Annu. Rev. Sociol. 1995. 21:289-321 N Copyright 0 1995 by Annual Reviews Inc. A rights reserved HERE AND EVERYWHERE: Sociology of Scientific Knowledge Steven Shapin Department of Sociology and Science Studies Program, University of California, San Diego, La Jolla, California 92093-0102 KEY WORDS: philosophy of science, relativism, realism, localism, natural attitude The sociology of scientific knowledge (SSK) is one of the profession's most marginal specialties, yet its objects of inquiry, its modes of inquiry, and certain of its findings have very substantial bearing upon the nature and scope of tihe sociological enterprise in general. While traditional sociology of knowledge asked how, and to what extent, "social factors" might influence the products of the mind, SSK sought to show that knowledge was constitutively social, and in so doing, it raised fundamental questions about taken-for-granted divisions between "social versus cognitive, or natural, factors." This piece traces the historical development of the sociology of scientific knowledge and its relations with sociology and cultural inquiry as a whole. It identifies dominant "localist" sensibilities in SSK and the consequent problem it now confronts of how scientific knowledge travels. Finally, it describes several strands of criticism of SSK that have emerged from among its own practitioners, noting the ways in which some criticisms can be seen as a revival of old aspirations toward privileged meta-languages. There is no shortage of reviews and assessments of the sociology of scientific knowledge (SSK). Most have been written by critics or by participants meaning to put their special stamp on a contentious and splintered field.' I too am a participant: My views about what the field is, and ought to be, are strongly held; they have been canvassed elsewhere, and they will be unavoidably evident in this survey. Yet my purpose here is less to score points than to offer a critical survey 'special note should be taken of Lynch (1993:Ch 2-4), which, while it argues a vigorous case for the virtues of ethnomethodology, is a detailed critical survey of recent trends in the social studies of science. 290 SHAPIN ofhow SSKdeveloped, and continues to develop, in relation to sociology, and to make the leading concerns of the field rather more comprehensible to sociologists in general than they have been. For this specialty, such a purpose is not banal, as neither the place of SSK in the sociological culture nor its implications for the future of sociology+specially social theory-have been adequately canvassed b e f ~ r eThe "here and everywhere" of my title refers at once to the .~ problematic place of SSK within academic sociology and to a central problem it has generated and now confronts-how to interpret the relationship between the local settings in which scientific knowledge is produced and the unique efficiency with which such knowledge seems to travel. Sociology of ScientiJic Knowledge and the Academic Culture SSK must count as one of sociology's notable recent successes. Emerging not more than 25 years ago, in the 1970s and early 1980s it was an almost exclusively British practice (Collins 1983a:26&71). Now there are influential practitioners throughout North America, as well as in France, Germany, the Netherlands, Scandinavia, Israel, and Australia; and key Anglophone works have been translated into French, Italian, Japanese, Polish, Russian, and Spanish. Programs in "Science Studies," "Science and Technology Studies," or "Science, Technology and Societyv-several elaborately funded by the US National Science Foundation-employing sociologists of scientific knowledge have sprung up at leading American universities; relevant professional societies flourish. Journals and academic publishers, once unaware of or uninterested in the field, now actively seek out contributions, creating a situation in which demand outstrips quality supply. Seminal monographs have been reprinted and advertised as "classics" (Bloor 1991, Collins 1992); anthologies, primers, synthetic surveys, and candidate textbooks have appeared and been superseded by new texts bidding to redefine a fast-changing field (Barnes 1972, 1985, Mulkay 1979, Barnes & Shapin 1979, Barnes & Edge 1982, Law & Lodge 1984, Yearley 1984, Woolgar 1988a, Cozzens & Gieryn 1990, Callon & Latour 1991, Jasanoff et a1 1994). Projects have been launched to intercalate the findings of SSK into programs of science communication and liberal education (Collins & Pinch 1993, Chambers & Turnbull 1989)3 and into the analysis and formulation of science and *ASI shall note below, many sociologists of scientific knowledge were not professionally trained in sociology, and neither was I. (My training and much of my work belong more to history than to sociology.) Such amateurism often betrays itself in naivetC, far less often in insight into fundamentals. Not coming to sociology through a normal career-route, I find myself "unprofessionally" interested in what it is to have a sociological understanding of science. 'A measure of the provisional success of these suggestions is a recent series on scientific experiment, published in The Economist (Morton & Cam 1993). which draws heavily upon SSK research. SOCIOLOGY OF SCIENTIFIC KNOWLEDGE 291 technology policy (Jasanoff 1990,1992, Wynne 1992,Collins 1985,Fuller 1993, Cambrosio et a1 1990, Travis & Collins 1991, Epstein 1993), while the potential of SSK (broadly construed) to recast the traditional categories of social and cultural theory as a whole has been asserted (Latour 1993, Law 1994).4 Most importantly, the general academic culture has shown great interest in what has been done in this field. Unlike many other sociological specialties, SSK has strongly engaged the attention of historians and philosophers (e.g. Shapin 1982, Shapin & Schaffer 1985, Rudwick 1985, Golinski 1990, Dear 1995, Fuller 1988, 1992, Rouse 1987, Toulmin 1990). and the boundary lines between what counts as historical or philosophical and what as sociological practice in the area have been blurred to the point of in~isibility.~ Meanwhile anthropologists, literary and feminist theoreticians, and a loosely defined but trendy "cultural studies" community have been attracted in significant numbers to the study of science largely through work in SSK. The social study of science is one of the modem academy's most unremittingly interdisciplinary projects. Twenty-five years ago it was a truth almost universally acknowledged that there might be a legitimate sociological understanding of scientific error, of "the blind alleys entered by science," of the state of scientific institutionalization, and, perhaps, of the overall dynamics of scientific foci, but that there could be no such thing as a sociology of authentically scientific knowledge (Ben-David 1971:11-13). Now, while assent to the validity of SSK is scarcely universal, a number of central claims have quietly passed into common academic currency, and the recent paths of the history and philosophy of science, technology, and medicine have been fundamentally shaped by practitioners' appreciation of opportunities opened up or problems posed by SSK research. At the same time, the field shows many signs of being in serious trouble: Some problems are of very long standing, while others must be seen as the bitter fruits of success itself. The very achievement of SSK in establishing the possibility, legitimacy, and interest of a thoroughly sociological (and social historical) understanding of scientific knowledge has attracted so great a range of scholars from other disciplines that neither the boundaries of the field nor 4 ~ e c e nsocial theorists continue to comment centrally on modern science and technology while t engaging only obliquely or not at all with the SSK literature (e.g. Bauman 1993:199-209, Giddens 1993:9-15, Bourdieu 1990,1991). 'when the speciality was last reviewed in this journal, HM Collins (1983a:272) acutely noted that the relationship between SSK and relevant history was seamless and that a "proper description" of the field "would treat the history of science as integral." Yet, largely restricting his treatment to ethnographic studies of contemporary science, Collins did not there attempt to offer such a "proper" account. Two years earlier (Collins 1981a). he had voiced doubts that historical work was capable in principle of attaining the ethnographer's understanding of science, and four years later (1987) he proclaimed that historical studies represented some of the best SSK. Shifting judgments are possibly best read as reliable reflections of shifting realities. 292 SHAPIN its intellectual goals and foci are any longer at all clear. What appears to some practitioners as an admirable "diversity of voices" seems to others lamentable incoherence and lack of seriousness of purpose. The "social study of science," as opposed to SSK "proper," has developed into one of the modem academy's most centrifugal, most argumentative (at times uncivil), as well as most vital terrains (e.g. Pickering 1992). Just because what is at stake is nothing less than the proper interpretation of our culture's most highly valued form of knowledge-its truth-the struggle for interpretative rights has become fraught and bitter. Names are called and mud is slung. The weight of the world's injustices is dumped firmly on the shoulders of those maintaining "incorrect" methodological views. This is not a practice for the cardiovascularly challenged. Fundamental issues of methodological propriety are fervently debated. Choices between Durkheimian objectivism and Weberian subjectivism, explanatory and interpretativegoals, stress on structureand agency,micro and macro foci, theoretical and empirical methods-all are often fought out in relative disengagement from the career of parallel debates in mainstream sociology, with results ranging from rediscovered wheels to important respecifications of the terms of debate (Callon & Latour 1981, Collins 198la, 1983b,Knorr-Cetina 1981a,b,Law 1974,1984,Tumer 1981). Metaphysical and ontological schemes are proffered, and it is asserted that sociology of science requires the adoption of the correct scheme, while skeptics wonder why interpretative projects should be supposed to require a metaphysics (Latour 1993, Shapin 1992:354-60). Leading sociologists of science discover that the practice has contained social-theoretical entities, such as "interests," and announce their gleeful despair that "definitive" descriptionsor explanations of science canever be attained,while other practitionersexpressbemusementthat anyone could ever think to construct accounts free of theorizing or pretending to definitiveness (Woolgar 1981, Barnes 1981). Relativism is attacked (far less often than it is actually commended) as an insidious threat to the fabric of social order, while advocates argue that methodological (not moral or ontological) relativism is simply necessary for the naturalistic interpretation of variation in belief.6 Practitioners agonize over the proper posture of the analyst, as between disinterested and committed. The original claim that SSK was just the extension of science to the study of itself (Bloor [I9761 1991) has been countered by the increasingly insistent-though perhaps not yet dominant-voices of writers meaning to "expose" science (as 6 ~ o an analyst to say that the credibility of two different beliefs about the world should be r interpreted using the same ntethods is, thus, not necessarily the same thing as saying that they are equally "true" or that the world(s) to which they refer is(are) multiple. Almost all SSK relativists set aside ontological questions and treat tnlth-judgments as topic rather than as resource. So far as morality is concerned, the dominant tendency here is not to celebrate moral anarchy but to interpret how locally varying moral standards acquire their obligatory character. SOCIOLOGY OF SCIENTIFIC KNOWLEDGE 293 "hegemonic," as "masculinist," as "dehumanizing," as "mystifying") and by those who reckon that a proper task for scholars is to open up alternative visions of what science might be and how its social relations ought to be constituted (Martin 1993, Restivo 1989, Lynch & Fuhrman 1991, Scott et a1 1990; cf Collins 1991, Lynch 1992b). Quite recently, small numbers of eminent natural scientists have become aware of SSK, and, cavalierly neglecting crucial differences in tone and intent among practitioners, have sought to expose them all as motivated by hostility to science, purportedly animated by hidden political agendas (Gross & Levitt 1994, Wolpert 1992). Alleged crises in public confidence in, and support for, science have been traced-incredible as it may seem-to the sinister influence of SSK and fellow-traveling philosophy of science (e.g. Theocharis & Psimipoulis 1987). The political vulnerability of one of the few sociological specialties that, so to speak, "studies up," that aims to interpret a culture far more powerful and prestigious than itself, and that offers accounts at variance with that culture's official myths, is only now being made manifest. As the Chinese proverb has it, he who rides on the back of the tiger may wind up inside. The number of sociologists working in the area continues to be very small. The rise of SSK to relative popularity coincided with the Thatcher government's systematic reduction in British university funding, from which several of the original homes of this sort of work suffered significantly, eroding or eliminating their ability to train the next generation. A surge of interest in this area among American institutions from the mid-1980s was also checked by recession and a consequent retrenchment in graduate student support and opportunity. Hard times discourage intellectual adventurousness, on the part of both students and recruitment committees. Time and improving economies may heal these wounds, but endemic structural difficulties beset SSK. First, the sociological study of science makes demands upon initiates which all but a handful find difficult to fulfill. Despite the continuing scientistic bent of North American sociology, few students come equipped with relevant competences in the natural sciences. The genuine incapacity of many to get to grips with the scientific technicalities involved is added to the fear of others that such competences will be extremely hard to acquire. Despite much liberal educational rhetoric and distribution requirements, the gap between the "Two Cultures" described by CP Snow in 1959 has not noticeably been bridged. There is a widespread, and partly justified, sense that SSK is "hard," and students searching for a secure career-track are encouraged to look el~ewhere.~ 'Yet it has to be noted that the study of any culture possessing esoteric knowledge--e.g. that of machinists, soldiers, nurses, or Azande magicians-demands similar dedication and similar commitment to technical mastery. It is arguably not the "difficulty" of science but its "prestige" and "sacredness" that beget this heightened anxiety. 294 SHAPIN Nor does the poor availability of undergraduate courses in the area do much either to inform students about what the specialty is like or to give them even a smattering of familiarity to set against structurally induced "technophobia." Moreover, scientistic North American sociological traditions and, to a lesser extent, traditions in Britain and Europe continue actively to disseminate a picture of scientific "method and scientific knowledge radically at variance with those offered by SSK. There is an argument that the last great bastion of faith in simplistic images of science and its "method is not to be found in the natural but in the social sciences. Students thus trained often find the body of recent SSK-when they encounter it-not just unfamiliar but shocking. Few sociology texts prepare them for the claim that scientific truth is amenable to a thoroughgoing sociological scrutiny, while some of the most senior and eminent authorities remain among the unconvinced or unapproving. Joseph Ben-David (1981:41-47, 54-55), judging work in the area to be largely "programmatic," pronounced SSK to be "sociologically irrelevant" and a "failure." Stephen Cole (1992: 81), while making irenic gestures toward SSK, nevertheless gave his opinion that it had "failed to generate a single example or case study" that shows that social processes "actually influence the specific cognitive content of science." And TS Kuhn (1992:8-9), dissociating himself from sociological appropriation of his work, has recently proclaimed that SSK, or, more ambiguously, what "has been widely understood" as its claims, is "an example of deconstruction gone mad."8 Compared to other specialties, SSK has few senior advocates or practitioners in the sociological profession, nor, despite its persistent characterization as "fashionable," is association with SSK evidently a sound strategy of career advancement. SSK and Sociology The founding father of the sociology of science, Robert K Merton, worked from the late 1930s through the 1960s to constitute the study of science as a legitimate branch of structural-functionalist sociology, while at the same time he attempted to constitute sociology as "scientific." What counted as "being * ~ a m eZuckerman's recent full-scale survey (1988) of the sociology of science is, by comt parison, notably informed about and courteous toward strands of SSK, while she labors to assimilate this work to the structural-functional tradition with which it is often seen to be in conflict. I should add that, as a participant, I am, of course, wholly satisfied that a sociology of scientific knowledge is both possible and necessary, and that it has accumulated a large body of outstanding empirical work. Likewise, I am satisfied that much-not all-criticism of SSK continues to proceed from an obtuse-and possibly willful-misrepresentation of its central methods and claims (cf Barnes 1994:22-25, Bloor 1991:163-85). Yet my purpose here is not merely to reiterate old arguments in defense of SSK but to try to note some features of the cultural framework in which that misrepresentation is so deeply entrenched. SOCIOLOGY OF SCIENTIFIC KNOWLEDGE 295 scientific" was overwhelmingly taken from formal and informal philosophical models of the natural scientific "method" (e.g. Parsons 1949:Ch 1). The same sensibility that persuaded Merton and his associates that sociological accounting had to stop at the door of scientific method and scientific knowledge (e.g. Merton 1970:xviii-xix, 75, 199-200) also supported the claims of sociology to be a genuine science. Accordingly, the very idea of a sociology of scientific knowledge butted against the self-understanding and legitimation of dominant strands of sociology. It is this circumstance, more than others, that makes the place of SSK so problematic within the overall sociological culture, especially in its North American form. Therefore, Peter Winch's (1958) critique of enterprises that tried to base an understanding of social action on the methods of natural science was decisive for several practitioners of SSK (Collins 1975:216, 1981a:373, Knorr-Cetina 1981a:148-49, Lynch 1993:40-41, 163, 183, 228). If social science could be construed as fundamentally different from natural science-in its objects and in its appropriate methods-then it followed that the opening up of the natural sciences to sociological understanding need not be seen as a threat to sociology. The pertinence of Winch's views indicates the importance to SSK of intellectual resources coming from the margins of the American sociological profession, and, indeed, from outside of sociology proper. !9 Winch'&ookignificantly stimulated curiosity about the later philosophy of Wittgenstein, especially its analysis of the indeterminacy of "rules," while other British practitioners disputed Winch's distinction between sociology and natural science (Bloor 1983). The intellectual mix that in the 1970s inspired the early sociological studies of such British writers as Barry Barnes, David Bloor, HM Collins, Donald MacKenzie, Michael Mulkay, Richard Whitley, and Steve Woolgar included, to be sure, elements of the classic sociological theory of Durkheim and, more diffusely, of Marx, but also the historiography of TS Kuhn, the comparative cultural anthropology of EE Evans-Pritchard, Mary Douglas, and Robin Horton, philosophical work on the categories of sociological explanation by Alisdair MacIntyre, Basil Bemstein's revisionist sociology of language and education, the relativist philosophy of Nelson Goodman, Mary Hesse's neo-Bayesian philosophy of science, and, especially, a vast body of detailed historical work on the natural sciences in their social and cultural contexts. In the early and mid 1980s, SSK received an infusion from practitioners trained in, or attracted by, phenomenological and ethnomethodological traditions. Studies by Michael Lynch, Steve Woolgar, Steve Yearley, and Eric Livingston drew significantly on work by Alfred Schutz and Harold Garfinkel %inch's work, while influential in British sociological theorizing, is referred to little or not at all in standard American surveys. 296 SHAPIN (Lynch 1988, 1993:Ch 4). [It was predominantly these writers who imported the tag "social construction" into SSK, most immediately by way of Berger & Luckmann (1966), though others not primarily indebted to phenomenology soon elaborated a modified conception of "social constructionism," different from both its theoretical begetters and from sociological "labeling theory."]10 More recently, such sociologists as Susan Leigh Star, Adele Clarke, Joan Fujimura, and Chandra Mukerji have effected links between SSK and "Chicago School" interactionist sociology of work, occupations, and culture (e.g. Clarke 1990, Clarke & Montini 1993, Star 1989, Star & Griesemer 1989, Mukerji 1989, Fujimura 1987, 1988). And all through the 1980s, social studies of science have been increasingly preoccupied by challenges to several central descriptive and explanatory categories emerging from a Parisian circle centered on Bruno Latour, whose work was itself fundamentally shaped by Nietzschean and Heideggerian philosophical traditions as well as by the techniques of semiotics and anthropological ethnography (Latour 1987, 1988a, 1993). Only in one respect is SSK typical of the sociological profession: Its practitioners disagree about the very identity of sociology, and, therefore, about the identity of a legitimate sociological framework for the study of their objects. In the late 1970s and early 1980s, Ben-David, meaning to be rude about SSK, observed (1978:203-08, 1981:43-47; cf Zuckerman 1988:513) that few of its leading practitioners were properly trained as sociologists, that they meddled with epistemological concerns best left to philosophers, and that, owing to their amateurism, they were unfamiliar with the history of disaster that was said to be the career of systematic sociological attempts to account for scientific knowledge. Ben-David's description was, to be sure, correct on several points. Few of the founding figures were professionally trained as sociologists (Collins 1983a:267-68). On the other hand, a number had natural science backgrounds that discouraged them from confusing the reality of scientific knowledge-making with textbook idealizations. The field was also particularly receptive to the sociological exploitation of historical and philosophical frameworks developed by such writers as Michael Polanyi (1958) and Thomas Kuhn (1962), who did have extensive natural scientific experience. Moreover, as Ben-David rightly noted, the leading concerns of "British" SSK were philosophical and, in particular, epistemological. If scientific judgment and the growth of knowledge could be adequately accounted for by impersonal canons of evidence, logic, rationality, and, especially, of "the ' O B ~theearly to mid 1970s,phenomenologically inclined sociologists were widely appropriating the tag, and it remains especially fashionable in work on sexuality, deviance, and crime. So far as I can discover, the first uses of the term in titles of studies concerned with science appear in 1976 and 1977: evidently the term reached its height of SSK popularity in the late 1970s andearly 1980s (e.g. Latour & Woolgar 1979. MacKenzie 1981a. and Collins & Pinch 1982). SOCIOLOGY OF SCIENTIFIC KNOWLEDGE 297 scientific method," then, indeed, neither a sociological nor an historically contextual account was appropriate for the interpretation or explanation of scientific knowledge. The "Great Tradition" of Vienna Circle logical empiricism was concerned with providing not a naturalistic account of scientific change and judgment but (as Rudolf Carnap and Hans Reichenbach said) its "rational reconstruction." Yet other philosophers wrote as if "method-stories" were historically adequate, and still others continued to conceive of sociological considerations as potential "pollutants" of authentic science, to be guarded against or relegated to the contingent domain of "contexts of discovery." Accordingly, early SSK took it as a primary task to create a legitimate space for sociology where none had previously been permitted, in the interpretation or explanation of scientific knowledge. In just that sense, SSK set out to construct an "anti-epistemology," to break down the legitimacy of the distinction between "contexts of discovery and justification," and to develop an anti-individualistic and anti-empiricist framework for the sociology of knowledge in which "social factors" counted not as contaminants but as constitutive of the very idea of scientific knowledge (e.g. Bloor 1975, Law 1975; cf Fuchs 1992:Ch 2). SSK developed in opposition to philosophical rationalism, foundationalism, essentialism, and, to a lesser extent, realism. The resources of sociology (and contextual history) were, it was said, necessary to understand what it was for scientists to behave "logically" or "rationally," how it was that scientists came to recognize something as a "fact," or as "evidence" for or against some theory, how, indeed, the very idea of scientific knowledge was constituted, given the diversity of the practices claiming to speak for nature (Bloor 1984a,b, Collins 1981b). The current philosophical tag corresponding to SSK is "social epistemology" (Fuller 1988, 1992). Analytic philosophers of science have not much appreciated, nor in many cases comprehended, the gesture-a "social epistemology" seemed to some a contradiction in terms-and the career of SSK continues to be marked by trench warfare between its practitioners and the dominant tendencies in the philosophy of science (e.g. Brown 1984, 1989, Bloor 1991:163-85). For these reasons, SSK developed partly through efforts to exploit some traditional and nontraditional sociological resources to show-both theoretically and empirically-how a sociology of scientific knowledge was possible, and not as a professional extension of mainstream disciplinary practices into this terrain. On the whole, mainstream sociological practitioners did not want sociology to go in such directions or did not believe that it could be so extended. The over-publicized "warfare" between SSK and the "Mertonians" was, in fact, but a brief early episode in the career of the field and was mainly concerned with such questions of possibility (Collins 1983a:266, 271). SSK practitioners soon found it more satisfying to do the sociology of scientific knowledge than to argue whether it was possible, and by the early 298 SHAPIN 1980s, they were content to point to a body of detailed empirical studies as strong evidence of that possibility (Shapin 1982). Indeed, early practitioners systematically argued that scientific knowledge could be understood in just the same way as one would go about interpreting any other area of culturethere were no special resources or methods required to account for science (Barnes 1974). So a number of important writers, having established that point of possibility to their satisfaction, saw no special reason to persist with the particular study of the natural sciences and moved on (in whole or in part) to applying the methods and resources of SSK to other areas of culture (notably technology and economics), to debates in the philosophy of knowledge and theories of representation, to social theory, and, notably, to participation within such scientific practices as artificial intelligence (e.g. Barnes 1988, Collins 1990, Ashmore et a1 1989). SSK and the Forms of Cultural Inquiry If intellectual influences on SSK and its diverse disciplinary affiliations make the field marginal to the profession of sociology, its preoccupations, circumstances, and several of its findings ought to make it central to the sociological enterprise, and, indeed, to cultural inquiry as a whole. On the one hand, SSK, like any descriptive or explanatory practice, inevitably deploys our current stock of knowledge about what the world, natural and social, is like. However much practitioners in this area may mean to show that such items as "neutrinos," "neurofibrillary tangles," or "social class" are theorized and socially constructed, the realist mode of speech is ineliminable in practice, and the "phenomenological bracketing" that allows analysts to be curious about how such items are constructed is dependent upon a robust realist idiom in speaking about other items. Skepticism, as Wittgenstein said, takes place on the margins of trusting systems, and radical skepticism is radically disruptive of communicative order (Douglas 1986, Shapin 1994:Ch 1). This is no more than to say that sociologists of scientific knowledge "know" the world that science has depicted as securely as any other competent members of the culture, and that they use this knowledge in producing their accounts. On the other hand, the practice that seeks to understand science as an historical and social enterprise also demands that analysts be curious about its findings, including the findings about the natural and social worlds that have to be used to implement that curiosity. The realist mode of speech itself becomes an object of curiosity. In this sense, SSK is prone to tension between how it speaks and what it says, and its practice is irremediably embedded in the objects of its inquiry (Barnes 1981:484, 493). While many philosophical and everyday forms of inquiry seek to justify our intuitions about science-its correspondence, its objectivity, its efficacy, and its progressiveness-SSK takes those intuitions as matters to be interpreted and explained (Collins SOCIOLOGY OF SCIENTIFIC KNOWLEDGE 299 1981b). That makes SSK at times uncomfortable-to both practitioners and readers of their work-but also fundamental to our culture's self-understanding. Uneasiness in inquiry is often-not invariably-a sign that the inquiry is nearing the heart of the matter, and the claimed hyper-awareness of "postmodernity" is played out in SSK in one of its most acute forms. The second reason SSK may arguably be central to the sociological enterprise and to cultural inquiry as a whole flows from the categories that traditionally comprised the sociology of knowledge and the changes wrought on these categories by work over the past quarter-century. As Merton surveyed the field in 1945, the sociology of knowledge was the practice that sought to show the influence of social (or "existential") factors upon "mental productions" (Merton 1973; cf Parsons 1949:14). How did social factors condition the form, content, and dynamics of cognitive products? There was social stuff and there was intellectual stuff, and there were (varying) narratives concerned to bridge the Cartesian gulf. That dualism, and that resulting problematic, were accepted by all theorists, no matter what scheme they proposed for doing the connecting (causal, functional, or symbolic), and no matter what exemptions (typically the mental productions of logic, mathematics, and the natural sciences) they stipulated.I1 The dualism that provided traditional sociology of knowledge with its frame of reference was inherited from ancient lay and philosophical discourse. From the Greek philosophical tradition to early Christianity and on into the culture of seventeenth-century English empiricism and nineteenth-century high romanticism, knowledge was considered to be properly philosophical, sacred, or genuine insofar as the circumstances of its attainment were removed from the domains of the practical and the political (Shapin 1991a,b). Disengagement and disembodiment were ancient tropes of value: Removing knowledge-making from the polis was seen as a technique of transcendence. Accordingly, to say that knowledge was produced in and through mundane interactions between people, as well as between people and reality, was taken just to say that its truth, objectivity, universality, and power were compromised. So far as genuine philosophical knowledge was concerned, the polity was a pollutant. In this way, interpretative and explanatory tasks were embedded-largely unwittingly-in traditional tropes of evaluation. Bacon's idols of the theatre and the marketplace marked the social contamination of knowledge no less "1t has often been insisted that Merton himself (1973) was the father of that "Copernican revolution" in the sociology of science which took tnie as well as false belief for its legitimate subject, from which it follows, in Bourdieu's opinion (1990:297-98), that writers like Barnes and Bloor were merely "crashing through an open door." It is, for all that, remarkable that Merton never purported to produce a sociological account of what has been called "the technical content" of scientific knowledge, while some of his followers continue to insist very vigorously on the impossibility of any such account. 300 SHAPIN than the presentation of Greek and early Christian thinkers as withdrawn and disengaged. From the late 1930s through the 1960s and beyond, the discourse of "internalism" and "externalism" that so fundamentally structured the practice of history and sociology of science took the dualistic juxtaposition of "social" versus "rational," "intellectual," and "evidential" for granted. "The social" was taken as that which was "external" to science, and it was persistently debated by what means authentic science kept "the social" at bay, how and to what extent "social influences" infiltrated science without deleterious effects, or how what seemed to be properly scientific knowledge was "in fact" socially marked ideology (Shapin 1992). Some strands of early SSK and related social historical work did indeed deploy the same society-mind vocabulary as traditional sociology of knowledge. Here the task was taken to be the showing of "social influences" on properly scientific knowledge where such "influences" had previously been reckoned not to act. The taken-for-granted equation between the social autonomy and the truth of knowledge was challenged, and a series of empirical studies sought to establish-without a tone of expost-that even the "hard cases" of claims within the physical and mathematical sciences, taxonomic sciences, and observation-reports were so "influenced": Society, and its concerns, nevertheless "got in" (Shapin 1979, MacKenzie 1978, 1981a). To a number of critics, that sums up the case that SSK argued: Its bearing upon the truth and objectivity of science was taken over from traditional schemes that conceived the social as a "contaminant" (Brown 1989). Where there was "social influence," there the roles of natural reality and rationality were regarded as compromised. However, this sensibility in fact grossly misrepresents SSK's case for "the social." Rather, the claim was that "the social dimension" of knowledge needed to be attended to in order to understand what counts as a fact or a discovery, what inferences are made from facts, what is regarded as rational or proper conduct, how objectivity is recognized, and how the credibility of claims is assessed. The target here was not at all the legitimacy of scientific knowledge but the legitimacy of individualist frameworks for interpreting scientific knowledge. Attention was drawn to "the social dimension," accordingly, not as a pollutant but as a necessary condition for making, holding, extending, and changing knowledge. In just that sense, the language of "the social" as a "dimension," an "influence," or a "factor" to be juxtaposed with the "factors" of evidence and rationality was rendered problematic (Lynch 1991b). And here, arguably, SSK was the primary field in which that challenge to the traditional dualism was laid down. The challenge was expressed in varying idioms. From 1979 Bruno Latour repeatedly pointed out that there was undeniably as much (and arguably more) "politics" within the walls of scientific workplaces as there was outside, and SOCIOLOGY OF SCIENTIFIC KNOWLEDGE 301 that the securing of credibility for scientific claims was a thoroughly social and political process. Thus he highlighted as a potential topic of inquiry the cultural scheme that simply assumed otherwise (Latour 1987: 30, 62). At the same time, the performance of modern political action fundamentally implicated scientific knowledge of what sorts of things existed in the world and how these things acted upon humans. The "missing masses" in existing social and political theory were the "nonhumans" predicated by science and technology. A defensible sociology of science and technology, therefore, had the potential to recast the terms of social theory generally. Signaling the sensibility that sought to remove "the social" from its status as "factor," the second (1986) edition of Latour & Woolgar's Laboratory Life deleted the word "social" from its original (1979) subtitle: The Social Construction of Scientific Facts. To remove "the social" from the idea of scientific knowledge was said to remove its status as knowledge. In a more familiar (anti-)epistemological idiom, Mulkay, Barnes, and Bloor sought, from the early 1970s, to establish the inadequacies of individualism for interpreting scientific, or any other form of, knowledge. Here the Kuhnian framework assumes central significance, not least for appreciating the place of SSK vis-a-vis existing sociological traditions. If, for Merton, the answer to the HobbesParsons social-order problem was supplied, in the case of science, by a set of allegedly unique social norms making up the "ethos of science," for Kuhn-inspired SSK, the regulative principles of social order in science were furnished by scientific knowledge itsev Within traditions of "normal science," authoritative socializing institutions schooled practitioners in exemplars ("paradigms") of what it was to do good science in particular domains. For early modern chernists, Robert Boyle's J-tube experiment defined a model problem and its model solution, including the embodied representation of what it meant for evidence to confirm or disconfirm a theoretical hypothesis; for late twentieth-century molecular biologists, the "central dogma" (by which DNA produces RNA produces protein) similarly structures practitioners' sensibilities about relevant domains of inquiry, about the directionality of molecular cause, and about the locus of biological meaning. From a sociological point of view, Kuhnian SSK is at once conservative and radical. On the one hand, it seeks inter alia to answer traditional questions about the grounds of a communal order, and it does so by pointing to the regulative role of norms. While the regulatory relevance of social maxims ("Be skeptical," "Be disinterested) is doubted, the significance of norms for ensuring order and for marking the boundaries of communities is vigorously respecified and reaffirmed in a new idiom. The solidarity of specialist communities--or such solidarity as is found to exist-is coordinated through their specialist knowledge. Good and bad, proper and improper, interesting and banal scientific behavior is recognized and sanctioned by members' knowledge 302 SHAPIN of the natural world. On the other hand, by arguing that the relevant norms are made of the same stuff as the community's technical knowledge, the Kuhnian move overturns the existing sociology-of-knowledge scheme that asks how "society might influence knowledge." Just because the sway of an evaluative individualism in interpreting our society's most esteemed knowledge has been so strong, SSK's insistence upon a quite elementary feature of the sociological sensibility has seemed to acquire a shockingly radical, even subversive, character. If sociology is the study of the collective aspects of human conduct, then a basic role for the sociological study of scientific knowledge is showing in what ways that knowledge has to be understood as a collective good and its application as a collective process. If there is a fundamental and irreducibly sociological point to be made about scientific knowledge, it is this one. Society-including the specialist societies of scientists-might properly be regarded as a distribution of knowledge, just as the very idea of knowledge depended upon the social relations of knowers (Barnes 1988, Shapin & Schaffer 1985). Following such writers as Simmel (1950: 313) and Polanyi (1958), it has been noted that modern systems of scientific and technical knowledge are highly differentiated and distributed: No one individual keeps the whole of a discipline's knowledge in his or her head, and even the technical knowledge involved in the conduct of a single experiment in modern physics or biology is typically distributed across a range of specialist actors. In a symbolic interactionist idiom, actors in different "social worlds" are invariably involved in the making of scientific goods (Star & Griesemer 1989). And, while this distributed character is very evident in modem scientific practice, in principle it is arguably just as pertinent as a description of the "simpler" scientific cultures of past centuries. The director of a large-scale experiment in high-energy physics does not have direct knowledge of every aspect of that experiment, just as an individual seventeenth-century English natural philosopher would typically not have direct evidential warrant for his knowledge of icebergs, cornets, or the flora of the Americas. As a general matter, practitioners rely massively upon others for their knowledge. For there to be solutions to the problem of knowledge there have to be practical solutions to problems of trust, authority, and moral order (Barnes 1985:49-58, 82-83). Individualist philosophies of knowledge at least since Locke have persistently argued that knowledge is genuine and secure when its warrants are direct, experiential, and individual (Shapin 1994:Ch 5). If that is the case, then the sociological sensibility would suggest that there is perishingly little genuine and secure scientific knowledge in the world. Yet that is not what sociologists of scientific knowledge have argued: Scientific knowledge is as secure as it is taken to be, and it is held massively on trust. The recognition of trustworthy persons is a necessary component in building and maintaining systems of knowledge, while SOCIOLOGY OF SCIENTIFIC KNOWLEDGE 303 the bases of that trustworthiness are historically and contextually variable. This core sociological insight into the collective nature of knowledge has enormous potential to generate detailed comparative studies of the moral economies of science, but, perhaps owing to the largely philosophical concerns of many sociologists of scientific knowledge, the point has as yet been made, for the most part, at a programmatic level.12 A fundamental sociological collectivism applies not just to describing the conditions in which it can rightly be said that individuals have knowledge but also to the means by which knowledge is acquired, applied, and changed. Scientists' knowledge of specialist domains of the natural world, like that of children, is for the most part initially acquired via trusted sources. The proper applications of terms like "chicken," "dog," "electron," and "ideal gas" are not logically fixed; rather, how such terms are used, whether by scientists or laity, is adapted to a range of contingent circumstances, including the weight of custom and convention and the purposes people may have in representing the world. This is the sense in which it is said, following Durkheim and Mauss, that the classification of things reproduces that of people (Bloor 1982). When people confront the experience of their senses, they do so within an already existing structure of knowledge given them by their community and within a structure of purposes sustained by their community. Nor, when new experience is confronted, is it logically determined how such experience is to be sorted out with respect to existing schemes: whether it is to be counted as evidence confirming or disconfirming some theory, whether it is to be bracketed, subjected to taboo, or filed away, to be dealt with another time. It is people's goal-orientation-the pragmatic structure of the community to which they belong-that judges among possible courses of action. Much of the theoretical devel- opment of SSK through the 1970s and early 1980s concentrated upon elaborating a fully general sociological framework for interpreting knowledgeacquisition and concept-application (Barnes 1982a,b,c, 1983). And, despite the fact that this work developed without evident specifiable intellectual "influence" from American pragmatist philosophy, it is wholly compatible with pragmatism, and, by extension, with strands of academic sociology-those of Mead, Blumer, and their progeny-that drew inspiration from James and Dewey. In this way, SSK opposed philosophical rationalism-the view that scientific judgment is sufficiently determined by unambiguous criteria of method-by asserting the contingency and the locality of judgment. Rules did not suffi' * ~ insistence that SSK should be concretely operationalizedin such ways has informed some n criticism that it "has not developed a fully-fledged sociological account of science" (Fuchs 1992:Ch 2, Hagendijk 1994:135). Once more, an accusing finger is pointed at excessively philosophical concerns. 304 SHAPIN ciently explain scientific judgment; the way in which rules were identified and used was itself a topic for contextual inquiry. Why is it that, since one can "rationally" continue the series 2,4,6, 8 ... in any number of ways, the "right way" of going on in an arithmetic class is "10" whereas at an American sporting event it is more likely to be "who do we appreciate?'(Collins 1992:12-16).13 Right conduct is tied to place and purpose. The in-principle "interpretative flexibility" of rules is securely settled in practice by local notions of decorum. By contrast with rationalism, such SSK writers as Barnes and Bloor explicitly endorse a robust realism and, indeed, have noted that the idiom of sociological realism presupposes a corresponding natural realism: "No consistent sociology could ever present knowledge as a fantasy unconnected with our experience of tlie material world around us" (Bloor 1991:33) or "[Tlhere is indeed one world, one reality, 'out there,' the source of all our perceptions ... " (Barnes 1977:25-26, cf 1992, Barnes & Bloor 1982).14What one cannot.do, if one proposes disinterestedly to interpret varying beliefs about nature, is to use one particular account-usually that of modem science-to gauge the validity of others. That would be to include the answer in the premises (Barnes 1992). All institutionalized beliefs about nature are causally connected to reality, and all are on a par with respect