American Behavioral Scientist 1976 Hagstrom 753 68

http://abs.sagepub.com/Scientist

American Behavioral

http://abs.sagepub.com/content/19/6/753The online version of this article can be found at:

DOI: 10.1177/000276427601900606 1976 19: 753American Behavioral Scientist

Warren O. HagstromThe Production of Culture in Science

Published by:

http://www.sagepublications.com

can be found at:American Behavioral ScientistAdditional services and information for

http://abs.sagepub.com/cgi/alertsEmail Alerts:

http://abs.sagepub.com/subscriptionsSubscriptions:

http://www.sagepub.com/journalsReprints.navReprints:

http://www.sagepub.com/journalsPermissions.navPermissions:

http://abs.sagepub.com/content/19/6/753.refs.htmlCitations:

What is This?

- Jul 1, 1976Version of Record >>

at European University in St Petersburg on February 28, 2012abs.sagepub.comDownloaded from

[753]

The Production of Culturein Science

WARREN O. HAGSTROMUniversity of Wisconsin

Hagstrom employs a production perspective to integrate what is knownand reveal what is not known about the relationship between the subjectmatter of scientific specialties and the structure of the research groups. Healso compares the structure of science and engineering work groups.

Metaphor has always played an indispensable role in thedevelopment of scientific theories as well as in the developmentof other types of culture (Kaplan, 1964: 258-288). The notionof the production of culture is such a metaphor, suggesting howwe can apply theories about the production of commodities andservices to culture and perceive similarities among diverse realmsof culture. I shall proceed by taking the metaphor seriously, andI will only briefly indicate some of the problems that may beproduced by taking it too seriously. However, I cannot forbearpointing out that this notion that culture may be produced likeanything else has long since been savagely criticized. Almost250 years ago Jonathan Swift satirized the notion in CullipersTralels. He describes the Grand Academy of Lagado, one of theprojects of which was &dquo;for improving speculative knowledge, bypractical and mechanical operations,&dquo; in which &dquo;the mostignorant person, at a reasonable charge, and with a little bodilylabour, might write books in philosophy, poetry, politics, laws,mathematics, and theology, without the least assistance fromgenius or study&dquo; (CulliFer 5 Trapel5, Part III, ch. V, first


Highlight

[754] AMERICAN BEHAVIORAL SCIENTIST

published 1726). How might Swift react to the fact that there isnow a small industry producing culture about him! In the 40years prior to 1967, about 600 books and articles about himand his work appeared (Mayhew, 1967: 187), and in 1973alone, 33 works on Swift were published by as many authors(Philological Quarterly, 1974).

Anyway, scientists work on nature, applying human, animal,and inanimate energy to produce information which, suitablyprocessed, is disseminated to consumers. I shall begin bydiscussing the organization of work groups in science, proceedto discuss the coordination of the activities of related groups,continue to consider the dissemination of products to ultimateconsumers, and conclude by considering some of the limitationsof the metaphor as it pertains to the consumption, destruction,or transformation of scientific culture. I will not trouble muchto make explicit comparisons with other realms of culture,though potential similarities and differences should be apparentthroughout.THE ORGANIZATION OF WORK GROUPS

Scientific research implies ignorance and uncertainty. Theuncertainty is primarily in the task or technology domain; theuncertainty about the provision of resources or the marketingof products is much less serious. The degree of uncertaintyvaries considerably among the sciences. When the production ofinformation about the empirical world is most highly rou-tinized, workers are likely to be characterized not as scientists,but as technicians.

Charles Perrow ( 1967, 1972), March and Simon (1958: 141 ff.), and others have pointed out how varying degrees ofuncertainty affect several characteristics of organizations. Un-certainty is reduced and behavior capable of being routinizedwhen workers deal with uniform types of problems and whenthe search for ways to deal with exceptions is seldom needed oreasily accomplished. These theories have been applied to scienceby Lowell Hargens (1975: ch. III) in a comaprative study ofmathematicians, chemists, and political scientists. He found thework of mathematicians least routinized (cf. Fisher, 1973).


Highlight

Highlight

Highlight

UserNote - , -, ,

[755]Hagstrom / PRODUCTION IN SCIENCE

Mathematicians will say such things as &dquo;I feel that Im stuckmost of the time. The real question seems to be whether or notthere is any period when Im not stuck,&dquo; and &dquo;Sometimes Ispend literally weeks to find a two-line proof. Usually, Im justwaiting around for something to happen&dquo; (Hargens, 1975: 40ff.). By contrast, work in political science and chemistry is moreroutine, chemistry being the most routinized of the three fieldson most indicators. Hargens found that the more routinized thefield, the better able workers were to get things done byplanning work in detail and by setting deadlines for themselves.Correspondingly, in the more routinized fields, a larger amountof the variance in research productivity could be explained bythe amount of time spent on research and the scientistssubjective evaluations of their own personal efficiency.When work is relatively routine, scientists are enabled to

delegate many tasks to less-skilled subordinates, usually gradu-ate students, but also technicians and postdoctoral fellows.Chemists especially, but political scientists too, have moreassistance than mathematicians, and the amount of assistancecan account for well over half the variance in the productivityof chemists, much less but still a significant amount of thevariance in productivity of political scientists, and essentiallynone of the variance in productivity of mathematicians.Chemists, then, rationally attribute low productivity to a lackof assistance. Mathematicians are more likely to attribute theirlow productivity to their personal deficiencies, and thusmathematicians are more likely to be depressed by periods oflow productivity than those in the other two fields.

As Randall Collins has noted (1975: 506-514), Joan Wood-wards (1965) theory of the relation between technology andthe structure of organizations can be applied readily to science.Given the typically nonroutine nature of scientific work, it canbe characterized as prototype, unit, or small batch production.Following Woodward, we would expect work groups to berelatively small, work to be relatively labor-intensive, and a highproportion of workers engaged in direct production instead ofsupport services. Work groups should have few hierarchicallevels, there should be a low span of control, and there should


Highlight

Highlight

[756]AMERICAN BEHAVIORAL SCIENTIST

be a high ratio of supervisory personnel to those supervised.Most science is little science, and the data we have about workgroups are consistent with the preceding hypotheses. In mysurvey of representative samples of academic scientists (Hag-strom, 1967, 1974), I found that work groups averaged in sizefrom about four in mathematics to more than nine inexperimental physics. While most scientists in all the fields Isurveyed reported collaborating with persons of faculty rank,graduate students were the most important components inevery field. The importance of young persons employed atsubstandard wages is common in almost all types of culture.

Consistent with the preceding discussion, I found that workgroups are smallest in the least routine field, mathematics.Other differences among fields are also affected by technology.Chemistry epitomizes a labor-intensive field. Professors ofchemistry can work quite autonomously with workers andapparatus under their exclusive control. The relative independ-ence of work groups in chemistry permits a lower degree ofcollaboration among faculty peers: 44% of the chemistsreported collaborating with peers, in contrast to 68% of theexperimental physicists, and 61 % of the biologists. When suchcollaboration exists in chemistry, it is likely to be informal andto involve a clear division of labor-chemists collaborate whenthey are compelled to do so because they lack the needed skills.

As in almost any labor-intensive area of work, attempts aremade to mechanize and automate work by substituting capitalfor labor. In chemistry this has tended to occur withoutproducing major changes in the organization of work groups.Thus, most of the chemists in my survey were using analyticinstruments invented in the preceding 15 years and sometimesbased on discoveries made in that inverval; but the apparatus issuch that it usually can be operated by the chemist himself andhis small number of relatively untrained assistants.

In other areas of science, the nature of research problems aswell as a desire to substitute capital for labor have led to thosemajor changes in the nature of work we call big science. There isno clear-cut demarcation between little science and big science,and not all big science is capital-intensive. Big science is


Highlight

Highlight


epitomized by high energy physics. Despite labor-saving devices,the work groups at high energy particle laboratories can be verylarge. For example, in 1964 the research staff at the LawrenceRadiation Laboratory in Berkeley was divided into about 11 Igroups; outside groups also used the Laboratorys accelerators.The smallest group had 12 professional personnel. The largestincluded 160 full-time-equivalent personnel-23 Ph.D.s, 20graduate students, and numerous others, including cryogenicengineers, mathematical programmers, computer technicians,and still others for more routine tasks (Swatez, 1970). As onemight expect, such large batch production in science involves agreater division of labor-but primarily among subordinatepersonel, not the professional physicists who lead the groups.There is a much higher proportion of highly trained auxiliarypersonnel, a broader span of control, and greater centralizationof decision-making. However, the &dquo;centralization&dquo; of workgroups in such large batch production tends to be like that ofinnovative firms in industries with uncertain environments (cf.Ritti and Goldner, 1969: B-240 ff.). The top decision-makersdo not so much initiate actions for subordinate scientists asmake choices among alternative courses of actions suggested bysubordinates. Subordinate Ph.D. scientists, and even graduatestudents, suggest experiments and are expected to do so.

In most cases big science has not meant an abandonment ofsmall batch production in science. Scarce and expensiveresearch facilities, such as computer centers or survey researchlaboratories, often can be organized in such a way as to providesmall work groups with routinized services on demand.

THE COORDINATION OF ACTIVITIES AMONG WORK GROUPS

The work groups I have been discussing are ordinarilyembedded in larger organizations. Universities and governmentalor industrial establishments can be seen as providing supportservices and as sanctioning or specifying goals; I will be unableto give such organizations the attention they deserve here. Inaddition to such formal organizations, each work group is likelyto be linked to other groups working on the same or related



research problems. Such links are likely to be informal, and thegroups are usually found in different formal organizations. Theplural number of groups working in a problem area probablystems mostly from the nature of the tasks confronting them: noone group is likely to be perceived capable of solving all thoseproblems felt to require solutions. It is also motivational: linkedgroups provided the most important audiences for the work ofeach.

These sets of groups are difficult to label or conceptualize.Considering their structure, one is led to call them networks or,more precisely, clusters in networks. Considering the intellec-tual content of their work, or their positions in encompassingdisciplinary organizations, one is led to call them specialties orsubspecialties. Considering the history of such groups, the mostfelicitous label might be &dquo;invisible colleges&dquo; (Price, 1963: ch.3). Different network clusters vary considerably in size,interconnectedness, internal stratification, clarity of boundaries,and visibility to members and non-members. Yet they areubiquitous and, perhaps, the most important level of the socialorganization of science.

These network clusters can be viewed as production organiza-tions. From the point of view of an organizational adminis-trator, they might seem to be hopelessly irrational &dquo;organiza-tions.&dquo; There is seldom any effective control over networkmembership; groups may enter or leave easily. There is almostnever coordinated planning or control; no agencies exist todirect component groups to work on particular problems. As aresult, there is considerable duplication of effort, and mostscientists are concerned that others may publish solutions totheir problems before they can succeed in doing so (Hagstrom,1974). Yet, given the uncertainty of the task environment andthe relative beneficence of the support and market environ-ments, this loose form of organization may lead to moreeffective coordination of the efforts of different groups thanany alternative form. Given task uncertainty, effective coordina-tion cannot be achieved through planning or centralizeddirection. The alternative is more or less rapid feedback throughformal and informal communication among the component


[759]Hagstrom / PRODUCTION IN SCIENCE i

groups in network clusters. This type of coordination requiresthat scientists spend a far larger fraction of their time incommunicative behavior than most other types of workers.

It is difficult to estimate the typical size of clusters in thesenetworks, since groups often enter or leave, and some groupsare only loosely attached. It is safe to say that they are typicallysmall. Derek Prices 1963 guess that the maximum size of aninvisible college is 100 cannot be far wrong. On the basis ofnumerous case studies, I would estimate that the averagenumber of individual professional workers closely linked to-gether in networks is less than 50; the effective networkcomponent is ordinarily a work group, and thus the averagenumber of closely linked groups in networks must be less than50. (For case studies and estimates, see Hagstrom, 1967 and1974; Crane, 1972; and Griffith and Miller, 1970.) Theautonomy of work groups, and the linking together by looseties of work groups throughout a discipline (cf. Granovetter,1973), make it possible for the effective size of network clustersto expand rapidly, a growth potential demonstrated by numer-ous studies, in particular Diana Cranes Invisible Colleges(1972).Members of scientific networks share with one another those

related commitments Thomas Kuhn has called &dquo;paradigms&dquo;-or,in words more true to Kuhn, the network cluster is one aspectof that &dquo;disciplinary matrix&dquo; constituting a paradigm (Kuhn,1970). Members share commitments to values, metaphysicalparadigms, heuristic models, and examplars. Thus, they cancommunicate with one another without first having to negotiatea consensus about fundamental values and meanings, and theycan coordinate their activity to solve scientific problemsbecause of shared convictions that the problems are paradigm-induced &dquo;puzzles&dquo; with assured solutions.

In the language of the organization theory of March andSimon (1958), one can view the network cluster as anorganization that both constrains and facilitates the work of itsmembers. While such network clusters are very much differentform the well-bounded and hierarchical firms about whichSimon and his colleagues wrote, it is clear that they operate to



socialize individual members, to reward conformity, and topunish deviance. Just as paradigms function to enable scientiststo solve puzzles in the absence of explicit rules and the presenceof anomalous data, so the organization, for March and Simon,facilitates local rationality and &dquo;satisficing&dquo; behavior in situa-tions where optimizing behavior is impossible. The organizationdoes this by specifying goals, fostering limited search processesthat are only mildly innovative, focusing attention on arestricted range of stimuli, providing programs and repertoiresof action that do not require a process of optimal decision-making at every turn, and in other ways not so relevant toscience. There is, thus, a close similarity in the oppositionbetween the March-Simon model of organizational rationalityand classical optimizing models, on the one hand, and theopposition between Kuhns notion of normal science and KarlPoppers ideal of permanently revolutionary science (seeLakatos and Musgrave, 1970), on the other. That is, debatesabout rationality in science are like debates about rationality inorganizations-perhaps because of fundamental similarities inthe nature of human rationality.

Kuhnian normal science is inadequate to characterize theorganization of science in at least two situations. The first isthat of crisis and revolution, which he discussed in detail. Whena paradigm cannot successfully resolve critical anomalies, arevolutionary situation may arise. At such times fundamentalvalues and beliefs can no longer be taken for granted, and atsuch times communication may break down. As Kuhn puts it,in revolutionary situations those committed to competingparadigms may seem to speak in different languages. It ispossible that productivity may decline in such critical periods,although in the one case I know of, such a situation coincidedwith increased productivity in the parity nonconservation crisisof weak interaction physics around 1956 (Sullivan et al., 1975).A second situation deviating from normal science may occur

when a preponderance of scientists are not linked to others innetwork clusters. This may be a common characteristic ofprcparadigmatic and multiparadigmatic fields, and it may resultfrom pathological overspecialization as in mathematics. Follow-



ing Durkheim, I have earlier suggested that such situations canbe characterized as anomic and will have the characteristicsexpected of anomic societies (Hagstrom, 1964, 1965; see alsoHargens, 1975). Workers have little social confirmation of thevalue of their own products and thus vacillate from attributingabsolute value to them-&dquo;mathematics for mathematics sake&dquo;-to having pervasive doubts about their value and, since scientistsidentify themselves with their work (cf. Stinchcombe, 1966),pervasive doubts about their own self-worth. They may becomeritualistic, continuing to produce when they have no confidencethat their work will be appreciated or used by others. It isreasonable to believe that productivity will decline in suchanomic situations.

While anomie is probably less common in science than in artand music, those sociologists of art and music among myreaders probably know from first-hand experience what it is liketo work in an anomic field. Anomie implies a lack of paradigmsin the Kuhnian sense: one does not find, simultaneously, anintegrated group, shared values, models, and exemplary works.Each worker in the sociology of art has tended, in Kuhnswords, to create the field anew. Few have found it possible toselect, as research problems, puzzles existing in prior exemplars,nor can they confidently use the methods of the exemplars tosolve their puzzles. Each worker has had to try to create anewhis own &dquo;invisible college,&dquo; and many have been tempted tooptimize by solving &dquo;fundamental&dquo; problems rather than&dquo;satisfice&dquo; by working on smaller soluble problems. It is noteasy. (On the other hand, work in nonparadigmatic fields is notnecessarily inferior, as evidenced by the work of such scientistsas Louis Pasteur, Gregor Mendel, and Sigmund Freud in suchfields.)DISTRIBUTION AND CONSUMPTION

Before the professionalization and institutionalization ofscience in the nineteenth century, scientists often disseminatedtheir discoveries directly to an elite clientele in settings thatrequired entertainment as well as edification. Michael Faradaylectured gentlemen and ladies at the Royal Institution ofLondon, and he wrote that polite audiences &dquo;expect to be


Highlight

Highlight

Highlight

Highlight

Highlight

KateNote culture


entertained not only by the subject of the lecture, but by themanner of the lecturer; they look for respect, for languageconsonant to their dignity, and ideas on a level with their own.&dquo;(Crowther, 1935: 81). Similar statements could have been madeby Robert Hooke about his work for the Royal Society around1700, when he was expected to prepare an interesting experi-ment about every other week, and by the workers in similarinstitutions in Europe in the eighteenth century. Hooke,Faraday, and others managed to be quite productive in suchsettings, in part because they were linked into the kind ofinformal networks of scientists that have been described above.

In modem science products are disseminated through chan-nels controlled almost completely by the scientific producersthemselves. Perhaps most important are the courses of schools,colleges, and universities, and their associated textbooks andlaboratories. In these settings the scientist is clearly dominantover his clientele; he repeatedly evaluates their performance,and his evaluations of them are far more consequential thantheir evaluations of him.

Scientists also control, in large part, the media of printedcommunication, of which the journals are the most importantsegment. Among other things, this entails that they are almostassured that what they write will be published. Althoughaverage journal rejection rates range from 24% in physics to84% in political science (Zuckerman and Merton, 1971: 76),scientists can submit articles repeatedly, and initial rejectionrarely prevents the eventual publication of a manuscript(Garvey and Griffith, 1971: 358). In addition, many articleshave extremely few readers. For example, it has been estimatedthat about half of the research reports in the core journals ofpsychology are likely to be read by less than I percent (or twohundred) psychologists (Garvey and Griffity, 1971: 358). Evenif read, papers may never be referred to; a study of the paperspublished in the major journal of a subfield of physics showedthat 10% of those published in 1967-1968 were never referredto in the succeeding seven years (Gillmore, 1975)! Clearly,&dquo;marketability&dquo; is not the primary criterion in the decision topublish.


Highlight

Highlight

Highlight

Highlight

Highlight

Highlight

Highlight

KateHighlight

KateHighlight

KateHighlight

KateHighlight

KateHighlight


Scientists spend a large fraction of their time evaluating thework of others. Most academic scientists serve as journalreferees at one time or another. About 12% of the scientists inmy 1966 survey had been editors or associate editors ofjournals, and more than one-quarter had served on advisorycommittees for governmental or other research fundingagencies. In addition, academic scientists spend much timeevaluating the work of students and candidates forappointments and promotions. This relatively great attentiondevoted to evaluating the work of others may be typical ofculture producers in all fields. For example, Batia Sharon(1969) surveyed an inclusive sample of Chicago artists andfound that a majority had served as jurors for art shows, morethan one-third of them doing so at least four times in theircareers.

The common small readership of articles in scientific journalssuggests that they often serve primarily as communicationchannels for the network clusters I have described above. Themeetings of scientific societies also have this as a majorfunction. Network clusters also involve frequent, less formalcommunication. For example, Hargens (1975: 29) found thatacademic chemists report spending an average of more than fourhours per week on professional correspondence; and 62%attended professional meetings, of one sort or another, for atleast six days a year. In tightly knit clusters, informalcommunication is much more frequent and intensive.

If journals often serve the interests of network clusters, theyare also accessible to anyone else with the qualifications to readthem. Thus, the literature in all the fields of basic science islinked by references from papers in each field to papers in oftenquite dissimilar other fields. The literature in basic science isalso potentially accessible to applied scientists and engineers.However, studies of the communication behavior of engineersshow that they seldom use the literature of the basic sciences(Marquis and Allen, 1966; Allen, 1970). Often that literature iswritten in a language they cannot understand, and often theycannot search the basic literature to find what they need toknow. Thus, references in patents are usually to other patents,


Highlight -


and references in applied science and engineering journals areusually to the same types of journals. Unlike basic scientists,engineers tend to be oriented less to external communicationthan to the unpublished technical reports of their employingorganizations and to interpersonal contacts within theseorganizations. Given this typical internal orientation, it is notsurprising that Allen (1970) found that a small number ofgatekeepers played a critical role in transferring scientific andtechnical information across the organizational boundaries todevelopment groups within. These gatekeepers tended to bescientists and engineers who were well-educated, more widelyread, and maintained closer personal ties with persons outsidethe firm. In addition to such gatekeepers, personal consultationwith academic scientists plays an important role in transferringinformation from basic science to technology; directconsultation reduces the costs of search and permits thenegotiation of mutually comprehensible languages. Despitecriticisms of the speed and extent to which basic research isbrought to bear on societal needs, and despite considerableinterindustry variability, the loose links between science andtechnology seem to have been remarkably effective in recentdecades (Nelson and Pollock, 1970).

The preceding paragraphs have stressed the dissemination ofscientific knowledge among professionals and have neglectedthe dissemination of such knowledge in the larger lay com-munity. This reflects the recent research interests of sociologistsof science, who have paid far less attention to folk culture ormass culture than have sociologists of art, music, or religion.However, research by anthropologists of ethnoscience suggeststhat modem science (among other things) has had deep andextensive effects on the belief systems of persons in developedsocieties. Comparative studies of developed and developingsocieties show that children and adults in the former are muchmore able than those in the latter to cope with abstractproblems, present coherent explanations of a range of naturalphenomena, and recognize the possibility that knowledge maybe obtained from empirical research instead of from traditionalauthorities alone (see, e.g., Gay and Cole, 1967; Dart andPradhan, 1967).



The same studies show that persons in developing nationsoften have a kind of dualistic belief system, asserting the&dquo;validity&dquo; of both Western school-based knowledge and ap-parently inconsistent traditional beliefs. For example, Nepaleseschool children readily assert the truth of both the propositionsthat &dquo;Lightning comes from the collision of clouds,&dquo; and&dquo;Lightning comes from the bangles of Indras dancers&dquo; (Dartand Pradhan, 1967: 651). Persons in developed societies aremuch less likely to have such a duality of viewpoint. This doesnot imply that the scientific culture of the professionals isaccepted without question by persons in developed societies.Laymen can recognize that professionals may have differentgoals and interests than themselves, and they may discount theteaching and advice of professionals after taking such differ-ences into account. The process is nicely illustrated in Beckers(1974) comparison of how the effects of drugs are assessedwhen the situation is controlled by users, by users agents (e.g.,physicians and pharmaceutical companies), and by externalagents (e.g., professional staff in mental hospitals); the main andside effects of drugs may be assessed quite differently in thethree situations. While Becker stresses the different rationalitiesof laymen and professionals, other types of lay rejection oforthodox science may require closer study of the expressivesignificance of deviant beliefs. The popular science distributedin mass media, such as newspaper Sunday supplements, theNational Enquirer, or womens magazines, often emphasizessuch unorthodox approaches as parapsychology, astrology, orvariants of flying saucer stories. We do not know howimportant such beliefs are in modem societies; explanation oftheir significance may be as much the task of the sociologist ofreligion as the sociologist of science.

In any case, scientists engaged in teaching, or whose workdepends on the actions of political leaders, can be quite awareof the reluctance of laymen to admit that the knowledge theyseek to offer is either relevant or valid. Lay skepticism may be afunction of such factors as the cost, in money or consumereffort, to acquire the knowledge; the degree to which orthodoxscience deviates from common sense; the relation of scientists



to dominant institutions; and the degree of consensus amongculture producers. It is reasonable to believe that lay skepticismwill erode the authority of scientists and societal support forscientific research.

TRANSFORMATIONS OF SCIENTIFIC CULTURE

Unlike many commodities, culture can be consumed withoutdestroying it: culture is not like cake. In fact, the &dquo;consump-tion&dquo; of culture in the classroom, theater, or elsewhere, canmean its &dquo;reproduction.&dquo; Nevertheless, scientific products are,if not destroyed, revolutionized, surpassed, and abandoned. Themetaphor of the production of culture can help understandsuch processes to some extent. For example, just as rapidchange leads to the obsolescence of commodities, it can lead tothe obsolescence of knowledge and those scientists and tech-nical workers possessing that knowledge. Such obsolescence hasbeen perceived as problematic by social science students orengineers and employers of engineers. Engineers are most likelyto become obsolete if they have worked in a relatively routinearea that becomes rapidly changed by new discoveries; they areless likely to become obsolete if they have advanced degrees andwork in areas of advanced and nonroutine technology (Perrucciand Rothman, 1969).A related kind of transformation in science is the &dquo;exhaus-

tion&dquo; of research opportunities in a specialty (cf. Crane, 1972).This may occur when workers in a field find difficulty indiscovering interesting problems for research that can beapproached with existing techniques. The obligation to doresearch that is both original and interesting tends to forceworkers to change specialties; preparing for research in a newspecialty may be burdensome and risky, carrying no assuranceof success, and some scientists abandon active research ratherthan take such risks. More generally, the difficulty of findingintrinsically important research problems may lead to &dquo;fash-ions&dquo; in science-scientists may select problems because othersdo, rather than because of any personal conviction in theimportance of the problem (Hagstrom, 1965: 177-184). Crane( 1969) has argued that such behavior, considered deviant byscientists, is actually uncommon.


Highlight

Highlight


More radical transformations in scientific culture are not soreadily interpreted with the production-of-culture metaphor.Since the appearance of Thomas Kuhns The Structure ofScientific Revolutions in 1962, thought about such transforma-tions has been more heavily influenced by political metaphors.Such a contrast between production and political metaphorsmay do justice to neither, however. Many others have recog-nized that not only the production of culture, but theproduction of other goods, requires consideration of conceptsand theories about political behavior and collective behavior.

REFERENCES

ALLEN, T. J. (1970) "Roles in technical communication networks," pp. 191-208 inC. E. Nelson and D. K. Pollock (eds.) Communication Among Scientists andEngineers. Lexington, Mass.: D. C. Heath.

BECKER, H. S. (1974) "Consciousness, power and drug effects." J. of PsychedelicDrugs 6 (January-March): 67-76.

COLLINS, R. (1975) Conflict Sociology: Toward an Explanatory Science. NewYork: Academic.

CRANE, D. (1972) Invisible Colleges: Diffusion of Knowledge in ScientificCommunities. Chicago: Univ. of Chicago Press.

(1969) "Fashion in science: does it exist?" Social Problems 16 (Spring):433-441.

CROWTHER, J. G. (1935) British Scientists of the Nineteenth Century. London:Routledge & Kegan Paul.

DART, F. E. and P. L. PRADHAN (1967) "Cross-cultural teaching of science."Science 155 (10 February): 649-656.

FISHER, C. S. (1973) "Some social characteristics of mathematicians and theirwork." Amer. J. of Sociology 78: 1094-1118.

GARVEY, W. D. and B. C. GRIFFITH (1971) "Scientific communication: its role inthe conduct of research and creation of knowledge." Amer. Psychologist 26(April): 349-362.

GAY, J. and M. COLE (1967) The New Mathematics and an Old Culture: A Study ofLearning Among the Kpelle of Liberia. New York: Holt, Rinehart & Winston.

GILLMORE, C. S. (1975) "Citation characteristics of the JATP literature." J. ofAtmospheric & Terrestrial Physics (in press).

GRANOVETTER, M. S. (1973) "The strength of weak ties." Amer. J. of Sociology78 (May): 1360-1380.

GRIFFITH, B. and A. J. MILLER (1970) "Networks of informal communicationamong scientifically productive scientists," pp. 125-140 in C. E. Nelson and D. K.Pollock (eds.) Communication Among Scientists and Engineers. Lexington, Mass.:D. C. Heath.

HAGSTROM, W. O. (1974) "Competition in science." Amer. Soc. Rev. 39(February): 1-18.

(1967) "Competition and teamwork in science." Final report to the National


Highlight


Science Foundation for research grant GS-657, University of Wisconsin, Madison.(mimeo).

(1965) The Scientific Community. New York: Basic Books. (1964) "Anomy in scientific organizations." Social Problems 12 (Fall):

186-195.HARGENS, L. L. (1975) Patterns of Scientific Research. Washington, D.C.:

American Sociological Association.KAPLAN, A. (1964) The Conduct of Inquiry. San Francisco: Chandler.KUHN, T. S. (1970) The Structure of Scientific Revolutions, second ed. Chicago:

Univ. of Chicago Press.LAKATOS, I. and A. MUSGRAVE (1970) "Criticism and the growth of knowledge."

Proceedings of the International Colloquium in the Philosophy of Science,London, 1965, Vol. 4. Cambridge: Cambridge Univ. Press.

MARCH, J. G. and H. SIMON (1958) Organizations. New York: John Wiley.MARQUIS, D. G. and T. J. ALLEN (1966) "Communication patterns in applied

technology." Amer. Psychologist 21: 1052-1060.MAYHEW, G. P. (1967) "Recent Swift scholarship," pp. 187-197 in R. McHugh and

P. Edwards (eds.) Johnathan Swift 1667-1967. Dublin: Dolman.NELSON, C. E. and D. K. POLLOCK (1970) Communication Among Scientists and

Engineers. Lexington, Mass. D. C. Heath.PERROW, C. (1972) Complex Organizations: A Critical Essay. Glenview, Ill.: Scott,

Foresman. (1967) "A framework for the comparative analysis of organizations." Amer.

Soc. Rev. 32 (April): 194-208.PERRUCCI, R. and R. A. ROTHMAN (1969) "Obsolescence of knowledge and the

professional career," pp. 247-276 in R. Perrucci and J. E. Gersel (eds.) TheEngineers and the Social System. New York: John Wiley.

Philological Quarterly (1974) Vol. 53 (Fall): 802-807.PRICE, D. J. de S (1963) Little Science, Big Science. New York: Columbia Univ.

Press.

RITTI, R. R. and F. H. GOLDNER (1969) "Professional pluralism in an industrialorganization." Management Sci. 16 (December): B-233-B-246.

SHARON, B. (1969) "Center and penphery in Chicagos art community." Mastersthesis, University of Wisconsin, Department of Sociology.

STINCHCOMBE, A. L. (1966) "On getting hung-up and other assorted illnesses."Johns Hopkins Mag. (Winter): 25-30.

SULLIVAN, D., D. H. WHITE and E. BARBONI (1975) "The state of a science:indicators in the specialty of weak interactions." Delivered at the annual meetingof the American Sociological Association, San Francisco.

SWATEZ, G. M. (1970) "The social organization of a university laboratory." Minerva8 (January) 36-58.

WOODWARD, J ( 1965) Industrial Organization: Theory and Practice. London:Oxford Univ Press

ZUCKI RMAN, H and R K. MERTON (1971) "Patterns of evaluation in science:institutionalization, structure and functions ot the referee system." Minerva 9.66-100


American Behavioral Scientist 1976 Hagstrom 753 68

Documents

production perspective

diverse realmsof culture

engineering work groups

types of culture kaplan

organization of work

research groups

structure of science

st petersburg