The Sociology of Knowledge: Retrospect and Prospect
Annual Review of Sociology
Volume 9 (1983), 287-310.
Before World War the intellectual climate of
American sociology was congenial to the growth of a sociology of knowledge
Surveys of work in the sociology of knowledge have a formulaic character. Invariably they begin by lamenting the relative neglect of the field by American sociologists, continue with programmatic definitions of its mandate, and conclude with optimistic forecasts of its future growth. In this essay I adhere to the conventions of the genre, for they are dictated by historical evidence. That is, there is no coherent, self-conscious research tradition in the sociology of knowledge as such. Much of the literature in the field is best described as philosophical anthropology - perhaps propaedeutic to analysis of ideas in their social context but hardly equivalent to such analysis. In part, this state of affairs has resulted from the lack of professional consensus on a charter mandate for this subdiscipline - essential to the realization of a research program. Then, too, for the past four decades the intellectual climate of American sociology has been unfavorable for the sociology of knowledge; I return to this problem shortly.
The substantive boundaries of the sociology of knowledge may be set either narrowly or broadly, either restricted to specialized knowledge or extended to cover the commonplace beliefs that inform everyday life. By the latter definition, the field’s theoretical model is comprehensive, delimiting the fundamental processes linking social structure, thought, and action; Durkheim and Mannheim stand as historical patrons of this comprehensive form of the sociology of knowledge, although they did not endorse it consistently, their explanatory expectations varying with the sorts of societies they analyzed and at different stages of their careers (Lukes 1973: 435-49; Douglas 1975; xi- xviii; Simonds 1978: 6). Such contemporary sociologists of knowledge as Berger and Luckmann insist that the field must treat ordinary and esoteric culture alike if it is to address issues of general sociological significance (1967: 12-14).
Nonetheless, my essay surveys the sociology of
knowledge in its narrow form, the analysis of esoteric knowledge; thus, it is
prescriptive as well as descriptive. I
survey research treating what
scientific theories are underdetermined by facts” (Hesse 1978: 1) and that collective agreement on the significance of evidence is achieved through various types of negotiation (e.g. see Bloor 1981). Such a delimitation of the sociology of knowledge serves a programmatic purpose: the definition of a potentially viable research area that has been unjustifiably neglected. Its neglect seems odd when viewed in the light of the history of social science over the past century and a half, for, as Clifford Geertz recently observed, during this period social science has benefitted from virtually every theoretical and methodological scholarly innovation, with the notable exception of one major trend in recent scholarship - the effort to interpret “symbolic action,” to explain symbols that serve as vehicles for conceptions (1973: 208). Certainly the sociology of religion has long been an established specialty in American sociology, and the sociology of science has flourished recently; but research in both fields would surely benefit from consideration of the similarities - and dissimilarities - that obtain among different types of objective culture.
My essay is divided into three sections. In the first I identify the impediments to the development of a research tradition in the sociology of knowledge, impediments based in the character of post-1945 American sociology. In the second I discuss recent theoretical formulations of the sociology of knowledge; though articulated by practitioners of several disciplines - philosophy, literary criticism, the social sciences, the history of art, the history of science, and intellectual history - they display remarkable agreement. Finally, in the third section I describe research illustrating the diversity of approaches to the sociology of knowledge, emphasizing work that realizes the programmatic injunctions expounded in the second section.
Because many of the founding fathers of American sociology were trained abroad, the sociology of knowledge here has both an indigenous and a European ancestry. From the end of the 19th through the first three decades of the 20th centuries, American social thinkers took for granted the basic assumptions of a sociology of knowledge, as many scholars have observed. Whatever their differences, the thinkers of this period agreed that the social world had to be understood in historical terms, not in terms of timeless laws; if laws of social behavior could be determined, they were laws of evolutionary development. Such diverse thinkers as J. Dewey, G. H. Mead, C. H. Cooley, L. F. Ward, A. Small, W. I. Thomas, E. Fans, and Robert Park were not all unilinear evolutionists, but they all charted the course of past and anticipated human history as the progressive triumph of reason over instinct in the guidance of behavior; increasing rationality derived from human adaptive capacity and was
the product of interaction between conscious purposes and material constraints. Different styles of thought prevailed during - and were appropriate to - different stages of evolution, and in any given society the various social strata displayed different outlooks (e.g. Berger 1966; Fuhrman 1980: 212-29; Lewis & Smith 1980: 100-1, 131; Merton 1967a: 562; White 1976).
From the late 1930s through the 1950s
sociologists set themselves tasks very different from those of their
predecessors, rejecting the intellectual tradition appropriate to the
sociology of knowledge (e.g. Kuklick 1973). Certainly
the older sociological approach survived in the still-thriving interactionist
school, but it was upheld only by a minority within the discipline; and
because it was wrenched from its evolutionist matrix it was transformed from a
macro- to a micro-sociological theory (for a survey, see Fisher & Strauss
1978). How had mainstream sociology
changed? Even before the 1930s
sociologists were turning away from analysis cast in historical terms, even
denying the value of historical information about their subjects on the
grounds that knowledge of “previous values and ideals could distort an
objective view of the present” (Karl 1969: 380; and see Barnes 1921).
They sought to establish timeless
sociological laws, transhistorical and transcultural.
Because the sociologists of the 1930s
were repudiating historicism in all its forms, they were unprepared to accept
No less important to their antirelativism was
the methodological stance sociologists adopted in the 1930s.
Determined to make their field truly
“scientific,” led by such propagandists as G. Lundberg and W. F. Ogburn, they
embraced the ideal of general “scientific method,” equally applicable to all
forms of scholarship, which to them entailed both methodological individualism
and collaborative research. Scientific
sociology so defined was institutionalized after World War II at
range” problems. Finally, the notion of a general scientific method was premised on a distinction - central to the philosophy of science in the 1930s - between the context of discovery and the context of justification. According to this formulation, sociological analysis can explain scientists’ choices of research problems, which are conditioned by sociohistorical factors, but is irrelevant to judgment of research results, which are evaluated by universal criteria. Nowadays this distinction is considered specious, for the same cultural factors inform problem selection and the cognitive content of science. Sociology was expected to gain prestige by becoming more scientific, and, indeed, through analogous reforms the natural sciences won increased respect in the mid-19th century [although one should note that many types of scientific research cannot be bureaucratically organized (see e.g. Lundberg 1942; Morrell 1972; Mulkay 1979: 1-26; Ogburn 1930; Shils 1970: 794; R. S. Turner 1982)].
Scientific sociology could not be reconciled
Given the reward structure of sociology, it is not surprising that sociologists of knowledge have practiced externalist analysis, for they thus defined their problems in the terms of more conventional (and more prestigious) sociological specialties. Hence, we may know little about the social meaning of “objective” culture, but we can analyze its production just as we do any form of complex organizational behavior (e.g. Reskin 1977; Swatez 1970); we can assimilate our understanding of the patterns of transmission of knowledge to general theories of communication (e.g. Crane 1972; Kadushin 1976); we can suspend judgment about the relationship between the allocation of prestige among cultural producers and the quality of their work, transmuting this question into the problem of determination of the formal mechanisms by which rewards are distributed (e.g. P. Clark, 1979; Cole et al 1977; Useem 1976), thus steering a neutral course between the irreconcilable conclusions of students of social stratification that high status is either fully deserved or virtually arbitrary; and we can translate analyses of the prestige structure of the scientific community into the terms of those sociologists who have attempted to relate social stratification patterns to the course of the life cycle (Zuckerman & Merton 1973).
How would sociology have to change so as to encourage research on the meaning of culture? As has already been suggested, what is required is the reconciliation of sociology with history. This platitude has long been a staple of textbook wisdom (e.g. Berger 1963: 20, 1681), but there is evidence that the long-awaited reconciliation is now under way, especially in the study of the family, education, and social stratification. In these fields, sociologists and historians are profiting from one another’s work, not merely developing their specialties on parallel courses within the traditional disciplinary division of labor. Of course, in part this disciplinary cooperation represents the triumph of sociological methodology in certain types of historical inquiry, belying earlier sociological objections that historical data could not be analyzed rigorously. But this interdisciplinary exchange also reflects sociologists’ appreciation of the virtues of the interpretive analysis traditional to history (e.g. see Aminzade & Hodson 1982; Aronson 1969: 295-99; Collins 1979; Elder 1974: 1-24; Hareven 1976; Katz 1975: 44-175; Pleck 1976). Why are sociologists now receptive to historical approaches? Since the mid-1960s the discipline has endured a crisis of identity, occasioned by discontent with the previously dominant paradigm - positivist methodology coupled with functionalist theory, which together were incapable of explaining social change (e.g. see Friedrichs 1970: 21). Sociolgists’ concern to comprehend change entails attention to history. The rejection of positivism has taken such extreme forms as the recently formulated ethnomethodological model of behavior, which permits an absolute idealism through which, to quote Gellner (1975: 441), “the objective world evaporates, asymptotically approaches zero.” (For a survey, see Frank 1979.) Hence, sociology should be receptive to the theoretical perspective to which I now turn.
Social scientists have frequently debated the
merits of patterning their fields after the stereotyped model of the natural
sciences partially described above. Both
scientistic social scientists and their opponents typically grant the accuracy
of this model, although the former have concluded that scientific social
science cannot deal with problems of subjective meaning, while the latter
argue that the social sciences must treat such problems because natural and
social phenomena are qualitatively different.
All assume, however, that natural scientific explanations are
monistic, uncompromisingly materialistic, and reductionistic.
But many (if not all) philosophers of
science have rejected the positivistic characterization of science.
Above, I alluded to current
philosophic trends, and below I describe historical and sociological research
that supports philosophic generalizations. I
emphasize here, however, that philosophers of science no longer see their task
as constructing an ideal scientific method, freed of all historical and
personal idiosyncracies, the use of which in any sort of research will yield
“brute data” - data that are “available without any personal discernment or
interpretation on the part of the observer” (Taylor 1973: 56; see also
Bernstein 1976: 5-6; and for a somewhat different narrative of recent
philosophical trends see Laudan 1981). Philosophers
of science whose theories are informed by knowledge of actual (as opposed to
ideal) research behavior agree that in both the social and natural sciences
agreement on evidence is not reached by appeal to the timeless and universal
standards of true scientific method but is contingent on contextual factors of
various sorts (e.g. Bloor, 1983; Finocchiaro 1977; Toulmin 1977).
Indeed, it is difficult to see how
social scientists could reject interpretive analysis of symbolic meaning on
the grounds that it cannot be reconciled with positivist scientific method,
for social scientists have enthusiastically embraced an argument that
discredits such a model of science. That
is, despite Kuhn’s assertion in The Structure of Scientific Revolutions
(1962) that his description of scientific behavior does not apply to the
social sciences, social scientists have cast the histories of their
disciplines in the form Kuhn outlined for the natural sciences.
Paradoxically, they have sustained
their claim to be truly “scientific” on the grounds that natural scientific
development is occasionally discontinuous, marked by irrational “paradigm
shifts” (e.g. Boissevain 1974; Bronfenbrenner 1971;
The positivist tradition has left another legacy to contemporary social science, a legacy more significant for our purposes than the positivist notion of
scientific method. This is the assumption that true science is necessarily reductionist, which excludes interpretive analysis of meaning from scientific activity. We rank the sciences in a hierarchy deriving from Comte’s developmental scheme: Moving from decreasing generality to increasing complexity, the sciences evolve in a specific order - mathematics, astronomy, physics, chemistry, biology, sociology - each science depending on its predecessor for its positive content. Hence, each science gains in specificity by further elaboration of its problems in the terms of the antecedent sciences (e.g. Dolby 1979: 174-85). If we require social science to realize this research program however, we cannot justify research in the sociology of knowledge. If we assimilate the sociology of knowledge to a reductionist program of scientific inquiry, we reduce ideas to epiphenomena of material factors; if ideas merely reflect the social order, and play no role as determinants of social action, the scholar should abandon the study of ideas and turn directly to analysis of material conditions (e.g. H. White 1980: 327). Unless we accept the dualist assumption that human behavior is the product of the interaction between subjective perceptions and material factors, the sociology of knowledge is a trivial enterprise.
Thus, because social scientists have so often appealed to the standards of the natural sciences to justify their activities, it is important to observe that much reputable science has been based on dualist assumptions. As Merz noted in his classic survey of 19th century science, “none of the great men to whom we are indebted for the real extension of our knowledge of biological phenomena” accepted the view that “[v]italism and animism were at an end; there only remained mechanism and materialism” (1965: 399). That is, such scientific giants as R. Virchow, X. Bichat, J. Liebig, C. Bernard, and M. Foster agreed that the characteristics of living matter could not be understood by reducing them to their physical and chemical elements; either the material components of organisms were animated by some extrinsic, inexplicable vital force, or life was an emergent property of the organization of matter - an idea introduced into biology by H. Spencer (Merz 1965: 368-464; and see Bynum 1981). Vitalist commitments explain certain apparent inconsistencies in scientific argument. For example, W. Bateson promoted Mendelism, yet rejected the chromosome theory because it represented materialistic reductionism; Bateson insisted that the organism must be analyzed holistically (e.g. Coleman 1970).
What has been the fate of vitalism in the 20th century? In the life sciences it is still legitimate to argue against physicochemical reductionism on the grounds that the character of an organism is determined by the organization of its constituent parts, rather than the sum thereof. The existence of some autonomous, spiritual life force is not debated openly, but one need not interpret this as evidence of the unqualified triumph of materialism. Rather, the problem of determining the possibilities and limitations of reductionism become nonsen-
sical through philosophic fiat: The positivist
My description of the differences between mechanists and vitalists should disillusion social scientists who imagine that all scientific inquiry proceeds in identical fashion. Further, I have described debates within the natural sciences because analogous disputes have raged within the social sciences. The mechanist position corresponds to the positivist social scientific model already outlined, entailing methodological individualism, the reduction of the behavior of all groups to the sum of their constituents. Employing the terms current among philosophers of social science, one would identify the object of positivist social science as explanation rather than understanding. In positivist social science, subjective elements in social action are ignored because they are irrelevant to the quest for general behavioral laws; hence, scholarly attention is restricted “to the material surface of phenomena, to only such evidence as can be perceived directly and without mediation, which can be measured, quantified, and hence assimiliated to entirely formal and timeless categories of ‘explanation’ (Simonds 1978: 196).
The alternative tradition of social inquiry is
directed toward understanding of social behavior.
This tradition, exemplified and
Purged of its Germanic metaphysics, can this social scientific program be
translated into terms acceptable to contemporary
Anglo Saxon theorists? The rejection
of methodological individualism does not require us to believe in any such
mystical entity as a group mind, affecting but not affected by individuals - a
statement with which
The proper method is said to be that of tracing out the line of the so-called hermeneutic circle, by placing the text to be interpreted within a field of assumptions and conventions to which it contributes and from which it derives its distinctively meaningful character. Finally, the goal of the method is said to be that of “making sense” of the text by exhibiting the coherence between what the agent is doing and the meaning the situation had for him (Skinner 1975:210).
The contextualist program is unexceptionable. Indeed, its leading proponents - Q. Skinner, J. Dunn, J. G. A. Pocock - acknowledge not only the indebtedness to one another and their joint obligation to R. G. Collingwood but also that there is nothing “particularly dramatic or original” in their argument (Skinner 1974: 283). These are the essential components of their scheme:
Purposive social actors are bound by habits of
thought and behavior reflecting their societies’ traditions, but they
manipulate the elements of these traditions; we can understand actors’
purposes only in the contexts of their particular historical circumstances,
and thus the disagreements of those acting at cross purposes are as important
to a recovery of their intentions as are their common assumptions (e.g.
Skinner 1969). Not only can the
contextualist program be reconciled with the method of hermeneutics, the
conception of the actor it entails can be equated with the evolutionism of
theorists such as
Agents act purposively in adaptation to their circumstances, but they are constrained by ecological factors [both cultural and natural (Toulmin 1981)]. Evolutionist explanations (which the historian of early 20th century social thought is bound to recognize), permit accounting for consciousness without recourse to dualist models, for consciousness can be regarded as an adaptive mechanism built into the organism. Furthermore, the gradual process by which cultural patterns are transformed is explicable in evolutionary terms. In response to experience, actors modify particular features of their behavior; they must then modify their use of apparently stable elements in their conceptual schemes in order to reconcile them with these adaptive innovations (see Bloor 1983, who relies heavily on the work of Mary Hesse).
What methodological difficulties are posed by the contextualist program? First is the difficulty of delimiting the context within which culture is to be analyzed. Second, the criteria of plausibility for an interpretive account are problematic. Third, the hermeneutic method may be inseparable from a world view that many find objectionable.
First, the relationships between specialized communities of culture producers and outsiders vary, whether these outsiders are a limited group of consumers or the society as a whole, and no formulae exist to determine the context of highly specialized knowledge. It is a simple matter to determine the context of Hobbes’s political thought: Its restricted context - the community of political theorists - may be defined to comprise those who joined in debate over the nature of political obligation; that its general context was society as a whole is evidenced by the practical issues theorists were addressing - i.e. the problems of the nature of political authority and of the citizen’s duties to the state (Skinner 1972). Yet an inward-looking community may produce esoteric culture explicable only in its own eccentric terms, and contextualist analysis will reveal only the meaning such work has for the community (e.g. see Pocock 1969). We cannot assume, however, that culture has this eccentric character either on the basis of its content or because the community that produces it is corporately organized - permitting but not necessarily resulting in self-regulation according to peculiar norms (e.g. Kuklick 1980). Even so esoteric a field as quantum physics can be seen as the embodiment of a world view held by many outside the scientific community (Forman 1971). Furthermore, as Becker has persuasively argued in his analysis of the differences between artists and craftsmen, the producers of culture may also consciously alter their relationships with their audiences (1978). This methodological obstacle to the practice of contextualism may be more apparent than real, however. Sustained research on any given topic should suggest appropriate boundaries for the communities of culture producers and audiences.
By what criteria do we judge the adequacy of an interpretive account? To quote Taylor, a leading proponent of contextualist analysis,
A successful interpretation is one which makes clear the meaning originally present in a confused, fragmentary, cloudy form. But how does one know that this interpretation is correct? Presumably because it makes sense of the original text: what is strange, mystifying, puzzling, contradictory is no longer so, is accounted for. The interpretation appeals throughout to our understanding of the “language” of expression, which understanding allows us to see that this expression is puzzling, that it is in contradiction to that other, etc., and that these difficulties are cleared up when the meaning is expressed in a new way (Taylor 1971: 5).
Such analysis cannot be expressed in quantitative terms. The relative importance of various factors cannot be weighted; one can only decide whether a factor contributed to the creation of a situation by asking whether its omission from an account would render the interpretation implausible (e.g. Pocock 1969: 191). One can only advise would-be practitioners of contextualist analysis that exhaustive research may permit them to provide the “thick description” Geertz recommends: delimitation of “a stratified hierarchy of meaningful structures in terms of which… [a society’s repertoire of habitual acts] are produced, perceived, and interpreted,” and without which any behavior is meaningless (1973:7). A “thick description” evidently relies on considerable ethnographic or historical evidence, which is presented for scholarly inspection. But our judgment of the interpretation of such evidence ultimately rests on our belief in the reliability of investigators rather than their methods, for their conclusions do not rest on procedures that meet the standard of reproducibility expected of science.
We thus come to the most problematic aspect of
hermeneutic analysis - the undemocratic character of its methodology.
This characteristic is ironic in view
of the objectives of the line of theorists from
assumption that individuals possessed complementary, not equal, native gifts, and therefore must be joined in hierarchical, not egalitarian social relationships (e.g. Manuel 1956). Certainly, organicist notions can be bent to different political purposes. They serve liberal ends if used to support the claim that the state is obliged to serve the needs of all of its citizens, conservative ends if employed to defend the subordination of individuals’ purposes to societal goals (e.g. Clarke, 1978 : 24, 42, 58, 146-53). Nevertheless, organicist arguments rest on an investigative method with unambiguously conservative political implications.
In contrast to the methodology of hermeneutic analysis, that of positivist social science is that supposedly employed by the natural sciences. It ought to be thoroughly impersonal, permitting scholars and research to be judged by universalistic criteria. Research conclusions should be accepted only if they can be reproduced by any competent investigator, who can analyze published data using specified procedures. And meritocratic standards determine individuals’ places in the professional status hierarchy. Furthermore, the reductionist program linked to positivist methodology makes the individual, not the group, the basic unit of social analysis and thus virtually dictates a relatively democratic model of society as the sum of its individual parts. Contextualist analysis illuminates the political significance of positivist social science for its advocates. Historically, middle-class opponents of aristocratic privilege have demanded that scientific conclusions be based on data accessible to everyone, while the elite have argued that their dominance is legitimate because they possess special intuitive powers of understanding (e.g. Jacyna 1980; Shapin 1979). Indeed, from the 19th century on in both Britain and the United States a tradition of middle-class cultural criticism has been directed toward reforming society according to the perfect democratic standards of the scientific community, a tradition exemplified in Merton’s formulation of the “scientific ethos” based on the norms of “universalism,” “disinterestedness,” “communism,” and “organized skepticism;” significantly, Merton defined this ethos in 1942, when powerful fascist regimes threatened the survival of more democratic societies (Hollinger 1983).
Sociologists of knowledge must attempt both explanation and understanding of their material. For though we may observe long-persistent patterns of behavior, explicable through quasi-laws of explanation appropriate to a particular epoch, we must understand that conscious calculations affect individuals’ conduct, not the least because they figure in disjunctive social changes (e.g. Taylor 1971: 32—51). Knowledge claims based on positivist and hermeneutical methods can be faulted. Standards of “rationality” in scientific judgment are not peculiar to science but depend on general cultural standards, and scientific results are frequently accepted for particularistic reasons (see e.g. Barnes & Dolby 1970; Latour & Woolgar 1979). Hermeneutical interpretations
may sometimes be persuasive only if we accept their producers’ pretensions to possess special visionary abilities. Paradoxically, however, our objections to positivism also oblige us to recognize that endorsement of the obvious scholarly virtues embodied in the “scientific ethos” cannot be entirely irrelevant to behavior. Contextualist analysis is necessary to understanding of the sociohistorical locus of ideas, but its practitioners can adhere to rigorous standards, even though these standards may be more akin to those of conventional historical scholarship than to those of scientific research.
The distinction between the contexts of discovery and justification may be philosophically indefensible, but it suggests two broad analytical approaches that sociologists of knowledge must employ simultaneously in order to explain and understand culture. In practical terms this means: first, that we must define the social structural properties of communities of knowledge producers and consumers - types of work organization, patterns of patronage, and audience characteristics; and second that we must interpret the meaning culture has for its producers and consumers. Students of culture have usually focussed on either the first or second of these goals. I now illustrate the range of approaches possible to the sociology of knowledge by describing a variety of studies, not all by sociologists.
In the first part of this essay I described sociological studies treating the social-structural aspects of culture production, ignoring the content of the culture. Research exemplifying this genre reduces behavior to functions of a limited number of factors, and its character can be conveyed with only a few further illustrations. If, for example, degrees of formal religiosity are associated with differences in birth order and gender, religious participation must somehow reflect family socialization practices (e.g. Nelson 1981). If class differences are correlated with attendance at various types of artistic events, culture is but one type of consumer good, its purchase a means to reaffirm status (e.g. DiMaggio & Useem 1978). If religious denominational affiliations are associated with educational, occupational, and financial characteristics, variations in behavior linked to denominational membership must be functions of class, not realizations of distinct religious values (for citations to the voluminous literature on this topic, see Stryker 1981: 212). The careers of culture producers can also be explained without reference to the cognitive aspects of their work. Cultural diversity or uniformity, and the prestige accorded different types of work, can be treated as functions of the organization of financial patronage, its centralization or lack thereof (e.g. P. Clark 1979; Gaston 1978). Even cultural innovation may be explained as the product of simple careerism,
the desire to create opportunities for personal advancement where none previously existed (a subordinate theme in the analysis of Becker 1978; the crux of the argument in Ben-David & Collins 1966).
The organizational structure and cognitive content of culture are interdependent, however, and much can be learned about the content of cultural enterprises from analysis focused on their organizational features. The well-known relationship between the organizational and doctrinal characteristics of groups classified along the sect-church continuum requires no elaboration here; types of political parties can be similarly described (e.g. Duverger 1954). The organization of scientific work is inextricable from its content: Advances in mathematics depend on unpredictable inspiration, and work groups in mathematics are the smallest of any scientific field; on the other hand, the nature of much research in experimental physics, and the high capital investment in technology required to solve many of its problems, are linked to the growth of enormous, hierarchically organized work groups, many of the members of which perform routinized tasks (e.g. Hagstrom 1976). Although the patrons of elite forms of musical and visual arts are similar, the institutional loci of these arts impede musical innovation and encourage innovation in the visual arts: Because concerts are highly organized, expensive activities, symphony orchestras cannot risk performing avante-garde music; in contrast, museums build their collections piecemeal, and art objects are admired individually, so museums can randomize the risks they take in purchasing novel works (Zolberg 1980).
Changes in the organization of cultural production affect the content of culture. For example, the limewood sculptors of Renaissance Germany belonged originally to guilds including other types of craftsmen and were prevented by guild regulations from exercising certain skills; their sculpture was highly conventional. When the sculptors achieved monopolistic control of sculpture production, the trend toward development of individual artistic styles accelerated (Baxandall 1980: 111-27). Because the French Impressionists’ social network was restricted to artists, art dealers, and art critics, they focussed on innovative painting techniques, rejecting the earlier artistic objective of conveying moral imperatives through narrative painting. In contrast, their Russian contemporaries, the Critical Realists, adhered to the aesthetic standard the Impressionists repudiated and belonged to the occupationally diverse social network of the politically committed Russian intelligentsia (Greenfeld 1981). The changed organization of American science after World War I, effected by the emergence of foundations as patrons of research, led to the creation of a new role in the scientific community. The “science manager,” a foundation employee, appraised scientific trends and allocated funds according to a hierarchy of research priorities, exaggerating such trends as the transfer of models and techniques from the physical to the biological sciences and encouraging the
growth of new disciplines (Kohler 1979). The American publishing industry has recently been reorganized. Once-independent small firms, concerned to reconcile artistic and commercial values, have become conglomerate subsidiaries. Consequently, publishers’ lists are dominated by books with mass market appeal - nonfiction works and popular novels suitable for sale to movie companies [which may share conglomerate ties (Whiteside 1981)].
Worthy though the study of the relationship between the organizational structure of and substantive content of culture may be, it tells us nothing about the meaning of culture to its producers and consumers. Can we understand how culture reflects world views and affects perceptions merely from examination of texts? In his classic Art and Illusion, E. H. Gombrich demonstrates how our perceptions may be conditioned, showing, for example, how Dürer’ s misrepresentation of the rhinoceros, based on secondhand evidence and published in 1515, influenced future representations of the beast for almost two centuries, even those drawn from life (1961: 8 1-82). Such analysis depends only on study of pictures, but it aids our understanding of the way knowledge is acquired and sustained. Or, it may be argued that the sociohistorical conditions that engender cultural forms are so fundamental to the traditions of our civilization that we understand them from our own experience. Inspecting representations of the female nude in Western Art from the Middle Ages to the present, J. Berger has argued that the consumer of such art is implicitly male and that women are depicted so as to reinforce male notions that they are fundamentally passive and seductive. Because this view of women has been - and, alas, remains - so characteristic of our culture, we can accept his interpretation of the social meaning of art on the basis of everyday knowledge (1972: 45-64). Scholarship of this sort supplements rather than supplants contextualist analysis, however, for psychological propensities and sustained traditions are modified by peculiar historical circumstances.
If texts are analyzed closely and seen against
the backgrounds of the traditions they sustain or overturn, little additional
sociohistorical evidence may be required to determine their sociological
significance. For example, from the
19th through the 20th centuries, the medical model of the body changed in
tandem with the changing doctor-patient relationship in American medicine.
Initially, disease was conceptualized
and the results of treatment interpreted in a fashion that made the
doctor-patient relationship relatively egalitarian; medical theories and
practice changed as doctors achieved superior professional status (
aesthetic standards promoted by government
patrons from the mid-18th century, artists intended their narrative pictures
to inspire behavior (T. J. Clark 1973: 78-99; Fried 1975).
If we analyze the avant-garde culture
One way to express the partisan character of
culture in the absence of exhaustive information about the characteristics of
society as a whole is to establish elective affinities between the world view
embodied in cultural constructs and the social interests of groups that
endorse them. For example, we can see
A thorough analysis of the production and reception of culture must treat both the organization of culture production and the subjective meaning of culture. None of the studies so far described in this section does this. I now discuss four contextualist exemplars, chosen because they deal with different
types of culture and illustrate different types of contextualist analysis: Radway’s interpretation of the appeal of “romance” fiction in the United States today (1984); MacKenzie’s study of the development of statistics in Britain between 1865 and 1930 (1981); Fleck’s description of the gradual emergence of the modern concept of syphilis from the 15th to the 20th centuries (1979); and Baxandall’s study of the 15th century Italian painting (1972).
Although some of its elements are as old as the novel itself, the contemporary romance is a commodity manufactured by commercial publishers. Standard market research techniques determine the preferences of readers who constitute the established market for romances: women between 20 and 40 who have had little, if any, higher education; manuscripts are tailored to suit readers’ preferences, and books are distributed in supermarkets and drugstores where their buyers routinely shop. Thus, Radway’s investigation of the publishing industry confirms Whiteside’s description [given above (1981)]. In order to determine the cultural significance of the romance, however, she also analyzed the books’ content and interviewed a network of romance readers. The conventional romance portrays an intense emotional relationship, leading to marriage, involving a successful man, the embodiment of traditional masculinity, whose steely exterior masks a capacity for nurturance and devotion, and a woman whose innocence and dependence embody traditional femininity. Textual analysis alone suggest that romance readers are persuaded to accept the limitations of their lives and the inevitability of patriarchial domination. Yet systematic analysis shows readers’ responses to the novels to be various. Certainly some accept the glorification of conventional marriage by glamorous fictional characters. For others, however, reading romances is an act of protest against their lot. Romances are utopian literature, revealing the inadequacies of their readers’ lives; the act of reading such books is one of protest against socially prescribed self-denial, an escape from the unremitting demands of husbands and children (Radway 1984).
During the late 19th through the early 20th century, statistics in Britain developed within a nested set of contexts: the statistics community itself, affiliated directly or indirectly with academic institutions and divided into identifiable groups or “schools”; the larger academic community, within which the statisticians competed for jurisdiction over a piece of intellectual terrain - the explanation of biological heredity; and the general social context, within which alternative research programs suited the objectives of different segments of the middle and upper classes. Within each of these defined populations, opinion was divided on a single issue - that practicability of eugenic reform. The followers of Pearson and Yule conducted an esoteric debate about the fundamental premises of statistical techniques: Should they assume that every variable had a bivariate normal distribution, as Pearson contended? Their debate had practical significance. Statistical findings supported policy recom-
mendations - eugenic or any other - only if population characteristics could be inferred from samples. The biometricians who followed Pearson used statistics to explain heredity in orthodox Darwinian fashion, in opposition to the Mendelian geneticists, and their different explanations corresponded to the middle-class model of social change in opposition to the aristocratic model, as described above (Coleman 1970). If evolution were unpredictable, as the Mendelians contended, it could not be controlled by human intervention, as the eugenicists expected. Why was eugenics a quintessentially middle-class reform movement? Implementation of its program would allow the elevation of the new class of university-trained “brain workers” to the positions of social managers, who would make society as a whole more “efficient” through the application of value-neutral technical “expertise.” Statistical techniques such as Pearson’s correlation coefficient could be used to demonstrate the legitimacy of their claims to power, for statistical evidence proved that ability was transmitted biologically and that it was differentially distributed among social classes. The existing pattern of social stratification was thus for the most part justifiable, a natural order, but aristocrats’ eminence was undeserved. A just society was a meritocratic one, governed by the biologically superior middle class (MacKenzie 1981).
The contextual explanation of the development of the modern concept of syphilis is less detailed than the above examples, for it deals with a centuries-long process. Yet this process must be treated as a whole, for the Renaissance conception of venereal disease as a “carnal scourge” manifested in “bad blood” was finally translated into the Wasserman blood test, which distinguished one type of venereal disease from all others. Popular notions thus informed the research of scientific communities, which Fleck terms “thought collectives” - groups defined both by theoretical orientations and by commitment to the employment of specific instruments and techniques. The “thought collective” is a diffuse concept, for it comprehends groups ranging from a few individuals, who will behave differently with one another than they would working with other individuals, to organized scientific disciplines. But by analyzing scientific thought in holistic terms, Fleck is able to integrate the diverse factors that figure in scientific judgment: “tacit knowledge” - procedural skills that can be learned only from personal contact with a successful scientist, rather than from impersonal formulae; specialized scientific as well as commonsensical standards that determine which scientific problems will be selected and which solutions will be accepted; and the “active” and “passive” elements in scientific discovery. His distinction between the “active” and “passive” components of research is especially valuable, for it qualifies the relativism that characterizes much sociology of scientific knowledge, to the discomfort of many (e.g. Horton 1979). Scientists confront the phenomena they investigate in a culturally mediated fashion; the solutions they achieve represent the interaction
between cultural factors and natural regularities (Fleck 1979; and for an exhaustive survey of the history of science literature susceptible to reanalysis in contextuahist terms, see Shapin 1982).
Michael Baxandall’s analysis of Renaissance Italian painting is contextualist explanation of the highest order. Paintings were relatively standardized commodities, produced in hierarchically organized workshops. Contractual agreements between artists and their merchant patrons specified the paintings’ content, and the patrons’ commercial identity and religious beliefs affected every aspect of painting. Paintings were vehicles for conspicuous consumption; patrons displayed their wealth by commissioning elaborate compositions, lavish use of expensive pigments, and work predominantly executed by masters rather than assistants. The message of the paintings cannot be understood fully without knowledge of the patrons’ commercial values and skills: Narratively significant elements in the paintings were indicated by then-expensive pigments, gold and ultramarine; and the commercial classes of Renaissance Italy possessed the perceptual skills to recognize the three-dimensional relationships conventionally portrayed in the paintings, as we do not, for these skills were routine to their commercial life, standard means of gauging volumes of goods with geometric techniques. The attitudes of figures in the paintings conveyed emotional significance through then-conventional body language. And the religious significance of the paintings was obvious to those who had detailed knowledge of the life of Christ, knew precisely what moment in the narrative was being portrayed, and understood the religious emotion this moment should evoke (Baxandall 1972).
Evidently, different cultural forms and different historical periods cannot be analyzed contextually in identical fashion. It is relatively simple to provide a contextualist explanation for political thought, for its relationship to current events is fairly straightforward. Networks of producers of other sorts of culture - science, religion, art forms - may be delimited, although the intensity of their interactions with one another will vary. Extrapolating from the content of these forms of objective culture to the larger society is problematic, however, and the degree to which culture serves particular social interests and mirrors commonsense notions will have to be determined for every case. Sociohistorical circumstances determine the degree to which cultural forms can be assumed to have collectively appreciated meanings. The producers of scientific culture necessarily agree on the meaning of their work to a considerable extent, for their activity represents a search for agreement. Nevertheless, the degree to which art production represents a collective enterprise similar to science varies with historical circumstances (e.g. Kuhn 1977). In a relatively homogeneous culture, such as that of Renaissance Italy, we may assume that the public’s appreciation of art’s meaning is relatively homogeneous. But the contemporary experience of religion and various art forms is highly privatized.
Sociologists might despair of the practicability of contextuahist analysis, given all of the above qualifications of possibilities of generalization. They should be consoled, however, by reflection on the accumulated results of the positivist program in the sociology of knowledge. What conclusions have we reached from repeated efforts to test general propositions? Is Protestantism associated with worldly achievement? Do scientists adhere to universalist norms? Our results are inconclusive; they must be, because such questions make no sense unless they are posed within particular sociohistorical contexts. In paradoxical contrast, historical research is cumulative. A study such as MacKenzie’s could not have been written unless historians and sociologists had devoted considerable attention to various elements that figure in his account - the development of evolutionary theories, middle-class reform movements, and British social stratification patterns. Persuasive contextualist explanation is difficult, requiring extensive documentation that a single scholar is rarely capable of providing; but its very difficulty makes it a task worth attempting.
I would like to thank Mark B. Adams, Richard P. Gillespie, and Robert Alun Jones for their help in the preparation of this essay.
Aminzade, R., Hodson, R. 1982. Social mobility in a mid-nineteenth century French city. Am. Sociol. Rev. 47: 441-57
Aronson, 5. 1969.
Obstacles to a rapprochement between history and sociology: a
sociologist’s view. In
Interdisciplinary Relationships in the Social Sciences, ed. M. Sherif, C.
W. Sherif, pp. 292-304.
Ashcraft, R. 1981. Political theory and political action in Karl Mannheim’s thought: reflections upon ideology and Utopia. Comp. Stud. Soc. Hist. 23: 23-50
Barnes, H. E. 1921. The development of historical sociology. Proc. Am. Sociol. Soc. 16: 17-49
Barnes, S. B., Dolby, R. G. A. 1970. The scientific ethos: a deviant viewpoint. Eur. J. Sociol. 11: 3-25
Baxandall, M. 1972. Painting and Experience in
Baxandall, M. 1980.
The Limewood Sculptors of
Becker, H. 1978. Arts and crafts. Am. J. Sociol. 83: 862-89
Ben-David, J., Collins, R. 1966. Social factors in the origin of a new science. Am. Sociol. Rev. 31: 451-65
Berger, J. 1972. Ways of Seeing.
Berger, P. 1963. Invitation to Sociology. Garden City, NY: Doubleday. 191 pp.
Berger, P. 1966. Identity as a problem in the sociology of knowledge. Eur. J. Sociol. 7: 105-15
Berger, P., Luckmann, T. 1967. The Social Construction of Reality. A Treatise in the Sociology of Knowledge. NY: Anchor Books. x + 219 pp.
Bergner, J. T. 1981. The Origin of Formalism
in Social Science.
Bernstein, R. J. 1976. The Restructuring of Social and Political Theory. NY: Harcourt Brace Jovanovich. xxiv + 286 pp.
Bloor, D. 1981. Sociology of (scientific)
knowledge, In Dictionary of the History of Science, ed. W. F. Bynum, E.
J. Browne, R. Porter, pp. 391-93.
Bloor, D. 1983. Durkheim and Mauss revisited. Classification and the sociology of knowledge. Stud. Hist. Philos. Sci. 13: 267-98
Boissevain, J. 1974. Toward a sociology of social anthropology. Theor. Soc. 1: 211-30
Bottomore, T. B. 1956. Some reflections on the sociology of knowledge. Br. J. Sociol. 7: 52-58
Bronfenbrenner, M. 1971. The ‘structure of revolutions’ in economic thought. Hist. Polit. Econ. 3: 136-51
Bynum, W. F. 1981. Vitalism. See Bloor 1981, pp. 439-40
Clark, P. P. 1979. Literary culture in
Clark, T. J. 1973. Image of the People.
Gustave Courbet and the
Clarke, P. 1978. Liberals and Social
Cole, S., Rubin, L., Cole, J. 1977. Peer review and the support of science. Sci. Am. 237 (4): 34-41
Coleman, W. 1970. Bateson and chromosomes: conservative thought in science. Centaurus 15: 228-314
Collins, R. 1979. The Credential Society. NY: Academic. vii + 222 pp.
Crane, D. 1972. Invisible Colleges.
Demerath, N. J. III, Roof, W. C. 1976. Religion - recent strands in research. Ann. Rev. Sociol. 2: 19-33
DiMaggio, P., Useem, M. 1978. Social class and arts communication: the origins and consequences of class differences in exposure to the arts. Theor. Soc. 5: 141-61
Dolby, R. G. A. 1979. Classification of the
sciences. The nineteenth century tradition, In Classifications in their
Social Context, ed R. F. Ellen, D. Reason, pp. 167-93.
Duverger, M. 1954. Political Parties.
Elder, G. H. Jr. 1974.
Children of the Great Depression.
Finocchiaro, M. A. 1977. The uses of history in the interpretation of science. Rev. Metaphys. 31: 93-107
Fisher, B., Strauss, A. 1978. Interactionism. In A History of Sociological Analysis, ed. T. Bottomore, R. Nisbet, pp. 457-98. NY: Basic Books. xvi + 717 pp.
Fleck, L. 1979 [orig. 1935]. Genesis and
Development of a Scientific Fact, ed. T. J. Trenn, R. K. Merton. Transl.
F. Bradley, T. J. Trenn. Foreward T.S. Kuhn.
Forman, P. 1971.
Frank, A. W. III. 1979. Reality construction in interaction. Ann. Rev. Sociol. 5: 167-91
Fried, M. 1975. Toward a supreme fiction:genre and beholder in the art criticism of Diderot and his contemporaries. New Lit. Hist. 6: 543-85
Friedrichs, R. 1970. A Sociology of Sociology. NY: Free Press. xxiii + 429 pp.
Fuhrman, E. R. 1980. The Sociology of
Furman, F. K. 1981. Ritual as social mirror and agent of cultural change: a case study in synagogue life. J. Sci. Stud. Rel. 20: 228-41
Gaston, J. 1978. The Reward System in British and American Science. NY: Wiley. 204 pp.
Geertz, C. 1973a [orig. 1964]. Ideology as a cultural system. In The Interpretation of Cultures, pp. 193-233. NY: Basic Books. ix + 470 pp.
Geertz, C. 1973b. Thick description: toward an interpretive ‘theory of culture, see Geertz 1973a, pp. 3-30.
Gellner, E. 1975. Ethnomethodology: the re-enchantment industry or the Californian way of subjectivity. Philos. Soc. Sci. 5: 43 1-50
Gombrich, E. H. 1961. Art and Illusion. NY: Pantheon Books. xxxi + 466 pp.
Graham, L. 1977. Science and values: the eugenics
Greenfeld, L. 1981. The social context of
artistic style: impressionism in
Hagstrom, W. 0. 1976. The production of culture
in science. In The Production of Culture, ed. R. A. Peterson, pp.
Hareven, T. 1976. Modernization and family history: perspectives on social change. Signs 2: 190-206
Hill, C. 1974. William Harvey and the idea of
monarchy. In The Intellectual Revolution of the Seventeenth Century,
ed. C. Webster, pp. 160-81.
Hollinger, D. A. 1983. The defense of democracy and Robert K. Merton’s formulation of the scientific ethos. Current Perspectives on the History of the Social Sciences, Vol. 4, Knowledge & Society, ed. R. A. Jones, H.
Kuklick, pp. 1-15.
Jacyna, L. S. 1980. Science and social order in
the thought of A. J. Balfour.
Horton, R. 1979. Material-object language and
theoretical language: towards a Strawsonian sociology of thought. In
Philosophical Disputes in the Social Sciences, ed. S. C. Brown, pp.
Kadushin, C. 1976. Networks and circles in the production of culture. See Hagstrom 1976, pp. 107-22
Karl, B. 1969. Presidential planning and social science research: Mr. Hoover’s experts. Perspect. Am. Hist. 3:347-409
Kohier, R. E. Jr. 1979. Warren Weaver and the
Rockefeller Foundation Program in molecular biology: a case study in the
management of science. In The
Sciences in the American Context: New Perspectives, ed. N. Reingold, pp.
Kuhn, 1. 5. 1962. The Structure of Scientific
Kuhn, T. S. 1977 [orig. 1969]. Comment on the
relations of science and art. In The Essential Tension, pp. 340-51.
Kuklick, H. 1973. A ‘scientific revolution’: sociological theory in the United States, 1930-1945. Sociol. Inq. 43: 3-22
Kuklick, H. 1980. Boundary maintenance in American sociology: limitations to academic professionalization. J. Hist. Behav. Sci. 16: 201-19
Laudan, L. 1981. The pseudo-science of science. Philos. Soc. Sci. 11: 173-98
Latour, B., Woolgar, 5. 1979. Laboratory Life.
Lewis, J. D., Smith, R. L. 1980. American
Sociology and Pragmatism: Mead,
Lukes, S. 1973. Emile Durkheim. His Life and
Work. A Historical and Critical Study.
Lundberg, G. A. 1942. Operational definitions in the social sciences. Am. J. Sociol. 47: 727-43
Manuel, F. E. 1956. From equality to organicism. J. Hist. Ideas 17: 54-69
MacKenzie, D. A. 1981. Statistics in
Merton, R. K. 1967a [orig. 1941]. Karl Mannheim and the sociology of knowledge. In Social Theory and Social Structure, pp. 543- 62. NY: Free Press. xxiii + 702pp. Enlarged ed.
Merton, R. K. 1967b [orig. 1945]. The sociology of knowledge. See Merton 1967a, pp. 510-42
Merz, J. 1. 1965 [orig. 1904]. A History of
European Thought in the Nineteenth Century, Vol. II. NY:
Morrell, J. B. 1972. The chemist breeders: the research schools of Liebig and Thomas Thomson. Ambix 19: 1-46
Mulkay, M. 1979. Science and the Sociology of
Nelson, B. 1970. Is the sociology of religion possible? J. Sci. Stud. Rel. 9: 107-11
Nelson, H. M. 1981. Religious conformity in an age of disbelief: contextual effects of time, denomination and family processes upon church decline and apostasy. Am. Sociol. Rev. 46: 632-40
Ogburn, W. F. 1930. The folkways of a scientific sociology. Proc. Am. Sociol. Soc. 24: 1-11
Pleck, E. 1976. Two worlds in one: work and family. J. Soc. Hist. 10: 178-95
Pocock, J. G. A. 1969. The history of political
thought. In Philosophy, Politics, and Society,
ed. P. Laslett, W. G. Runciman, pp. 183-202.
Radway, J. A. 1984. Reading the Romance:
Women, Patriarchy, and Popular Literature.
Reskin, B. F. 1977. Scientific productivity and the reward structure of science. Am. Sociol. Rev. 42: 491-504
Rosenberg, C. E. 1979. The theraputic revolution.
In The TheraputicRevolution, ed. M. J. Vogel, C. E. Rosenberg, pp.
Rueschemeyer, D. 1977. On the non-reception of Karl Mannheim’s Wissenssoziologie in American sociology. Paper presented to the Oxford Conf. Res. Comm. Hist. Sociol. Int. Sociol. Assoc.
Schorske, C. E. 1980. Explosion in the garden:
Kokoschka and Shoenberg. In Fin-de-Siècle
Shapin, 5. 1979. The politics of observation:
cerebral anatomy and social interests in the
Shapin, 5. 1982. History of science and its
sociological reconstructions. Hist. Sci. 20: 157-211
Shils, E. 1970. Tradition, ecology, and institution in the history of sociology. Daedalus 99: 760-825
Simonds, A. P. 1978. Karl Mannheim’s Sociology
Skinner, Q. 1969. Meaning and understanding in the history of ideas. Hist. Theor. 8: 3-53
Skinner, Q. 1972. The context of Hobbes’s theory of political obligation. InHobbes and Rousseau, ed. M. Cranston, R. S. Peters, pp. 109-42. Garden City, NY: Doubleday. ix + 505 pp.
Skinner, Q. 1974. Some problems in the analysis of political thought and action. Polit. Theor. 2: 277-302
Skinner, Q. 1975. Hermeneutics and the role of history. New Lit. Hist. 7: 209-32
Stryker, R. 1981. Religio-ethnic effects on attainments in the early career. Am. Sociol. Rev. 46: 212-31
Swatez, G. 1970. The social organization of a university laboratory. Minerva 8: 36-58
Taylor, C. 1971. Interpretation and the sciences of man. Rev. Metaphys. 25: 3-51
Taylor, C. 1973. Peaceful coexistence in psychology. Soc. Res. 40: 55-82
Toulmin, S. 1977. From form to function: philosophy and history of science in the 1950s and now. Daedalus 106: 143-62
Toulmin, S. 1981. Evolution, adaptation, and
human understanding. In Scientific Inquiry and the Social Sciences, ed.
M. B. Bressler, B. E. Collins, pp. 18-36.
Turner, F. M. 1978. The Victorian conflict
between science and religion: a professional dimension.
Turner, R. 5. 1982. Justus Liebig versus Prussian
chemistry: reflections on early institute-building in
Useem, M. 1976. State production of social knowledge: patterns in government financing of academic social research. Am. Sociol. Rev. 41: 613-29
Weimer, W. B.,
White, H. 1980. Literature and social action: reflections on the reflection theory of literary art. New Lit. Hist. 11: 368-80
White, M. 1976 [orig. 1949]. Social Thought in
Whiteside, T. 1981. The Block-Buster Complex,
Conglomerates, Show Business, and Book Publishing.
Wisdom, J. 1970. Situational individualism and
the emergent group-properties. In Explanation in the Behavioural Sciences,
R. Berger, F. Cioffi, pp. 271-96.
Wolff, J. 1975. Hermeneutic Philosophy and the
Sociology of Art.
Zolberg, V. 1980. Displayed art and performed music. Selective innovation and the structure of artistic media. Sociol. Q. 21: 219-31
Zuckerman, H., Merton, R. K. 1973 [orig. 1972].
Age, aging and age structure in science. In The Sociology of Science,
ed. N. W. Storer, pp. 497-559.