Among the topics discussed in Thomas Kuhn's The Structure of Scientific Revolutions, those of education, training, and pedagogy are apt to seem the least compelling. Certainly, the earliest debates about Structure focused on other, more controversial, matters: incommensurability, meaning change, the rationality of theory choice, normal science—the list goes on.Footnote ² Over the past two decades, however, a growing concern among historians and sociologists of science with the nature of scientific apprenticeship has stimulated greater appreciation of the importance of questions of teaching and learning to the philosophical position sketched in Structure.Footnote ³ This paper seeks to develop further our understanding of these issues as they bear on Kuhn's theory of science.

Nothing is more crucial to the description of scientific development presented in Structure than the claim that scientists-in-training acquire the tools and skills of their discipline by learning to work with model puzzle solutions. According to Kuhn, a “community's paradigms” were constituted by “a set of recurrent and quasi-standard illustrations of various theories in their conceptual, observational, and instrumental applications.” These standard illustrations, in turn, were “revealed in [the community's] textbooks, lectures, and laboratory exercises.” So crucial was this view of learning-through-paradigms to Kuhn's historical epistemology that he framed around it his description of the major stages of a scientific revolution. Kuhn was frank that what he called “normal science” could be explained only by the pedagogy of paradigms. This alone could account for the persistence of a research tradition in what Kuhn insisted was the evident absence of consensus on methodological, theoretical, and ontological fundamentals. In the process of research, meanwhile, anomalies could emerge only because paradigms conditioned scientists to expect certain kinds of solution to certain kinds of problem; nothing could be said to have “gone wrong” in the conduct of an experiment or calculation without the violation of expectations set by socialization into a normal-scientific tradition. Likewise, the chronic “technical breakdown” associated with a scientific crisis was explicable only if the habitual manner of modeling research on standard solutions—a habit acquired in the earliest stages of training and carried beyond the PhD—repeatedly failed to account for the phenomena produced in the course of inquiry. In such cases the stage was set for a revolution in which a new paradigm replaced the old.Footnote ⁴

Only the most casual readers of Structure will be surprised by these observations. Nevertheless, we have yet fully to appreciate the commitments and experiences that drew Kuhn to this particular pedagogical conception of scientific knowledge. We can begin to grasp the importance of this historical context for understanding Kuhn's intervention when we note that Structure was conceived and written during a transformative period in the history of educational theory and practice in the United States. When, in the autumn of 1945, Kuhn began jotting down his earliest thoughts on the epistemology of the sciences, American higher education was undergoing profound structural changes caused by the Second World War and its immediate aftermath. During the war, an infusion of funds from the state dramatically scaled up the operations of research in the physical, biological, and engineering sciences. Wartime research on nuclear weapons, radar, digital computing, psycho-acoustics, and much else besides demanded not only high capital expenditures of the sort only the state could afford, but also increased professional manpower in the form of professors, postdocs, graduate students, and rank-and-file technicians.Footnote ⁵ In the decade after 1945, disciplines like physics and engineering became institutional behemoths that consumed resources and spewed out skilled practitioners, research papers, and new technologies at unprecedented rates.Footnote ⁶

If the embryonic national security state stimulated much of the demand for Big Science in the early years of the Cold War, the supply side of the equation was taken care of by the Servicemen's Readjustment Act of 1944. Under Title II of the G.I. Bill, some 2.25 million veterans attended America's colleges and universities in the years following the end of the war. Of that number, nearly 40 percent, or 867,000, became credentialed professionals.Footnote ⁷ The specialization inherent in the veteran-led boom in higher education heightened existing worries among academics about the cultural fragmentation that appeared to be a central feature of the new world of capital- and labor-intensive science. In the immediate postwar years, Harvard president James Bryant Conant was one among many educationalists calling for programs of “general education” in America's schools and colleges that could allow future generations of public and business leaders to make sound judgments in an era of hyperspecialization and increasing reliance on scientific expertise.Footnote ⁸ The Harvard “red book,” otherwise known as General Education in a Free Society (1945), was perhaps the sole instance of wartime Big Humanities, and set the scene for a running debate about general education in an era of unprecedented enrolments at colleges and graduate schools across the nation.Footnote ⁹ Throughout the 1950s, the emergence of a pedagogy of scale in the liberal arts and sciences continued apace, with physics in particular taking the lead in developing the institutional forms associated with educational Fordism.Footnote ¹⁰

A further round of educational reform was prompted by the launch of the Soviet Sputnik satellite in 1957.Footnote ¹¹ The National Defense Education Act, passed by Congress in 1958 in response to perceived deficiencies in American science and language teaching, pumped money into training and research programs in physics, mathematics, engineering, and area studies.Footnote ¹² Whatever its intellectual achievements, this investment strategy led to exponential increases in enrolments and PhD awards in these fields.Footnote ¹³Sputnik also brought to the surface a critique of pedagogy that had been building in American psychology and educational theory for a generation. Popular bromides like Mortimer Smith's And Madly Teach (1949) and Rudolf Flesch's Why Johnny Can't Read—And What You Can Do about It (1955) laid the foundations for a thoroughgoing attack on the tradition of progressive education associated with John Dewey and his followers.Footnote ¹⁴ Many considered the student-led Deweyan curriculum lacking in the rigor and theoretical grounding necessary to produce innovative scientists and engineers. The Woods Hole Conference of 1959 was an especially important staging post in the battle for a new curriculum in science education, not least because it provided a forum for the reception of work in child development and the psychology of learning carried out by the Swiss psychologist Jean Piaget and by a founding figure in postwar cognitive psychology, Jerome Bruner.Footnote ¹⁵

It is without hyperbole, then, that one can describe the fifteen or so years following the end of the war as a time of major upheaval in American education and pedagogical theory. A good deal of Kuhn's involvement with these changes on the educational scene is either well covered in the literature, or else obvious after a quick scan of Structure and related writings. It is widely known, for example, that Kuhn got his start in the history of science as an assistant on James Bryant Conant's natural-science course in Harvard's fabled postwar general education program.Footnote ¹⁶ On the other hand, much of Kuhn's early writing about the stages of scientific development abounds with references to work in developmental and cognitive psychology. Around the time that Structure appeared in print, Kuhn was kibitzing with psychologists, learning theorists, economists, sociologists, and historians who were animated by the great post-Sputnik obsession with the social, economic and psychological conditions of technological innovation.Footnote ¹⁷

Taken together, these observations suggest, if they do not yet explain, important connections between the theory of science limned in Structure and the transformation of American education at multiple levels during the 1940s and 1950s. For historians of science pedagogy, as well as for intellectual historians and historians of science generally, Structure is, or ought to be treated as, more than a useful methodological resource: it is itself to be considered a significant text in the intertwined histories of education, science, and epistemology during the passage from world war to Cold War. Having made this rather bold claim, I must admit that there are some considerable exegetical difficulties standing in the way of our appropriating Structure as a work of philosophy of science suffused with contemporary concerns with pedagogy, education, and training. In particular, there remains in Structure a significant hermeneutic puzzle concerning Kuhn's conception of pedagogy and its place in scientific epistemology. My aim in the rest of this paper is to describe the puzzle and then show how it can be solved by examining the pedagogical traditions in which Kuhn's early theory of science was formed.

* * *

Although he appealed to an exemplar-based theory of learning to explain why most scientists stuck to the normal-scientific tradition in which they had been socialized, Kuhn did not, as David Kaiser has noted, “explore in any historical or anthropological detail how such exemplars emerged, how students in various generations actually learned to solve exemplars and build upon them in their own research, or whether students at different training centers learned about and learned upon exemplars in distinct ways.”Footnote ¹⁸ Lack of detail is the least of it: when we turn to the passages of Structure in which Kuhn asserted the primacy of exemplars in binding scientists into a normal-scientific tradition, we see that Kuhn offered no examples whatsoever of pedagogical regimes in the history of science.Footnote ¹⁹ In Structure, as well as in important essays touching on science pedagogy like “The Essential Tension” (1959) and “The Function of Measurement in Modern Physical Science” (1961), Kuhn presented some distinctive views about the role of demonstration and textbook exercises in the formation of normal-scientific practice. But in none of these cases did his remarks follow from an historical assessment of traditions of scientific education; at most, he assumed that treatises like Newton's Opticks had functioned as the “books from which men learned science” and left the matter there.Footnote ²⁰ What we need to understand is what, given the lack of scholarly machinery driving Kuhn's otherwise determined discussion of pedagogy and professional training, could have motivated his account of learning-through-paradigms.

One explanation is readily available. In Structure, Kuhn's description of exemplar-based learning appeared to hinge on an analogy with the account of the learning of kind-terms given in Wittgenstein's Philosophical Investigations (1953). Kuhn took Wittgenstein's problem to be how terms like “chair,” “leaf,” or “game” came to be applied “unequivocally and without argument” by members of a speech community—even though the members of that community could not be said to agree upon some universal rule of application or criterion of identification governing the connections between a word and its referents.Footnote ²¹ Wittgenstein's point had been that rules or criteria were not themselves sufficiently robust or unambiguous to provide a secure basis for mutual understanding.Footnote ²² For Kuhn, Wittgenstein's answer to this problem was that what looked like explicit agreement on rules or criteria was in fact an overlap of the model applications or samples through which each individual member of the speech community learned the term in question. A sufficient overlap between these sets of model applications was sufficient to ensure mutual understanding between speakers and, in most cases, agreement on how to apply terms in new cases. The standard illustrations or paradigms acquired in the course of scientific training, Kuhn suggested, functioned in much the same way: scientists did not learn theories, laws, or concepts as fully explicit and autonomous systems of rules, despite what contemporary proponents of logical positivism or operationism may have suggested; instead, “these intellectual tools [were] from the start encountered in a historically and pedagogically prior unit that display[ed] them with and through their applications.”Footnote ²³

When we reflect upon these crucial passages in Structure, it is tempting to conclude that the Wittgensteinian philosophy is driving the appeal to pedagogy. Footnote ²⁴ A little more digging into the history of the composition of Structure, however, shows us that Wittgenstein's cameo appearance is a red herring—at least insofar as it makes Kuhn's invocation of a paradigm-based pedagogy into a consequence or working-through of his indebtedness to Wittgenstein's philosophy. I aim to show that the matter is better put the other way around: that Kuhn's whirlwind romance with the Investigations represented the philosophical aftereffects of a long and complex effort to strike a balance between competing experiences of pedagogy, professional training, and epistemology.

For Kuhn, questions of education and professional training were far from academic. His path to Structure took him through more or less every major pedagogical program the United States had to offer in the tumultuous middle decades of the twentieth century. Before he took up a place at Harvard College in 1940, Kuhn spent a substantial part of his childhood and adolescence in progressive schools in the Northeast. At the Lincoln School of Teachers College in New York, a bastion of Deweyan pedagogy, and later at the Hessian Hills School in Croton-on-Hudson, Kuhn absorbed the progressive ethos that would later come under attack in the 1950s. This was a pedagogy without textbooks, rote learning, or any of the other facets of “dogmatic initiation” that Kuhn would later identify with professional training in the sciences.Footnote ²⁵ In contrast, during his years at Harvard, Kuhn encountered intensive mass training at its most intensive: although he had chosen to major (or, in local parlance, to “concentrate”) in physics, Kuhn was rapidly inducted into a compressed wartime program in electrical engineering, which provided Kuhn with a grounding in electronics rather than physics. Graduating in three rather than four years, Kuhn was immediately set to work on radar countermeasures in Harvard's Radio Research Laboratory, where he spent most of his time, as he later recalled, “cooking standard formulas . . . on doing radar profiles as a function of distance.”Footnote ²⁶ While the dominant theme of his college education, intensive training was mixed in Kuhn's undergraduate career with a lingering interest in the literary and the philosophical: he served as president of Harvard's literary Signet Society and editor of the Harvard Crimson, and availed himself of some courses in the history of philosophy. Faced with choosing a career after the Allied victory, Kuhn found a way of hedging professional training and avocational interests. At the same time as beginning graduate work at Harvard on solid-state physics, under the direction of John Van Vleck, Kuhn received special permission to take courses in philosophy that could count as credits toward the doctoral degree. In addition, he provided the alumni voice in response to the Harvard report on general education, and was eventually asked by Conant, in the spring of 1947, to prepare a class for his general education course. This brief led Kuhn to yet another unconventional turn in his professional training. His work for Conant allowed him to jump ship from physics to the history of science, and from the physics department to the resolutely cross-disciplinary Society of Fellows.

The Society of Fellows had been established by senior Harvard faculty in the early 1930s specifically to serve as a bastion against vocationalism in the arts and sciences; in the society's first incarnation, Junior Fellows were not, at least in principle, permitted to take advanced degrees.Footnote ²⁷ Kuhn already had a PhD when he took up his fellowship in 1948, but he certainly used the freedom granted him by three years of research to undertake an eclectic training program that encompassed reading across five or six different fields, from psychology to sociology, history, philosophy, and logic. It is more accurate to say of this postdoctoral fellowship that Kuhn trained not as a historian of science, and certainly not as a professional historian, but instead as a professor of general education in science with interests across a range of fields. By the terminal year of his fellowship, Kuhn was working full-time on the history of the experimental-sciences course, Natural Sciences 4, in the Harvard general education program. During the years that he held an assistant professorship in general education, Kuhn worked on historical and sociological problems concerning the study of scientific theory. Even when denied tenure at Harvard, Kuhn's peculiar, hybrid existence as a nonspecialist continued when he was given a unique, and ultimately unhappy, joint appointment in history and philosophy at Berkeley, where he stayed until a couple of years after the publication of Structure.

It is against the background of this eclectic educational career that we can begin to understand the role that notions of pedagogy and professional training would have in Kuhn's earliest work as a historian and philosopher of science. Kuhn seems to have registered strongly the contrast between the open-ended, independent-minded character of his adolescent studies in New York state progressive schools and the compressed period of intensive training provided by his initiation into electrical engineering as a physics major at wartime Harvard. The liberal and aesthetic leanings of the former made the rigidity of his truncated training in physics and electronics clear enough to instill in Kuhn a yen for, and conviction in, the wider philosophical and even metaphysical significance of science. This formative set of experiences, in turn, led Kuhn to a postwar education in physics, history, and philosophy that at once was deeply informed by the practices and norms of professionalization and yet, for one reason or another, escaped its exigencies.

Kuhn's dissatisfaction with the limits of his undergraduate crash course in physics emerged early. In the summer of 1943, having just turned twenty-one and on the cusp of taking up his position at the Radio Research Laboratory, Kuhn admitted in a letter to his aunt that he felt “if anything younger than before: that is further from any satisfactory creative intellectual or social functionings, not freer from responsibility.” Referring to an earlier letter, which would appear to have discussed his hopes for a future career, Kuhn confessed that the “dreaming goes on . . . I lean, in short, increasingly toward the academic, and to philosophy of & through science rather than to pure physics. But I am to [sic] far from knowing where it will end to waste your time here.” It will perhaps be evident that Kuhn's inclination toward “philosophy of and through science” reflected his desire to access the philosophical, speculative, or psychological features of science that would have been lacking in his Harvard education up to that point. It was therefore significant that he followed his expression of yearning for philosophy by voicing his disappointment with his then-current reading: Freud's Introductory Lectures on Psychoanalysis. “I like the frankness,” Kuhn wrote of the book, “I believe in the therapy . . . but I don't trust the generalizations. Above all, I fail to find even a fragmentary basis for a psychological epistemology, and my selfish hope had been to discover some such.”Footnote ²⁸

Remarks like these can be interpreted in terms of Kuhn's recurrent experience of intellectual crisis during his college and graduate years.Footnote ²⁹ Here I want to take a different tack and emphasize Kuhn's already firm commitment to identifying the psychological and philosophical underpinnings of science—and this just as he had graduated and was set to embark upon puzzle-solving work in radar countermeasures. Something in Kuhn resisted the professional mindset. It is reasonable to speculate that a combination of an unsystematic initiation in physics, due to the exigencies of the wartime Harvard curriculum, in combination with the trust in his own intellectual instincts engendered by progressive schooling, produced in Kuhn this unformed but strong impulse to philosophize. Whatever the exact source of Kuhn's unschooled philosophical interests, they intensified during his graduate and postdoctoral studies in physics and the history of science. In the immediate postwar years, Kuhn's inclination toward “philosophy of and through science” led to a dalliance with professional philosophy. At the outset of his graduate training in physics, in the autumn of 1945, Kuhn was given permission to take undergraduate philosophy courses. This experiment was swiftly curtailed. As Kuhn later recollected, he felt he had no clear point of access to the discipline:

Once again the spectre of specialist training had raised its head, and once again Kuhn recoiled. But he was far from ready to give up on the idea of a philosophical or epistemological or metaphysical something that was needed to account for scientific knowledge. In one of the philosophical papers he produced that autumn, entitled “The Metaphysical Possibilities of Physics,” Kuhn spoke directly to the sense of vertigo he experienced during the war when he lamented that “the philosophic base of the physical sciences has been lost by the physicist so often that it has become necessary to convince people that physics is more than an adjunct of engineering and a complex means for answering trick problems.” “Too seldom is it realized,” Kuhn declared, “that physics is also a philosophical science searching for absolute truths concerning the nature and structure of the universe.”Footnote ³¹

Enter James Bryant Conant and the Harvard general education program. As someone tutored in both extremes of mid-century American pedagogy—the progressive tradition and wartime intensive training—Kuhn was already interested, by the end of the war, in “the need for the redesign of science instruction.”Footnote ³² When asked by the Harvard Alumni Bulletin in 1945 to offer a recent undergraduate's perspective on the general education report, Kuhn produced an opinion piece that focused primarily on the tension between specialist training and liberal education, as this was registered in the student experience of what were called “distribution” courses: required courses taken in addition to the major. Both survey and advanced courses for distribution had been designed for those intending to major in the field, with the consequence that, for non-concentrators in these subjects, there was small chance of gaining either a specialist or, more importantly, a liberal education. Hence both the introductory course, composed of a compendium of basic facts and techniques, and the upper-level course, which was so narrow as to be of no general benefit to the nonspecialist, failed to serve as a “vehicle for general education.” Kuhn, signaling his agreement with the central tenets of the report, argued that “the increase in the bulk of knowledge during the last century necessitates a basic alteration of educational methods, for liberal education is no longer a probable end-product of conventional teaching methods.”Footnote ³³ These pedagogical problems arising from postwar debates about general education and the fissiparous college curriculum spoke directly to Kuhn's ongoing attempt to educe philosophical or liberal meanings from the resolutely professionalized enterprise of Big Science. By the time Conant asked Kuhn to prepare a case study on seventeenth-century mechanics for Natural Sciences 4, the disillusioned young physicist and philosopher manqué was ready to find in his pedagogical brief a philosophical vocation. The Harvard general education program offered Kuhn a space in which to pursue his version of philosophy after professional physics and orthodox philosophy had led him into dead ends. Crucially, however, this also meant that he began to frame his epistemological concerns in terms of the pedagogical distinctions that his new role as a teacher of general education in science encouraged him to draw.

As soon as he received Conant's invitation, Kuhn began to sketch out thoughts about general education in science that covered his long-standing epistemological interests in the non-mechanical, ideological aspects of scientific inquiry. In a May 1947 memorandum on the “Objectives of a General Education Course in the Physical Sciences,” Kuhn was already importing his philosophical commitments into the prerequisites of a nonspecialist teaching program in physics. Students taking the course, Kuhn stated, would need to appreciate the central role played in experimental science by selective “abstraction” and also by the “acceptance of a conceptual structure” in the form of a systematic, predictive theory. In addition, Kuhn was beginning to insist upon the “limited role of pure logic” in understanding science. Pace the advocates of formal approaches to the assessment of scientific theories, induction was for Kuhn “a name for ‘the experimental method’ rather than for a logically, or even operationally, defined route from the particular to the general.” “Scientific generalization” was therefore to be considered “a creative imaginative process.”Footnote ³⁴ Kuhn was certainly prepared for his famous epiphany upon grasping the meaning of “motion” in Aristotle's Physics, which occurred a couple of months later. In cracking Aristotle's code, Kuhn found a concrete illustration of the extra-logical dimensions of scientific theory posited in his earlier statements about scientific epistemology.Footnote ³⁵ Through his interpretation of Aristotle, Kuhn had hit upon a way of extracting philosophy from history, and he had done so in considering the teaching of science to nonspecialists.

* * *

Throughout the following fifteen years, as Kuhn followed a meandering path toward the publication of Structure, this fusion of pedagogical, professional, and philosophical concerns shaped the manner in which the neophyte historian of science sought to understand scientific development. In particular, Kuhn's instinctual sense of the limits of the professional self-understanding of science, evident as early as 1943, was reinforced in both his pedagogical and his philosophical endeavors. This sense showed up pedagogically in Kuhn's distinction, common in the Harvard discussion about general education, between specialist training and liberal learning in the arts and sciences. A fundamental challenge for Conant and the authors of General Education in a Free Society had been to rethink the foundations of liberal education in the face of an ever more specialized curriculum.Footnote ³⁶ The central tenet of this way of thinking was that, whatever else a general education in the sciences might be, it could not resemble the usual combination of survey and advanced courses in a discipline, for this had become the first step on the path to a professional career. As a teacher of science to laymen, one's objective was to convey what Conant called the “tactics and strategies of science” without recourse to the pedagogical infrastructure of the sciences themselves: textbooks, lab work, and the rest. In unpublished talks given on science and general education delivered during the late 1940s and early 1950s, Kuhn repeatedly underscored the point that the general education courses in science at Harvard were for the “non-scientist” and involved “no labs.”Footnote ³⁷ “No one of these courses is a survey course,” he told a group of educators at a 1951 conference; “all sacrifice portions of ‘content’ or subject matter that would be essential to the specialist.”Footnote ³⁸

The substitute for the specialist's pedagogical toolkit, for Kuhn and his mentor Conant, was the case history. Many of the cases employed in Conant's, and later Kuhn's, general education course, Natural Sciences 4, were later written up and published in two volumes as the Harvard Case Histories in Experimental Science (1957).Footnote ³⁹ Each of the case histories was designed to reveal what Conant described as “the evolution of a new conceptual scheme” in a particular research tradition.Footnote ⁴⁰ Students would not acquire new technical skills or a comprehensive body of knowledge; rather, the case histories would develop their sense of the practical wisdom involved in the conduct of exemplary scientific investigations. This connection between case studies and experienced judgment in practical affairs had been the motif of the development of the so-called “case system” of pedagogy at Harvard during the decades after Christopher Columbus Langdell first introduced it in the Law School in the 1870s. At the time Conant took up the case method in general education, it was employed both in the Harvard Business School and by Conant's uncle by marriage, the biochemist and Pareto disciple L. J. Henderson, who taught courses in history of science and applied sociology.Footnote ⁴¹ Kuhn adopted Conant's case system wholesale, and noted that the case method was “not [an] economical approach to science teaching,” which was why it was “quite properly omitted from standard science courses.”Footnote ⁴² This way of framing the difference between specialist and generalist education reaffirmed at a pedagogical and philosophical level Kuhn's existing distinction between the mechanisms of professional training, on the one hand, and the wider values and commitments that such training obscured from the trainee scientist, on the other. Textbooks and practical instruction animated science as a vocational enterprise, but case histories were needed to call attention to the creative revisions of conceptual schemes that defined scientific revolutions, and which were the bread-and-butter of a general education in science.

These professional and pedagogical orientations had important consequences for Kuhn's emerging theory of science in the early 1950s. With the critical distance he had gained on the professional scientific enterprise both from his own education and from the demands of creating an independent domain for general education in science, Kuhn gradually came to see that the dogmatic procedures of specialist training in the sciences could explain how what he called the “ideological” features of science were able to keep particular research traditions within fairly narrow and stable parameters. Nevertheless, the moments of dramatic change in conceptual schemes assessed in the Harvard case histories were, in this period of Kuhn's intellectual development, simply not be accounted for by the effects of scientific pedagogy; the explanation of revolutions had to work with the generalist's toolkit: intellectual history, sociology, psychology, and so on. When Kuhn suggested in his earliest years as a lecturer for Natural Sciences 4 that “creative science” was “totally different” from the kind of scientific practices resulting from a specialist education, he was marking the limits of pedagogy as an engine of scientific development.Footnote ⁴³ What I want to show, in conclusion, is that this view led Kuhn into what he came to consider a cul-de-sac in his philosophy of science, and that he overcame the impasse by identifying the dogmatic procedures of professional training as the very source and condition of possibility of revolution. Although he gave Wittgenstein some of the credit for this transformation in the published version of Structure, he resolved his problem by writing the case-based teaching method he acquired in the general education program into his theory of specialist training.

In each of these steps on the road to Structure, textbooks and the necessarily blinkered nature of professional training in science played an important, but subtly shifting, role. Very soon after Kuhn began his stint in the Society of Fellows, he started to translate his convictions regarding the pedagogical function of textbooks in specialist education into a philosophical critique of attempts to “formalize” or “logicize” scientific knowledge. Readers of Structure will know that when Kuhn undertook to criticize positivist philosophies of science like operationism and logical empiricism, he did so not through systematic assessments of these positions, but rather by consistently associating them with the textbook vision of science as a static, logically interrelated set of concepts and laws.Footnote ⁴⁴ The difficulty Kuhn had with these philosophical positions, to the rather limited extent he discussed them in Structure, was that they were epistemologies of the textbook; by contrast, everything Kuhn had learned in thinking about a general education in science pointed to the significance of non-formalizable elements in shaping the actual process of research and discovery. More than a decade before Structure was published, Kuhn was in the habit of associating textbooks with formalization programs in the philosophy of science. In one of earliest outlines for Structure, likely written in the late 1940s, Kuhn attacked the treatment of scientific theories as logical calculi or mechanical operations, but drew his reflections to a conclusion by observing that “text-books represent a relatively logicized form of language,” and were thus the root of the philosophical error represented by the formalism of contemporary philosophy of science.Footnote ⁴⁵ A 1951 letter took a similar line: “the major methodological tradition since Bacon and Descartes,” Kuhn wrote in outlining his chosen topic for a series of public lectures, “has drawn its conclusions from a study of the logical structure of the end-products of scientific research and has proceeded as if it assumed the study of science to be synonymous with a study of textbooks reporting on completed scientific conceptualizations.”Footnote ⁴⁶ When Kuhn gave those lectures at the Lowell Institute in Boston, also in 1951, he once again lumped together formal treatments in philosophy under the heading of “textbook science.”Footnote ⁴⁷

In the Lowell lectures, the antithesis of textbook science was “creative science,” understood by Kuhn as the actual process of scientific research, and involving especially the preconceptions, conceptual schemes, and ideological commitments that animated it. This was of course exactly the sort of thing that the case histories given in general education science courses were designed to uncover. In practice, however, Kuhn demonized the textbook as a source of philosophical data, only to bring the dogmatic pedagogical function of textbooks in scientific training back as the explanation for the hold of conceptual schemes on the conduct of science. That would become the double life of textbooks and professional training in Structure: at once the source of illusion concerning the nature of scientific knowledge and the indispensable means by which the preconceptions and ideologies that made normal-scientific inquiry possible were inculcated within a scientific community. What is usually passed over in assessments of this feature of Structure is that, before the final draft of Structure went to press, Kuhn piloted and then abandoned an account of what kind of shared understanding was brought about by educational initiation into the conceptual schemes of a field.

Much of this turned on Kuhn's evolving view of what the non-formalizable, ideological elements of scientific theories were composed of. When he was in the Society of Fellows between 1948 and 1951, Kuhn read extraordinarily widely in the search for a characterization of the conceptual frames that conditioned the scientific pursuit of knowledge.Footnote ⁴⁸ However, by the early 1950s his understanding of the efficacy of professional education became increasingly sociological. Although they receded into the background as mental sets or gestalts, the ideologies conveyed in textbook training were to be considered stocks of beliefs or rules on which all members of a scientific community agreed. Some of this was already evident in Kuhn's Lowell lectures, as when he claimed that the “predispositions” that conditioned inquiry had “social sources.”Footnote ⁴⁹ Responding in 1952 to an invitation from the physicist and philosopher Philipp Frank to participate in a study group on the sociology of science, Kuhn suggested that “in twentieth-century western science socially conditioned, implicit, professional ‘faiths’ have assumed many of the roles in the guidance of research and the acceptability of scientific theories which religions and metaphysical systems played in the physical science of the seventeenth century.” As Kuhn saw it, “this professional consensus has important bearing upon the problems which a scientist considers worth attacking, the experiments which he employs to resolve his problems,” and also on the various “logical and experimental criteria” demanded of a “‘valid’ argument.”Footnote ⁵⁰ Similarly, when Kuhn sent his editor an informal prospectus for Structure in 1953, he emphasized, in contrast to “the formal content of a theory,” the “functions of a theory as a profession[al] ideology for the practicing scientist.” Theory qua ideology, Kuhn claimed, “dictates preferred techniques of interpretation, and . . . sets standards of precision in experiment and of rigor in reasoning”; it was, in sum, “a source simultaneously of essential direction and of disastrous inhibition to the creative imagination.”Footnote ⁵¹

It was this radical disjuncture between ideology-driven ordinary science and the creative imagination that would collapse as Kuhn moved, during the late 1950s, to identify the “essential tension” between dogma and innovation in scientific research. The cause of this collapse was precisely the notion that ideology rested on consensus on rules or shared beliefs and value-commitments. Even as Kuhn began to speak, in his drafts for Structure, of the paradigms involved in textbook-based pedagogy, he continued to believe that what those paradigms delivered to the initiate were common values and beliefs. As late as the summer of 1960, when he completed the “penultimate draft” of Structure, Kuhn was still labeling his discussion of the sociology of professionalization “Normal Science as Rule-Determined.” Revealingly, Chapter 5 of the published edition of Structure, “The Priority of Paradigms,” was not in the penultimate draft—and nor was any of the material, including the discussion of Wittgenstein, that it contained. At this juncture, however, Kuhn was eager to confess that his earlier view—that “periods of normal science were periods of consensus, during which the entire scientific community agreed about the rules of the game,” and that revolutions were therefore “episodes through which the rules of the game were changed”—was “very probably wrong.” His highlighting of the role that standard textbook examples and laboratory demonstrations played in forging agreement had revealed that “the number of rules that can be educed by such study never seems sufficient to define the puzzles that scientists normally undertake or to restrict scientific attention to their pursuit.”Footnote ⁵²

Although Kuhn had nothing to say at this stage in the drafting process about Wittgenstein, he did something more revealing: having held textbook and creative science apart for the better part of a decade, along with the mechanisms of specialist and generalist training in science, he now implied that the standard samples and applications given in scientific training regimes played a more subtle role in forming the scientist's mental set than merely hammering in consensus on rules and standards. Instead, the model puzzle solutions found in textbooks and practical demonstration possessed some of the pedagogic properties of the exemplary cases hitherto reserved for studying moments of creative ferment in science in the general education program: these exemplars honed the judgments of students, their facilities in practical reasoning in a given area, such that their judgments on salient problems and appropriate solutions would tend to converge in the absence of explicit agreement on theoretical or methodological fundamentals. It was because they learned in this case-based manner, Kuhn argued, “that scientists can so regularly agree in their evaluations of particular problems and particular solutions without manifesting any similar agreement about the full set of rules that appear to underlie their judgment.”Footnote ⁵³ This was exactly the sort of practical wisdom about the nature of the scientific research process that Conant's case histories were supposed to inculcate in non-scientists. Moreover, the training in this sort of non-mechanical, practical judgment had defined the case method in teaching at Harvard for almost a century. In a much overlooked passage of the published version of Structure, Kuhn invoked this tradition of the case method as the pedagogy of practical wisdom when he proposed that a paradigm in science “like an accepted judicial decision in the common law” was “an object for further articulation and specification under new or more stringent conditions.”Footnote ⁵⁴

The way was now clear for Kuhn, in the final draft of Structure, to insist upon the priority of paradigms both in fixing the course of normal science, and in explaining the emergence of anomalies, crises, and revolutions as internally related to the predispositions instilled by specialist education. Kuhn's lifelong alternation between two traditions of pedagogy had led to an account of scientific training, and of scientific knowledge, that combined them both. No doubt Kuhn's encounter with Wittgenstein, that other great theorist of practical reason in human affairs, gave him a way of framing these matters philosophically.Footnote ⁵⁵ But when Kuhn spoke, in Structure and elsewhere, about paradigms and professional training, without any direct historical study of training regimes in science, he was riffing on pedagogical themes that had marked his personal and professional life during the long march to Structure.