalpositivists of the Vienna Circle and logical empiricists of the BerlinCircle, who have too often been caricatured as simple, cumulativeempiricists. As Friedman (2001) and others have shown, severalfounders of twentieth-century academic philosophy of science extendedthe neo-Kantian attack on simple empiricism. The basic idea here isthat, just as Kant regarded any account of perception and knowledge asnaïvely empiricist insofar as it left no room for underlyingcognitive organizing principles, so any account of the sciences thatprovided no analogous underlying social-cognitive framework was acontinuation of simple empiricism, i.e., a version of“positivism.” In this particular respect, W. V.Quine’s “Two Dogmas of Empiricism” and Word andObject (Quine 1951, 1960) were throwbacks to simple empiricism intheir attempt to eliminate the Kantian formal component.
The German Marburg School of Hermann Cohen, Paul Natorp, and ErnstCassirer was especially important in the emergence of modernphilosophy of science in the form of the logical positivism andlogical empiricism. Friedman (2001 and elsewhere) has explored itsinfluence on young Reichenbach’s attempt to interpret thesignificance of the new relativity theory. Rudolf Carnap had beeninfluenced by Ernst Cassirer, among others. (See the entries onlogical empiricism, Reichenbach, Carnap, Cohen, Natorp, and Cassirer.)Cassirer’s central theme was the fundamental epistemologicalimplications of the replacement of Aristotle’s subject-predicatelogic and substance ontology by the new relational logic of themodern, mathematical-functional approach to nature. This could not bea priori in Kant’s original sense, since the emergence ofnon-Euclidean geometry had shown that there are alternative organizingprinciples. But the very fact that we still needed organizingstructures that are constitutive or definitive of the cognitiveenterprise in question meant that Kant was still basically correct.Like Moritz Schlick, the first leader of the Vienna Circle,Reichenbach of the Berlin school parlayed his engagement withrelativity theory and non-Euclidean geometries into a conception ofthe “relativized a priori.” All of this by around 1920,with the younger Carnap’s views developing during the 1920s and‘30s. In the USA, meanwhile, C. I. Lewis (1929) was defendinghis “pragmatic a priori.”
In his famous paper of 1950, “Empiricism, Semantics, andOntology,” Carnap made his two-tiered view of inquiry quiteexplicit. Starting from the problem of the existence of abstractentities, Carnap distinguished internal questions, that is, questionsthat can arise and be answered within a particular logico-linguisticframework, from external questions, that is, meta-level questionsabout which framework to prefer. External questions cannot be answeredin the same, disciplined manner as internal, for choice of frameworkis ultimately a pragmatic decision based on the expected fertility ofusing one framework rather than another. Although it is difficult toequate Carnap’s fruitfulness decisions with Kuhn’srevolutionary breaks (Kuhn 1993, 313f), Carnap regarded his vision ofscience as similar to Kuhn’s, and he liked the manuscript ofStructure that Kuhn submitted to the InternationalEncyclopedia of Unified Science, as George Reisch (1991) hasshown. Thus the clash of Kuhn’s work with that of Carnap and thepositivist movement has been exaggerated.
Although both defended two-tiered conceptions of inquiry, there areimportant differences between Kuhn and Carnap (as Friedman, 2001,2003, 2010, among others, observes). For Carnap, as for Reichenbach,the choice of framework or coordinating definitions was conventional,a matter of convenience or heuristic fertility, whereas for committedKuhnian normal scientists the foundational tenets of their paradigmare deep truths about the world, principles not subject to empiricaltest. (However, in a crisis situation, fertility becomes a key elementin theory and paradigm choice.) In Carnap’s frameworks thesewere explicit systems of logical rules, whereas Kuhn’s accountof normal science largely jettisoned rule-based knowledge in favor ofa kind of case-based tacit knowledge, the cases being the concreteexemplars. Third, Kuhn himself emphasized that his approach washistorical, whereas Carnap’s was not. Although a Carnapianchange from one logical framework to another could, in principle, bequite revolutionary, Carnap himself never emphasized this point,suggested nothing of Kuhn’s radical discontinuity, and wassimply not interested in the history of science.
Meanwhile, Friedman himself has extensively developed the idea ofhistorically contingent but constitutive a prioris (e.g., 2001, 2003,2008). He is sympathetic to Kuhn’s view that revolutions occurwhen the constitutive principles change. From the old point of view,there is disruptive and incommensurability, but defenders of the newviewpoint manages to establish a kind of continuity. Friedman goeswell beyond Kuhn in stressing the role of philosophical ideas inestablishing this continuity.
Since deep conceptual revolutions or paradigm-shifts are a fact ofscientific life (and, I would argue, a necessity), we are never in aposition to make our present constitutive principles as trulyuniversal principles of human reason–as fixed once and for allthroughout the evolution of science.
In recent work, Friedman devotes more attention to thesocial dimension, and he notes that even the standards of rationalitymay continue to change historically. (See the entry “HistoricalTheories of Rationality”. See also DiSalle 2002.)
Another development that appeals to Germanic themes in its criticismof naive empiricism is the idealization movement of the PoznańSchool in Poland, associated especially with Leszek Nowak. This groupregards science as developing in an idealizational and dialecticalmanner, ideas that they trace back to Karl Marx’s analysis ofeconomics, inspired by his own study of Galileo’s fruitful useof abstract, idealized models against the Aristotelians. ThePoznań group regards idealization as the secret to modern scienceand finds it remarkable that virtually all previous analyticphilosophy of science remains Aristotelian in treating proposed lawsand theories not as ideal models but as true or false statementsdirectly about the real world (Nowak 1980 and later writings; see alsoKrajewsky 1977). As models, these constructions must be concretized tosome degree before they can be applied to the real world. While theidealizationists tend to reject Kuhnian revolutions as toodiscontinuous and irrational, they do see a resemblance to theirinternalistic, dialectical conception of scientific development. Forthem a revolution consists in a new theory or model that reveals apreviously unnoticed idealizing assumption built into its predecessor,a change that alters scientists’ conception of what is essentialversus peripheral to the domain of phenomena in question. Hence therecan be a significant change of world-conception.
There is some affiliation of Poznań with the Europeanstructuralist account of theories, based on a set-theoretical analysisof theory structure and theory relations. Kuhn himself was muchattracted to Joseph Sneed’s approach (Sneed 1971), soon extendedby Wolfgang Stegmüller (1974/1976) and others. Given theinformality of Kuhn’s own approach and his explicit shunning ofrules and rational reconstructions, his attraction to thestructuralist line was initially puzzling. However, the structuralistswere and are interested in intertheory relations, and models arecentral to their non-sentential conception of theories. These aremodels in the formal sense, but Kuhn found insightful connections tohis own use of models in the form of exemplars. For both Kuhn and thestructuralists it is the collection of exemplars or models, not anabstract statement of a theory, that carries the weight in scientificinquiry.
Already the early Kuhn, especially in the postscript to the secondedition of Structure, largely abandoned the traditionalconception of theories as deductive systems, even in physics, andsubstituted informal collections of models of various, exemplarykinds, along with a toolbox of expert practices for constructing andapplying them (Cartwright 1983, Giere 1988, Teller 2008). He alwaysliked Margaret Masterson’s remark that “a paradigm [in thesense of preferred models or exemplars] is what you use when thetheory is not there” (Baltas et al. 2000, 300). Such a view,like Nowak’s, anticipates the move from theory-centered tomodel-centered accounts of scientific work. However, Kuhn’snormal scientific practitioners presumably hold the models to be truein their original application, as is the grand theory incorporated inthe paradigm, whereas today the emphasis is on modeling practicesacross the sciences, in which the models are almost always known inadvance to be false because of their employment of idealizations,approximations, abstractions, etc.
Meanwhile, important French thinkers had already taken a historicalapproach, one that explicitly characterizes science as a series ofbreaks or coupures. The principal genealogy includesLéon Brunschvicg, Gaston Bachelard and his student, GeorgesCanguilhem, and the latter’s student, Michel Foucault. NeitherKuhn’s historicism nor his talk of revolutionary breaks was newsto the French (Gutting, 2001, 2003). The French tradition of sciencestudies, going back to Auguste Comte and including later figures suchas Pierre Duhem and Henri Poincaré, possessed a historicaldimension that positivism lost after Ernst Mach, as it became logicalpositivism. However, the French and Germanic traditions have someroots in common. As Gutting points out, Brunschvicg, like ÉmileMeyerson, was a science-oriented idealist. For him the mind is not apassive wax tablet; rather, it actively forges internal links amongideas, yet it is also often surprised by the resistant exteriority ofthe natural world. Against traditional metaphysics, philosophy ofscience should limit itself to what the science of the timeallows—but not dogmatically so. For dogmatic attempts to extracttimeless principles and limitations (such as Kant’s denial ofthe possibility of non-Euclidean geometry) may soon be embarrassed byfurther scientific advances. Einstein’s general theory ofrelativity exemplifies the revolutionary nature of the most impressivedevelopments.
Bachelard, French physicist and philosopher-historian of science, alsobelieved that only by studying history of science can we gain anadequate understanding of human reason. He stressed the importance ofepistemological breaks or discontinuities (coupuresépistémologiques). In Le Nouvel EspritScientifique (1934), Bachelard argued that the worldview ofclassical physics, valuable in its own time, eventually became anobstacle to future progress in physics. Hence a break was needed.Here, then, we already find the idea that a successful theory can loseits luster by being considered exhausted of its resources and thuslacking in fertility. Like Brunschvicg, Bachelard held that adefensible, realist philosophy had to be based on the science of itsday. Hence, scientific revolutions have (and ought to have) broughtabout epistemological revolutions. The reality we posit, he said,ought to be that of the best science but with the realization that ourconcepts are active constructs of our own minds, not imported fromnature’s own language, as it were. Future mental activity aswell as future empirical findings are likely to require anotherrupture. As Gutting points out, Bachelard’s account ofdiscontinuity was not as radical as Kuhn’s. Bachelard waswilling to speak of progress toward the truth. He made much of thefact that successor frameworks, such as non-Euclidean geometry orquantum physics, retain key predecessor results as special cases and,in effect, contextualize them.
Canguilhem was more interested in the biological and health sciencesthan Bachelard and gave great attention to the distinction between thenormal and the pathological, a distinction that does not arise inphysical science. For this and other reasons, in his view, we canexpect no reduction of biology to physics. Canguilhem provided a morenuanced conception of obstacles and ruptures, noting, for example,that an approach such as vitalism that constitutes an obstacle in onedomain of research can simultaneously play a positive role elsewhere,as in helping biological scientists to resist reductive thinking. Herewe find context sensitivities and heuristic resources difficult tocapture in terms of a context- and content-neutral logic of sciencesuch as the logical empiricists espoused.
Bachelard and Canguilhem also had less disruptive conceptions ofscientific objectivity and scientific closure than Kuhn. Canguilhemcriticized Kuhn’s (alleged) view that rational closure could notamount to more than group consensus. Both Frenchmen emphasized theimportance of norms and denied that disciplinary agreement was as weakas Kuhnian consensus. Kuhn replied to this sort of objection (in“Postscript” and elsewhere) that his scientificcommunities do possess shared values, that their agreement is notsomething arbitrary, say, as whipped up by political ideologues.
Foucault’s archaeology of knowledge (Foucault 1966, 1969) positsa distinction between a superstructure of deliberately madeobservations, claims, and arguments and a deep structure, mostelements of which we are probably unconscious. Once again we meet atwo-level account. Writes Hacking:
Foucault used the French world connaissance to stand for suchitems of surface knowledge while savoir meant more thanscience; it was a frame, postulated by Foucault, within which surfacehypotheses got their sense. Savoir is not knowledge in thesense of a bunch of solid propositions. This “depth”knowledge is more like a postulated set of rules that determine whatkinds of sentences are going to count as true or false in some domain.The kinds of things to be said about the brain in 1780 are not thekinds of things to be said a quarter-century later. That is notbecause we have different beliefs about brains, but because“brain” denotes a new kind of object in the laterdiscourse, and occurs in different sorts of sentences. [2002, 77]
Given the influence of Foucault, we may also locate our discussion ofHacking’s own work on historical ontology here (Hacking 2002).Hacking (1975, 1990, 1995, 2012) has studied in depth the emergence ofprobability theory and (later) of statistical thinking and theconstruction of the modern self as key examples of what he terms“historical ontology.” He acknowledges inspiration fromboth Foucault’s discursive formations and Crombie’s stylesof thinking (Crombie 1994), with a dose of Feyerabend thrown into themix. Like Kuhn (and Friedman), Hacking returns to Kant’s“how possible?” question, the answer to which sets out thenecessary conditions for a logical space of reasons in whichpractitioners can make true or false claims about objects and poseresearch questions about them. Hacking, too, historicizes the Kantianconception. He likes the term ‘historical a priori’, whichCanguilhem once applied to the work of his erstwhile student,Foucault.
The historical a priori points at conditions whose dominionis as inexorable, there and then, as Kant’s synthetic apriori. Yet they are at the same time conditioned and formed inhistory, and can be uprooted by later, radical, historicaltransformations. T. S. Kuhn’s paradigms have some of thecharacter of a historical a priori. [Hacking 2002, 5]
…See AlsoTélécharger Vcredist_x86.exe et réparer les erreurs EXEMen's Dress Code GuideWhat to Wear to Work
[S]cientific styles of thinking & doing are not goodbecause they find out the truth. They have become part of ourstandards for what it is, to find out the truth. They establishcriteria of truthfulness. … Scientific reason, as manifested inCrombie’s six genres of inquiry, has no foundation. The stylesare how we reason in the sciences. To say that these stylesof thinking & doing are self-authenticating is to say that theyare autonomous: they do not answer to some other, higher, or deeper,standard of truth and reason than their own. To repeat: No foundation.The style does not answer to some external canon of truth independentof itself. [2012, 605; Hacking’s emphasis]
Hacking describes changes in historical a prioris as“significant singularities during which the coordinates of‘scientific objectivity’ are rearranged” (2002,6).
Although reminiscent of Kuhn’s positions in some ways, there arestriking differences. As noted above, Hacking’s constructedformations are much broader than Kuhn’s. Thus he feels free toemploy telling bits of popular culture in laying out his claims, andhe admits to being whiggish in starting from the present and workingbackward to find out how we got here. Moreover, in mature, modernscience, unlike Kuhnian paradigms, several of Hacking’s stylesof thinking and doing can exist side by side, e.g., the laboratory andhypothetical modeling traditions. Yet people living before and afterthe historical crystallization of a style would find each othermutually unintelligible. Hacking recognizes that Kuhnian problems ofrelativism (rather than subjectivism) lurk in such positions.“Just as statistical reasons had no force for the Greeks, so oneimagines a people for whom none of our reasons for belief haveforce” (2002, 163). This sort of incommensurability is closer toFeyerabend’s extreme cases (as in the ancient Greek astronomersversus their Homeric predecessors) than to Kuhn’s “nocommon measure” (2002, chap. 11). This sort of unintelligibilityruns deeper than a Kuhnian translation failure. It is not a questionof determining which old style statements match presumed new styletruths; rather, it is a question of the conditions for an utterance tomake a claim that is either true or false at all. WritesHacking,
Many of the recent but already classical philosophical discussions ofsuch topics as incommensurability, indeterminacy of translation, andconceptual schemes seem to discuss truth where they ought to beconsidering truth-or-falsehood. [2002, 160]
By contrast, Kuhnian paradigms include a set of positive assertionsabout the world. Yet Kuhn himself was attracted by Hacking’s wayof putting the point about truth-and-falsity (Kuhn 2000, p. 99).
To what extent was Kuhn indebted to these thinkers? As noted above, hetook Kant but not Hegel very seriously. He was largely self-taught inphilosophy of science. Among his contemporaries, he was familiar withPopper but not in any detail with the various strains of logicalpositivism and logical empiricism, in particular the positions ofCarnap and Reichenbach. Apparently, he was only slightly acquaintedwith the work of Bachelard while writing Structure, and theynever engaged in a fruitful interchange (Baltas et al. 2000, 284f).Kuhn did acknowledge, in print and in his classes, the crucialinfluence on his historical and philosophical thinking of the twoRussian émigrés, Émile Meyerson, author ofIdentité et Realité (1908) and AlexandreKoyré, especially his Études Galiléenes(1939), and that of Annaliese Maier, the German historian of medievaland early modern science. He had read Ludwik Fleck’s Genesisand Development of a Scientific Fact (originally published inGerman in 1935) and Michael Polanyi’s PersonalKnowledge (1958) and had had some discussion with Polanyi (Baltaset al. 2000, 296). Kuhn was also indebted to Wittgenstein, early(“The limits of my language are the limits of my world,”1922, 148) and late (on language games and forms of life). (SeeSharrock and Read 2002, the introduction to Harris 2005, and Kindi2010 for Kuhn’s relation to Wittgenstein and others.) He knewsomething of Toulmin’s work.
Kuhn more than anyone in the Anglo-American world pointed out the needfor larger-sized units than individual theories in making sense ofmodern science. Nonetheless, as we have seen, others in the Teutonicand Francophone worlds had previously postulated even largersocio-intellectual units and correspondingly deeper changes thanKuhn’s, on somewhat different scales of intellectual space andtime. If we think of authors such as the Annales historianFernand Braudel, with his distinct time-scales, we recognize that theattribution of transformative change clearly depends heavily on thechoice of time-scale and on how fine- or course-grained is ourapproach. Hacking (2002, 76) makes this point with reference to theFrench context:
There are two extremes in French historiography. The Annalesschool went in for long-term continuities or slowtransitions—“the great silent motionless bases thattraditional history has covered with a thick layer of events”(to quote from the first page of Foucault’s 1969 Archeologyof Knowledge). Foucault takes the opposite tack, inherited fromGaston Bachelard, Georges Canguilhem, and Louis Althusser. He positssharp discontinuities in the history of knowledge.
Although Kuhn emphasized the importance of skilled scientificpractice, his paradigms remained closer to the articulate surface ofscientific culture than Foucault’s discursive formations, whichare better located in the unconscious than in the Kuhniansubconscious. Foucault does not speak of revolution.
Oliver Wendell Holmes, Jr. (1861) remarked that “Revolutionsnever follow precedents nor furnish them.” Given theunpredictability, the nonlinearity, the seeming uniqueness ofrevolutions, whether political or scientific, it is thereforesurprising to find Thomas Kuhn attempting to provide a General Theoryof Scientific Revolutions (Kindi 2005). Early Kuhn did seem to believethat there is a single, underlying pattern to the development ofmature sciences that is key to their success, and late Kuhn adifferent pattern. Has either early or late Kuhn found such a pattern,or has he imposed his own philosophical structure on the vagaries andvicissitudes of history? Kuhn’s Kantianism always did live intension with his historicism, and in his late work (e.g., 2000c) hesurprisingly gave up the pretense of deriving his pattern of taxonomicchange and speciation from history of science, on the ground that itlargely followed “from first principles.”
Numerous philosophers, scientists, and other commentators have madeclaims about scientific change that differ from Kuhn’s. (For arecent selection see Soler et al. 2008.) Some, as we have seen, areskeptical of revolution talk altogether, others of Kuhn’s inparticular. Still others accept that some revolutions are Kuhnian butdeny that all of them are. One common criticism is that not allrevolutionary advances are preceded by an acute crisis, that is, bymajor failures of preceding research. Kuhn himself allowed forexceptions already in Structure. Another is thatrevolutionary changes need not involve discontinuities in all ofKuhn’s levels at once (especially Laudan 1984). Yet another isthat there need be little logical or linguistic discontinuity. Arapid, seemingly transformative change in research practices mayinvolve simply a marked gain in data accessibility or accuracy orcomputational processing ability via new instrumentation orexperimental design. And on the later Kuhn’s own view,revolution need not be a game of creative destruction. Only a fewexamples can be considered here.
Do revolutions consist, according to Kuhn, of major new materials(experimental facts, theories, models, instruments, techniques)entering a scientific domain or, instead, of a major restructuring orrearrangement of materials, practices, and community affiliationsalready present? Kuhn states that the relativity revolution mightserve as
a prototype for revolutionary reorientation in the sciences. Justbecause it did not involve the introduction of additional objects orconcepts, the transition from Newtonian to Einsteinian mechanicsillustrates with particular clarity the scientific revolution as adisplacement of the conceptual network through which scientists viewthe world. [1970, 102]
The reader may find this claim confusing, however, because in thejust-preceding paragraphs Kuhn had emphasized the ontological andconceptual changes of precisely this revolution, e.g., the radicalchange in the concept of mass. Einstein’s masses are notNewtonian masses, he insisted. They are newly introduced entities;hence, we may infer, new content. Yet Kuhn surely does have a pointworth saving, in that relativity theory still deals with most of thesame kinds of phenomena and problems as classical mechanics andemploys immediate successors to the classical concepts. But, if so,then reorganization of familiar materials implies a disciplinarycontinuity through revolution that Kuhn minimized.
That reorganization dominates Kuhn’s conception of revolutionsis apparent throughout his work. As a young scholar he had an epiphanywhen Aristotle’s seemingly radically false or unintelligibleclaims suddenly came together for him as a coherent, comprehensiveworldview. This experience became Kuhn’s psychological model forrevolutionary transformation from one paradigm to its successor andinformed his later talk of Gestalt switches. But he also emphasizedthat revolution involves social reorganization of the field(not merely the cognitive reorganization of an individual), from oneform of scientific life to another, incompatible with it. Byimplication, his structural or formal conception of revolutionexcluded the alternative idea of revolution as extraordinary bursts insubstantive content.
In Conceptual Revolutions, Paul Thagard (1992) retainssomething of Kuhn’s idea of conceptual transformation and themore specific idea of taxonomic transformation. He distinguishes twokinds of reclassification, in terms of the language of tree structuresused in computer science: branch jumping and tree switching. Branchjumping reclassifies or relocates something to another branch of thesame tree, e.g., reclassifying the whale as a mammal rather than afish, the earth as a planet, or Brownian motion as a physical ratherthan a biological phenomenon. New branches can appear and old branchescan be eliminated. Meanwhile, tree switching replaces an entireclassification tree by a different tree structure based on differentprinciples of classification, as when Darwin replaced the staticclassification tree of Linnaeus by one based on evolutionary genealogyand when Mendeleev replaced alternative classification systems of thechemical elements by his own table. Taking a computational approach tophilosophy of science, Thagard employs his computer program ECHO toreconstruct and evaluate several historical cases of allegedconceptual revolution and arrives at a tamer conception ofrevolutionary breaks than Kuhn’s.
The Cognitive Structure of Scientific Revolutions by HanneAndersen, Peter Barker, and Xiang Chen (2006) also devotes a good dealof attention to cognition and categorization issues, in a defense ofthe later Kuhn’s approach. The work of cognitive psychologistLawrence Barsalou and of philosopher-historian Nancy Nersessian (thefounder of the “cognitive historical” approach to science)plays a significant role in their account. Nersessian herself (2003,2008) emphasizes model-based reasoning. These are no longer staticcases or exemplars, for they possess an internal dynamic.
Howard Margolis (1993) distinguishes two kinds of revolutions,depending on which kinds of problems they solve. Those revolutionsthat bridge gaps, he contends, differ from those that surmount orpenetrate or somehow evade barriers. His focus is on barriers, aneglected topic even though it fits Kuhn’s account of cognitionwell. Margolis develops Kuhnian themes in terms of deeply ingrained“habits of mind.” While such habits are necessary forefficient scientific work within any specialty discipline, theyconstitute barriers to alternative conceptions. More broadly, deeplyingrained cultural habits of mind can close off opportunities that,according to the perspective of later generations, were staring themin the face. Margolis is struck by the apparent fact that all thematerials for Copernicus’ new model of the solar system had beenavailable, in scattered form, for centuries. No new gap-crossingdevelopments were needed. He concludes that, rather than a gap to bebridged, the problem was a cognitive barrier that needed to beremoved, a barrier that blocked expert mathematical astronomers frombringing together, as mutually relevant, what turned out to be thecrucial premises, and then linking them in the tight way thatCopernicus did. If Margolis’ account of the CopernicanRevolution is correct, it provides an example of revolution asholistic reorganization of available materials, hence thenon-piecemeal, noncumulative nature of revolutions. The developmentsthat lead to a barrier’s removal can be minor and, as in thecase of Copernicus, even quite peripheral to the primary subjectmatter that they ultimately help to transform. Here one thinks of amodel popular with mystery writers, where an everyday observationleads to a sudden change in perspective.
Davis Baird (2004) contends that there can be revolutions in practicethat are not conceptual revolutions. He emphasizes the knowledgeembodied in skills and in instruments themselves. His central exampleis analytic chemistry.
There is little doubt that analytical chemistry has undergone aradical change. The practice of the analyst, who now deals with large,expensive equipment, is different than it was in 1930. Moderninstrumental methods are by and large more sensitive and accurate,have lower limits of detection, and require smaller samples; differentkinds of analyses can be performed. Analytical chemistry is much lessa science of chemical separations and much more a science ofdetermining and deploying the physical properties of substances.… The revolutionary phase of Thomas S. Kuhn’s Structureof Scientific Revolutions starts with a crisis, a problem that theestablished methods of normal science cannot solve (Kuhn  1970;1996, ch. 5). There was no such crisis in analytical chemistry. Whileone might imagine that analytical chemistry underwent a change ofparadigm, there was no crisis that provoked this change. Pre-1930analytical chemists did not bemoan the inability of their chemistry tosolve certain problems. Instead, new methods were developed that couldsolve established “solved” problems, but solve thembetter: more efficiently, with smaller samples, greater sensitivity,and lower limits of detection. These changes in analytical chemistrydo not suffer from any kind of incommensurability: today, one caneasily enough understand what analytical chemists were doing in1900–although the idea that the analytical chemist is one whocan quantitatively manufacture pure chemicals is startling on firstencounter. … The transformation in analytical chemistry passesall of Cohen’s tests.
Recently, Rogier De Langhe (2012, 2014a and b, 2017) has beendeveloping a broadly Kuhnian, two-process account of science from aneconomics standpoint. Instead of doing a series of historical cases,De Langhe and colleagues are developing algorithms to detect subtlepatterns in the large citation databases now available. De Langueemploys economic arguments to illuminate such themes as the divisionof cognitive labor, models of scientific progress, andscientists’ decisions about whether to specialize or toinnovate.
The account of the dynamics of science in Structure ill fitthe rapid splitting and recombining of fields in the post-World War IIera of Big Science, as Kuhn recognized. So he excluded from hisaccount the division and recombination of already mature fields suchas happened with the emergence of biochemistry. (This exclusion istroubling, given the universal thrust of his account. It is as if Kuhnadmitted that his account applies only to a particular historicalperiod that is now largely past; yet he also wrote as if thenormal-revolutionary model would apply to mature disciplines into thelong future.) In his later work he did devote careful attention to thedivision of fields into specialties and subspecialties (see §5).However, he still gave little attention to the more-or-less reverseprocess of new fields coming into existence by combinations ofpreviously distinct fields as well as to cross- and trans-disciplinaryresearch, in which a variety of different specialists somehow succeedin working together (Galison 1997, Kellert 2008, Andersen 2013).
And what can we make, on Kuhn’s account, of the explosion ofwork in molecular biology following the Watson-Crick discovery, in1953, of the chemical structure of DNA and the development of betterlaboratory equipment and techniques? Molecular genetics quickly grewinto the very general field of molecular biology. Less than twodecades after Watson and Crick, Gunther Stent could already write inhis 1971 textbook:
How times have changed! Molecular genetics has … grown from theesoteric specialty of a small, tightly knit vanguard into anelephantine academic discipline whose basic doctrines today form partof the primary school science curriculum. [Stent 1971, ix]
There is something paradigmatic about molecular biology and alsosomething revolutionary about its rapid progress and expansion. It isnot clear how to characterize this and similar developments. Was thisa Kuhnian revolution? It did involve major social and intellectualreorganization, one that conflicted with the previous ones in somerespects but without undermining the Darwinian paradigm. Quite thecontrary. Or is molecular biology more like a style of scientificpractice than a paradigm? Such an explosive development as molecularbiology hardly fits Kuhn’s description of steady, normalscientific articulation of the new paradigm by puzzle solving.Instead, it seems better to regard it as a large toolkit of methods ortechniques applicable to several specialty fields rather than as anintegrative theory-framework within one field.
Should we then focus on practices rather than on integrative theoriesin our interpretation of Kuhnian paradigms? The trouble with this moveis that practices can also change so rapidly that it is tempting tospeak of revolutionary transformations of scientific work even thoughthere is little change in the overarching theoretical framework (seePart II of Soler et al. 2008). Moreover, as Baird (2004) points out,the rapid replacement of old practices by new is often a product ofefficiency rather than intellectual incompatibility. Why continue todo gene sequencing by hand when automated processing in now available?Replacement can also be a product of change in research style, giventhat, as Kuhn already recognized, scientific communities are culturalcommunities.
Similar points can be made about the rise of statistical physics,mentioned above in relation to Hacking’s work. (See also Brush1983 and Porter 1986.) This was an explosion of work within theclassical mechanical paradigm rather than a slow, puzzle-by-puzzlearticulation of precisely that paradigm in its own previous terms. Orwas it? For Kuhn himself recognized that modern mathematical physicsonly came into existence starting around 1850 and that Maxwellianelectrodynamics was a major departure from the strictly Newtonianparadigm. In any case, there was much resistance among physicists tothe new style of reasoning. The kinetic theory of gases quickly grewinto statistical mechanics, which leapt the boundaries of its initialspecialty field. New genres as well as new styles ofmathematical-physical thinking quickly replaced old—anddisplaced the old generation of practitioners. Yet on Kuhn’sofficial theory of science it was all just “classicalmechanics.”
Furthermore, the biological and chemical sciences do not readilyinvite a Kuhnian analysis, given the usual, theory-centeredinterpretation of Kuhn. For biological fields rarely produce lawfultheories of the kind supposedly found in physics. Indeed, it iscontroversial whether there exist distinctly biological laws at all.And yet the biological sciences have advanced so rapidly that theirdevelopment cries out for the label ‘revolutionary’.
What of the emerging field of evolutionary-developmental biology(evo-devo)? It is too soon to know whether future work in thisaccelerating field will merely complete evolutionary biology ratherthan displacing it. It does seem unlikely that it will amount to acomplete, revolutionary overturning of the Darwinian paradigm. (Kuhnmight reply that the discovery of homeobox genes overturned a smallerparadigm based on the expectation that the genetic makeup of differentorders of organisms would have little in common at the relevant levelof description.) And if it complements the Darwinian paradigm, thenevo-devo is, again, surely too big and too rapidly advancing to beconsidered a mere, piecemeal, puzzle-solving articulation of thatparadigm. Based on work to date, evo-devo biologist Sean B. Carroll,for example, holds precisely the complement view—complementaryyet revolutionary:
Evo-Devo constitutes the third major act in a continuing evolutionarysynthesis. Evo-Devo has not just provided a critical missing piece ofthe Modern Synthesis—embryology—and integrated it withmolecular genetics and traditional elements such as paleontology. Thewholly unexpected nature of some of its key discoveries and theunprecedented quality and depth of evidence it has provided towardsettling previously unresolved questions bestow it with arevolutionary character. [2005, 283]
Eva Jablonka and Marion Lamb (2005) make even strongerKuhnian-revolutionary claims for evo-devo, which they see as a partialreturn to a Lamarckian perspective. It was in his review of their bookthat Godfrey-Smith (2007) suggested that recent biological progress isa deluge rather than a Kuhnian revolution.
Kuhn treated a scientific field (and perhaps science as a whole) as asystem with a far more interesting internal dynamics than eitherPopper or the logical empiricists had proposed. The famous openingparagraphs of Structure read as though Kuhn had analyzed ahistorical time series and extracted a pattern from it inductively asthe basis for his model of scientific development. The broadly cyclicnature of this pattern immediately jumps out at dynamical systemstheorists. Yet despite this perhaps promising start as an earlydynamical modeler of science, Kuhn apparently paid little attention tothe explosion of work in nonlinear dynamics that began with“chaos” theory and widened into such areas as complexadaptive systems and network theory. This is unfortunate, since thenew developments might have provided valuable tools for articulatinghis own ideas.
For example, it would appear that, as Kuhnian normal science becomesmore robust in the sense of closing gaps, tightening connections, andthereby achieving multiple lines of derivation and hence mutualreinforcement of many results. However, that very fact makes normalscience increasingly fragile, less resilient to shocks, and morevulnerable to cascading failure (Nickles 2008). Kuhn claimed, contraryto the expectations of scientific realists, that there would be no endto scientific revolutions in ongoing, mature sciences, with no reasonto believe that such revolutions would gradually diminish in size asthese sciences continued to mature. But it would seem to follow fromhis model that he could have made a still stronger point. ForKuhn’s position in Structure arguably implies that,when considering a single field over time, future revolutions canoccasionally be even larger than before. The reason is that justmentioned: as research continues filling gaps and further articulatingthe paradigm, normal science becomes more tightly integrated but alsoforges tighter links to relevant neighboring fields. Taking thesedevelopments into account predicts that Kuhnian normal science shouldevolve toward an ever more critical state in which something that wasonce an innocuous anomaly can now trigger a cascade of failures(Nickles 2012a and b), sometimes rather quickly. For there will belittle slack left to absorb such discrepancies. If so, then we have animportant sort of dynamical nonlinearity even in normal science, whichmeans that Kuhnian normal science itself is more dynamic, less static,than he made it out to be.
It seems clear that Kuhnian revolutions are bifurcations in thenonlinear dynamical sense, and it seems plausible to think thatKuhnian revolutions may have a fat-tailed or power-law distribution(or worse) when their size is plotted over time on an appropriatescale. Each of these features is a “hallmark of nonlineardynamics” (Hooker 2011A, 5; 2011B, 850, 858). To elaborate abit: one intriguing suggestion coming from work in nonlinear dynamicsis that scientific changes may be like earthquakes and many otherphenomena (perhaps including punctuated equilibrium events of theGould-Eldredge sort as well as mass extinction events in biology) infollowing a power-law distribution in which there are exponentiallyfewer changes of a given magnitude than the number of changes in thenext lower category. For example, there might be only one magnitude 5change (or above) for every ten magnitude 4 changes (on average overtime), as in the Gutenberg-Richter scale for earthquakes. If so, thenscientific revolutions would be scale free, meaning that largerevolutions in the future are more probable than a Gaussian normaldistribution would predict. Such a conclusion would have importantimplications for the issue of scientific realism.
To be sure, working out such a timescale of revolutions and theirsizes in the history of science would be difficult and controversial,but Nicholas Rescher (1978, 2006) has begun the task in terms ofranking scientific discoveries and studying their distribution overtime. Derek Price (1963) had previously introduced quantitativehistorical considerations into history of science, pointing out, amongmany other things, the exponential increase in the number ofscientists and quantity of their publications since the ScientificRevolution. Such an exponential increase, faster than world populationincrease, obviously cannot continue forever and, in fact, was alreadybeginning to plateau in industrialized nations in the 1960s. Amongphilosophers, Rescher was probably the first to analyze aggregate dataconcerning scientific innovation, arguing that, as researchprogresses, discoveries of a given magnitude become more difficult.Rescher concludes that we must eventually expect a decrease in therate of discovery of a given magnitude and hence, presumably, asimilar decrease in the rate of scientific revolutions. Although hedoes not mention Schumpeter in this work, he expresses a similarview:
Scientific progress in large measure annihilates rather than enlargeswhat has gone before—it builds the new on the foundations of theruins of the old. Scientific theorizing generally moves ahead not byaddition and enlargement but by demolition andreplacement. [1978, p. 82, Rescher’s emphasis]
This broadly Kuhnian position position on the number and magnitude ofrevolutions contrasts sharply with Butterfield’s, who sawrevolutions only as founding revolutions, and also with that of thoseepistemological realists who grant that revolutionary conceptual andpractical changes have occurred but who believe that they will becomesuccessively smaller in the future as science approaches the truetheory. Kuhn’s own later position, in which specialties areinsulated from one another by taxonomic incommensurability, presentsus with a somewhat less integrated conception of science and thus oneless subject to large-scale revolutionary disruption. Since we canregard scientific practices and organization as highly designedtechnological systems, the work of Charles Perrow and others ontechnological risk is relevant here. (See Perrow 1984 for entry intothis approach.)
Margolis (1993) notes the importance of the phenomenon of“contagion,” in which new ideas or practices suddenlyreach a kind of social tipping point and spread rapidly. Contagion is,of course, necessary for a revolt to succeed as a revolution. Today,contagion is a topic being studied carefully by network theorists andpopularized by Malcolm Gladwell’s The Tipping Point(2000). Steven Strogatz, Duncan Watts, and Albert-LászlóBarabási are among the new breed of network theorists who aredeveloping technical accounts of “phase changes” resultingfrom the growth and reorganization of networks, including socialnetworks of science—a topic dear to the early Kuhn’s heartas he struggled with the themes of Structure (see Strogatz,2003, chap. 10; Watts 1999; Newman 2001; Barabási 2002;Buchanan 2002).
Does the emergence to prominence of “chaos theory”(nonlinear dynamics) itself constitute a scientific revolution and, ifso, is it a distinctly Kuhnian revolution? In recent years severalwriters, including both scientists and science writers, have attemptedto link Kuhn’s idea of revolutionary paradigm shifts to theemergence of chaos theory, complexity theory, and network theory(e.g., Gleick 1987, chap. 2, on the chaos theory revolution; Ruelle1991, chap. 11; Jen in Cowan et al. 1999, 622f, on complexity theory;and Buchanan 2002, 47, on network theory). Interestingly, some authorsreapply these ideas to Kuhn’s account itself, theoreticallyconstruing revolutionary paradigm shifts as phase changes or asnonlinear jumps from one strange attractor or one sort of networkstructure to another.
Steven Kellert (1993) considers and rejects the claim that chaostheory represents a Kuhnian revolution. Although it does provide a newset of research problems and standards and, to some degree, transformsour worldview, it does not overturn and replace an entrenched theory.Kellert argues that chaos theory does not even constitute theemergence of a new, mature science rather than an extension ofstandard mechanics, although it may constitute a new style ofreasoning.
Kellert’s position hangs partly on how we construe theories. Ifa theory is just a toolbox of models, something like an integratedcollection of Kuhnian exemplars (Giere 1988, Teller 2008), then theclaim for a revolutionary theory development of some kind becomes moreplausible. For nonlinear dynamics highlights a new set of models andthe strange attractors that characterize their behaviors. In addition,complex systems theorists often stress the holistic, anti-reductive,emergent nature of the systems they study, by contrast with thelinear, Newtonian paradigm. Kuhn wrote that one way in which normalscience articulates its paradigm is by “permitting the solutionof problems to which it had previously only drawn attention.”But had not classical dynamics suppressed rather than drawn attentionto the problems of chaos theory and the various sorts of complexitytheory and network theory that are much studied today? Still, it iseasy to agree with Kellert that this case does not fit Kuhn’saccount neatly. To some readers it suggests that a more pluralisticconception of scientific revolutions than Kuhn’s is needed.
Kellert also questions whether traditional dynamics was really in aspecial state of crisis prior to the recent emphasis on nonlineardynamics, for difficulties in dealing with nonlinear phenomena havebeen apparent almost from the beginning. Since Kuhn himselfemphasized, against Popper, that all theories face anomalies at alltimes, it is unfortunately all too easy, after an apparentlyrevolutionary development, to point back and claim crisis.
Kuhn’s work called attention to what he called “theessential tension” between tradition and innovation (Kuhn 1959,1977a). While he initially claimed that his model applied only tomature natural sciences such as physics, chemistry, and parts ofbiology, he believed that the essential tension point applies, invarying degrees, to all enterprises that place a premium on creativeinnovation. His work thereby raises interesting questions, such aswhich kinds of social structures make revolution necessary (bycontrast with more continuous varieties of transformative change) andwhether those that do experience revolutions tend to be moreprogressive by some standard.
Some analysts agree that casting the net more widely might shedcomparative light on scientific change, and that Kuhn’s model istoo restrictive even when applied only to the mature sciences. We havealready met several alternative conceptions of transformative changein the sciences. Kuhn believed that innovation in the arts was oftentoo divergent fully to express the essential tension. By contrast, thesciences, he claimed, do not seek innovation for its own sake, atleast normal scientists do not.
But what about technological innovation (which is often closelyrelated to mature science) and what about business enterprise moregenerally? There are, of course, important differences between theproducts of basic scientific research and commercial products andservices, but there are enough similarities to make comparisonworthwhile—the more so with today’s emphasis ontranslational science. And in the sciences as well as economic lifethere would seem to be other forms of displacement than the logicaland epistemological forms commonly recognized by philosophers ofscience. Consider the familiar economic phenomenon of obsolescence,including cases that lead to major social reorganization astechnological systems are improved. Think of algorithmic data miningand statistical computation, robotics, and the automation to be foundin any modern biological laboratory. In The Innovator’sDilemma (1997), economist Clayton Christensen denies that majortechnological breakthroughs are either necessary or sufficient fordisruptive innovation. In that and later work he distinguishessustaining technologies that make incremental improvements of acompany’s sales leaders from two kinds of disruptivetechnologies. “New-market disruptions” appeal to apreviously non-existent market, whereas“low-market” or “low-end disruptions” providesimpler and cheaper ways to do things than do the leading products andservices. Such companies can sometimes scale up their more efficientprocesses to displace the major players, as did Japanese steel makersto the big U.S. corporations. There would seem to be parallels in thehistory of science.
Speaking of technological developments, philosophers, including Kuhn,have undervalued a major source of transformative developments,namely, material culture, specifically the development of newinstruments. There is, however, a growing literature in history andsociology of science and technology. A good example is AndyPickering’s discussion of the conception and construction of thelarge cloud chamber at Lawrence Berkeley Laboratory (Pickering 1995).Pickering’s Constructing Quarks (1984), PeterGalison’s How Experiments End (1987) and Image andLogic (1997), and Sharon Traweek’s Beamtimes andLifetimes (1988) describe the cultures that grew up around thebig machines and big theories of high-energy physics in the U.S.,Europe, and Japan. As he himself recognized, Kuhn’s model ofrapid change runs into increasing difficulty with the Big Science ofthe World War II era and beyond. But a similar point extends tosmaller-scale material practices as documented by much recentresearch, as in Baird (2004), discussed above. One line of fruitfulinvestigation has been that of the Social Construction of Technology(SCOT) program of Trevor Pinch and Wiebe Bijker (see Bijker et al.1987 and a good deal of later work). Such work takes place on allscales.
In Structure and later writings, Kuhn locates revolutionarychange both at the logico-semantical and methodological level(incompatibility between successor and predecessor paradigm) and atthe level of form of community life and practice. But does the latteralways require the former? Perhaps expressions such as “theproblem of conceptual change” and “breaking outof the old conceptual framework” have led philosophers toover-intellectualize historical change. As we know from the history ofeconomics and business, one form of life can replace another invarious ways without being based directly upon a logical or semanticincompatibility. The old ways may be not wrong but simply obsolete,inefficient, out of fashion—destroyed by a process that requiresmore resources than simple logical relations to understand it. Therecan be massive displacement by non-logical means. Many have arguedthat Kuhn’s semantic holism, with its logical-relationalunderpinnings, led him to underappreciate how flexible scientists andtechnologists can be at the frontiers of research (Galison 1997).Having distinguished the working scientists’ point of view fromthose of the historian and the philosopher, looking down from above,he proceeded to confuse them. Retrospectively, as many commentatorshave noted, we can view Kuhn on scientific revolutions as atransitional figure, more indebted to logical empiricist conceptionsof logic, language, and meaning than he could have recognized at thetime, while departing sharply from the logical empiricists and Popperin other respects.