(Routledge Library Editions - Social Theory 60) Michael Mulkay - Science and The Sociology of Knowledge-Routledge (2014)

ROUTLEDGE LIBRARY EDITIONS:
SOCIAL THEORY
Volume 60
SCIENCE AND THE SOCIOLOGY

OF KNOWLEDGE
This page intentionally left blank
SCIENCE AND THE SOCIOLOGY
OF KNOWLEDGE
MICHAEL MULKAY
Routledge
Taylor & Francis Group
LONDON AND NEW YORK
First published in 1979
This edition first published in 2015
by Routledge
2 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN
and by Routledge
711 Third Avenue, New York, NY 10017
Routledge is an imprint of the Taylor & Francis Group, an informa business
© 1979 George Allen & Unwin (Publishers) Ltd
All rights reserved. No part of this book may be reprinted or reproduced or
utilised in any form or by any electronic, mechanical, or other means, now
known or hereafter invented, including photocopying and recording, or in any
information storage or retrieval system, without permission in writing from the
publishers.
Trademark notice: Product or corporate names may be trademarks or registered

trademarks, and are used only for identification and explanation without intent
to infringe.
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library
ISBN: 978-0-415-72731-0 (Set)

eISBN: 978-1-315-76997-4 (Set)
ISBN: 978-1-138-78247-1 (Volume 60)
eISBN: 978-1-315-76340-8 (Volume 60)
Publisher’s Note
The publisher has gone to great lengths to ensure the quality of this reprint but
points out that some imperfections in the original copies may be apparent.
Disclaimer
The publisher has made every effort to trace copyright holders and would
welcome correspondence from those they have been unable to trace.
Science
andthe
Sociology
of
Knowledge
MICHAEL MULKAY
London
GEORGEALLEN & UNWIN
Boston Sidney
First publishedin 1979
Secondimpression1980
This book is copyright under the BerneConvention.All rights

are reserved.Apart from any fair dealing for the purposeof
privatestudy, research,criticism or review, as permittedunder
the Copyright Act, 1956, no part of this publication may be
reproduced,storedin a retrieval system,or transmitted,in any
form or by any means, electronic, electrical, chemical,
mechanical, optical, photocopying, recording or otherwise,
without the prior permissionof the copyright owner. Enquiries
shouldbe sentto the publishersat the undermentionedaddress:
GEORGE ALLEN & UNWIN LTD

40 Museum Street, London WClA lLU
© GeorgeAllen & Unwin (Publishers)Ltd, 1979
British Library Cataloguingin PublicationData
Mulkay, Michael Joseph

Scienceand the sociology of knowledge
(Controversiesin sociology).
1. Science-Socialaspects
I. Title I I. Series
301.24'3 Q175.5 78-40852
ISBN 0-04-301093-8
ISBN 0-04-301094-6 Pbk
Typesetin 10 on 11 point Times by Red Lion Setters,London

and printed in Great Britain
by Biddies Ltd, Guildford, Surrey
Contents
Acknowledgements pageviii
1 The CustomarySociologicalView of Science 1
The classicview of science:Durkheim and Marx;
more recentvariants: Mannheimand Stark;
the standardview of scientific knowledge;
the sociology of science
2 Revisionsof the StandardView 27
The uniformity of nature; fact and theory;
observationin science;the assessment of
knowledge-claims;sociologicalimplications
3 Cultural Interpretationin Science 63
The social rhetoric of science;the dynamics
of knowledge-production;the interpretation
of cultural resources
4 Scienceandthe Wider Society 96
Scientists'use of 'external' cultural
resources;the use of scientific culture in
externalpolitical settings;brief summary
and concludingremarks
References 123
Index 130
Acknowledgements
I wish to thank Tom Bottomoreand John Law for readingthe first

draft of this book very carefully and for making a number of
helpful commentsand suggestions.I would also like to thank
Elizabeth Chaplin for her help in preparingthe final manuscript
and Michael Holdsworth for giving me a special dispensationto
exceedthe usuallength for contributionsto this series.
Possiblythe world of externalfacts is much more fertile and plastic
than we haveventuredto suppose;it may be that all thesecosmo-
logies and many moreanalysesand classificationsare genuineways
of arrangingwhat natureoffers to our understanding,and that the
main condition determining our selectionbetweenthem is some-
thing in us ratherthansomethingin the externalworld.
E. A. Burtt, TheMetaphysicalFoundationsofModern Science,
p.305
If we cannotbearthe paradoxof acceptingthat genuineknowledge

may be fallible, then we must ban the term altogether from
productionsof the humanintellect.
J. R. Ravetz,ScientificKnowledgeandIts SocialProblems,p. 236
This page intentionally left blank
1
The Customary Sociological

View of Science
The sociologyof knowledgehasa long history. Its origins are sometimes

traced back as far as the writings of Francis Bacon, and it certainly
appearsas an importantelementin the work of the 'founding fathers' of
sociology, such as Marx, Pareto and Durkheim. It continuesto be a
thriving area of investigation today, although its practitioners have
tendedto concentrateincreasinglyon detailedstudiesof specific bodies
of knowledge and belief, instead of presentingthe kind of general
analytical formulae favoured by their forebears. Despite this long
tradition, little agreementhas beenachieved.The field is still characteri-
sedby a greatdiversity of aims andinterpretativeschemes.This is true to
suchan extent that someauthorsare unwilling to offer any definition at
all of the sociologyof knowledge(Curtis and Petras,1970, p. 7), whilst
those who risk making the attempt are forced to devise very general
formulationsin orderto coverthe entirecorpusof relevantliteratureand
encompassthe full range of phenomenato be studied. Thus Merton
(1957, p. 456), having noted that the term 'knowledge' must be inter-
pretedvery broadlyindeedin this contextas covering 'virtually the entire
gamut of cultural products',writes that the sociology of knowledge 'is
primarily concernedwith the relations between knowledge and other
existential factors in the society or culture. Generaland even vague as
this formulation of the centralpurposemay be, a morespecific statement
will not serve to include the diverse approacheswhich have been
developed.'Within the wide rangeof issuescoveredby such a definition,
one reasonablyclear distinction can be made, namely, that between
popular belief and commonsenseor everyday knowledge, on the one
hand, and systematised,specialised knowledge, on the other hand
(Berger and Luckmann, 1967). In this book, I shall be concerned
exclusively with the latter, that is, with the sociological analysis of
specialisedknowledge.
One of the central concernsof sociologistsof knowledgehas beento
show how specialisedbodiesof thought and knowledge,such as aesthe-
tic, moral and philosophical systems, religious creeds and political
principles, are influenced by the social and cultural contexts in which
they are produced.The guiding assumptionbehind this kind of analysis
2 Scienceand the SociologyofKnowledge
is clearly expressedby Mannheim. It is that the sociology of knowledge
explores the 'varying ways in which objects presentthemselvesto the
subject according to differences in social settings... when and where
social structurescome to expressthemselvesin the structure of asser-
tions, and in what sensethe former concretely determine the latter'
(1936, p. 265). This generalstatementleadsimmediatelyto a number of
more specific questions.We are led to ask, for instance:what kinds of
social and cultural factors exercisean influence on mental productions
and in what degree?What kinds of connectionsare there betweensocial
and cultural influencesand mental productions?Which aspectsof these
mental productions are we trying to account for-their form, their
content,their incidence,their generationor their acceptance?And most
important of all, for the purposes of this chapter, which mental
productionsare open to this kind of sociological analysis?Are we to
includeall cultural productsor only certainclassesof them?
When we examine which areas of knowledge have actually been
subjectto empirical investigation,we find that scientific and mathemati-
cal thought has beenalmostcompletelyignored by sociologists.I do not
meanby this that therehave beenno sociologicalstudiesof scientistsor
of the scientific community. What has beenabsent,until very recently,
has been the empirical investigation from a sociological perspectiveof
scientific knowledgeand its social construction. In addition, although
most sociologistsof knowledgehave discussedsciencein generalterms,
they have repeatedlyrejectedin principle the possibility that the form or
content of scientific knowledge,as distinct from its incidenceor recep-
tion, might be in someway socially contingent.Instead,they haveargued
strongly, albeit with occasional uncertainty, that the substanceof
scientific knowledge is independentof social influence and they have
tried to justify this assertion on philosophical grounds. They have
claimed,in short, that scienceis a specialsociologicalcasebecauseit has
a specialepistemologicalstatus.Becausethis line of reasoninghas been
generallyaccepted,sociologistshave left the close analysis of scientific
knowledgeto the philosophersof scienceand to the historiansof ideas.
After many yearsin which it seemedto have beenconclusivelysettled
that scientific knowledgewas exempt from sociologicalanalysis,debate
hasrecentlybegunagain.This hasbeen,in part, a consequence of radical
changesin the views of science held by historians and philosophers.
Throughoutthe 1960s,a number of historians and philosophersfound
themselveseither borderingon or actively engagedin sociologicalinter-
pretationof scienceasthey respondedto the traditionalproblemsof their
own disciplines. Gradually, these new ideas have entered sociology,
helping to underminethe epistemologicalassumptionswhich had vir-
tually requiredthe sociology of knowledgeto treat scienceas a special
case.As the restrictionsimposedby the old epistemologyhave become
weaker,so sociologistshavesoughtto extendand modify the work of the
The CustomarySociologicalView oj Science 3
philosophersand historians in order to produce, for the first time, a
genuine sociology of scientific knowledge. In later chapters I will
describesomeof the recentchangesin the philosophyand historiography
of scienceas well as certain parallelchangesin sociologicalanalysis.For
the remainderof this chapterI will examineseveralmajor contributions
to the sociology of knowledgeand the sociology of science,in order to
show that, despite occasional speculation and dissent, science has
customarily been regardedas a special sociological case, and also to
clarify the philosophicalrationale which underlay this position.
THE CLASSIC VIEW OF SCIENCE: DURKHEIM AND MARX

All the major contributors in the nineteenth century to the incipient
sociologyof knowledgewere doubtful about the possibility of including
natural sciencewithin its scope. Let me illustrate this with respect to
Emile Durkheim and Karl Marx. It is true that Durkheim attemptedto
providea sociologicalaccountof the genesisof man'sbasiccategoriesof
thought and his forms of reasoning.He argued,for instance,that ideas
of time and space, force and contradiction vary from one group to
another and within the same group from one period to another. For
Durkheim, the existenceof such cultural variation showedthat our basic
categoriesand our rules of logic dependto someextent on factors that
are historical and consequentlysocial (1915, pp. 12-13). This appearsto
make an analysisof the cognitive content of sciencedistinctly possible,
becauseit seems that the conclusions of every particular intellectual
community will be constrainedat least partly by such factors as their
cultural resources,the structureof their social group and their place in
the wider society.But althoughDurkheim doesnot abandonthis general
position, he does modify some of its details so as to remove scientific
knowledgefrom analytical consideration.
In the first place, he takes steps to avoid a completely relativist
position, in which the social origin of categoriesand forms of reasoning
could be seento renderthem wholly arbitrary so far as their application
to nature is concerned.
From the fact that the ideasof time, space,class, causeor personality
are constructedout of social elements,it is not necessaryto conclude
that they aredevoid of all objectivevalue. On the contrary, their social
origin ratherleadsto the belief that they are not without foundation in
the natureof things. (1915, p. 19)
Durkheim reachesthis conclusion by postulating a unity between the

physical and social worlds. Thus a group's conception of time, he
suggests,will be derivedfrom the social rhythms of its collective life. But
these social rhythms will be linked to, and in a sensepart of, more
inclusive periodicitiesin the physical world. The physical and the social
worlds constituteone overall realm of natural phenomena.Accordingly,
in Durkheim's view, it follows that conceptionsarising out of social
regularities will be applicable to parallel regularities in the physical
world. However, Durkheim's argument here is extremely general. It
appliesequally to all humangroups.It establishes,at most, only that all
socially derivedcategorieswill havesome,unspecified,degreeof 'object-
ivity'. The problemremainsof how to judge betweendivergentaccounts
of the physicalworld offered by the membersof different social groups.
Durkheim does not appearto recognisethis difficulty explicitly. But it
becomesclear in his account of social evolution that he can and does
employ a more discriminating criterion of objectivity. Objectivity, he
argues,becomesincreasinglyattainableas social evolution unfolds and
as sciencereplacesreligion as the basis for human thought about the
natural world .
. . .the essentialideas of scientific logic are of religious i.e. social

origin ... [but] sciencegives them a new elaboration;it purgesthem of
all accidentalelements;in a generalway, it brings a spirit of criticism
into all its doings, which religion ignores; it surroundsitself with pre-
cautions to 'escapeprecipitation and bias', and to hold aside the
passions, prejudices and all subjective influences... Having left
religion, sciencetendsto substituteitself for this latter in all that which
concernsthe cognitiveand intellectualfunctions. (1915, p. 429)
Durkheim describesat a rather general level the social conditions

which he believes are responsible for this fundamental transition in
humanthought. His central claim is that the growth in size of human
societiesand their progressiveinternal differentiation increasinglyliber-
ate intellectual activity from social constraint. Scientific thought is an
outcomeof this liberation and its conclusionsare, therefore, compara-
tively unaffectedby direct social influences.Religious thought about the
natural world, originating in cohesive, small-scalesocieties, was per-
meatedby categoriesand presuppositionsderivedfrom social life. But as
societiesbecomemorecomplexandthe form of social solidarity becomes
less restricting, so it is increasinglypossiblefor certainsectorsof society
to refine their conceptions and their techniques of observation in
accordancewith the actual realities of the natural world. The concepts
and conclusions of science, he maintains, are increasingly adopted
becausethey are true and not simply, as is primarily the case with
religious beliefs, becausethey arecollective(1915, p. 437).
A sociologicalanalysisof scienceis possible,then, for Durkheim; but
in a more limited form than is the casefor other areas of intellectual
endeavour.In principle, we can show how certain social developments
havebroughtaboutthe emergenceof science;we can investigatewhether
The CustomarySociological View oj science 5
the scientific community has certain distinctive features which make
possiblethe institutionalisationof the scientific method and the virtual
elimination of bias, prejudice and intellectual distortion; and we can
observehow the minority views of the scientific specialistare receivedby
the other sectorsof highly differentiatedsocieties.All this we can do, and
perhapsmore. But we cannot give a sociological account of scientific
knowledgebecause,to the extent that it is truly scientific, it is indepen-
dent of its social context. Genuinesciences,such as astronomy,physics
and biology, are basedon observablefacts about the physical world. The
conclusionsof thesesciencesare derived from the facts, insteadof being
imposed upon them. Science representsphenomenanot in terms of
culturally contingent ideas, 'but in terms of their inherent properties'
(1938, p. 35).
Durkheim'sanalysisof the social origins of knowledgeand belief was
undertakenas an explicit exercisein the sociology of knowledge. As a
result, his verdict with respectto scienceis relatively unambiguous.In
contrast,Marx's view of scienceas a social phenomenonemergesin a
piecemeal fashion in the course of his wide-ranging examination of
consciousness, ideology and modesof production.His conclusionsabout
science are, therefore, less clear-cut and there have been somewhat
different interpretationsof how far he saw social factors as determining
the contentof science.Consequently,it will be necessaryto look briefly
at two different approaches,both of which claim to be forms of Marxist
analysis. (There are numerous important writings on science in the
Marxist tradition which I have too little space to examine here: for
example, Marcuse, 1962; Habermas, 1972.) Let me begin with those
relevantfeaturesof Marx's work aboutwhich thereis little disagreement.
The history of mankindtakesplace within the natural setting provided
by the objective world, a setting which is continuously transformedby
humanactions. By acting on the natural world man producesthe means
of his own existence.The repetitive relationshipsbetweenpeoplewhich
grow out of this productive, economicactivity are fundamentalto all
societies,constitutingthe major influence upon their overall structures
and their mental productions.In the courseof acting upon the natural
world man generatesknowledge about that world. This knowledge is
formulatedin responseto the interestsand economicconcernsof various
social groups; and is constrainedby the ideological assumptionscurrent
within particular modes of production. Such knowledge is used to
manipulatenatural phenomenaand to support, or, in certain circum-
stances,to transform existing social relationships.
The growth and elaboration of scientific knowledge of the natural
world was greatly stimulatedby the emergenceof capitalist society. The
economic tasks faced by the bourgeoisiein the seventeenthand eight-
eenth centuries drew attention to certain technological issues which
stimulated an increasingly practical approach to questions of natural
6 Scienceand the Sociologyoj Knowledge
philosophyon the part of the intellectualrepresentatives
of this class.In
due course, the new natural philosophy began to generatepractically
effectivescientific knowledge,which was usedby the capitalist classas a
direct meansof economicproduction.As scientific knowledgeimproved
production, so more resourceswere made available for the support of
scientific investigation.Consequently,throughoutthe nineteenthcentury
and up to the presentday, sciencehas becomeintimately bound up with
the capitalist economy and the continuous technological innovation
which capitalismappearsto require.
The bourgeoisie... hascreatedmore massiveand more colossalforces

thanhaveall precedinggenerationsaltogether.Subjectionof Nature's
forces to man, machinery, application of chemistry to industry and
agriculture, steam navigation, electric telegraphs... what early cen-
tury hadevena presentimentthat such productiveforces slumberedin
the lap of sociallabour?(Marx and Engels,1965,p. 47)
Thus capitalism needs and promotes 'the developmentof the natural

sciencesto their highestpoint' (Marx, 1973, p. 409).
Initially natural science,like capitalism itself, was a liberating force,
setting men free from superstition and the ideological confusions of
religious thought. But in due course science necessarily became an
exploitativeresourcefor the bourgeoisie.Particularlywithin the realm of
industrial production,sciencecontributedsignificantly to the 'dehuman-
isation of man' (Marx, 1974, p. 97). Objective,scientific knowledgewas
increasingly used to create economic and administrative technologies
which restricted narrowly the actions and initiatives available to their
'operatives'.
The unity of thought and action, conceptionand execution,hand and

mind, which capitalism threatened from its beginnings, is now
attacked by a systematicdissolution employing all the resourcesof
scienceand the various engineeringdisciplines based upon it. The
subjectivefactor of the labor processis removedto a place amongits
inanimate objective factors. To the materials and instruments of
productionare addeda 'labor force', another 'factor of production'
... This is the ideal toward which management... uses and shapes
every productiveinnovation furnished by science.(Braverman, 1974,
pp.171-2)
The centraltheme,then, of the Marxist analysisof scienceis to seethe

latter as a social creationand to stressthat its consequences,
its usesand
the direction in which it developscan only be understoodin relation to
the wider social context. There are some similarities with Durkheim;
scienceis viewed from an evolutionaryperspective,as beingset in motion
TheCustomarySociologicalView oj Science 7
by changesin the structureof society and as underminingthe effective-
nessof the religious beliefs which helped to hold togetherpre-capitalist
social forms. But Marx goes farther than Durkheim, although by no
means far enough, toward analysing the production of science in
complex, differentiated societies.This is possible for Marx becausehe
avoids Durkheim'srelianceon an inherentlyambiguousnotion of simple
'correspondence'betweenconceptsand thought, on the one hand, and
generalfeaturesof social structure,on the other. Marx offers in addition
a dynamicaccountof social processeswhich canbe usedto describesome
of the links betweenscienceand society. In particular, he stressesthat
societiesare composedof relatively distinct groupings, the membersof
which have opposing interests as well as an unequal capacity for
controlling the actions of others. Consequently,the direction taken by
modern science,its rapid rate of growth and the mannerof its applica-
tion in industry and government can be seen to have been largely
determined by the technological objectives of a particular dominant
group, namely, the bourgeoisie.The bourgeoisiehas beenthe one group
in capitalistsocietyable to deploy surpluseconomicproduct to generate
new scientific knowledgedirectly relevant to its own objectives.
But what of the form and content of scientific knowledge? In the
following passage,Marx seemsto come close to arguing that the very
laws of natural science are merely a device for achieving socially
contingentobjectives.Under capitalism,he claims:
... nature becomespurely an object for mankind, purely a matter of

utility; ceasesto berecognisedas a powerjor itself; and the theoretical
discovery oj autonomouslaws appears merely as a ruse so as to
subjugateit underhumanneeds,whetheras an object of consumption
or as a meansof production.(1973, p. 410, emphasisadded)
This readingof Marx has been developedmost fully within the Russian
Marxist tradition. For example, in a now famous paper, Boris Hessen
(1931)tries to interpret Newton'sPrincipia within a Marxist framework.
He presentsevidenceto show, first of all, that there was a close identity
betweenthe central technical problems facing the entrepreneursof the
emergentcapitalist economy during Newton's period and the major
scientific problemsformulated by natural philosophersat that time (see
also Merton, 1936). He also tries to show that these same technical
problems provided the focus of Newton's influential work. Conse-
quently, Newton'swork can be seenpartly as an indirect responsefrom
the intellectualsof the bourgeoisclassto difficulties arising in the course
of economic production. However, the content of Newton's Principia
cannot be explained quite as simply as that. Although the economic
factor is fundamentalto the materialist conceptionof history, this does
not meanin Hessen'sview that it is the sole determininginfluence upon
any particular set of ideas. Accordingly, he attempts to complete his
analysis of Newton's work by showing how Newton drew selectively
upon the cultural resourcesavailable to a member of his class, for
example,in the form of political, juridical, philosophical and religious
beliefs, and by showing how theseideological elementsinfluenced and
limited Newton's thought.
Although Hessenstrongly urgesthat scienceis not 'a passive,contem-
plative acceptanceof reality, but ... a meansto effect its active recon-
struction', he managesto reconcile this view with commitment to the
Marxist-Leninistnotion 'that genuinescientific knowledgeof the laws of
the historical process leads with irrefutable iron necessity' to certain
political conclusions(Hessen,1931, p. 211). Hessen'sposition is clearly
not without ambiguity or irony. However, I do not intend to examine
further Hessen'sessayor to assessits merits and defects.It merely serves
hereto illustratethat Marx canbe interpretedin a strongsense,that is, as
implying that the contentof establishedscientific knowledgeshould be
treated to a considerableextent as the outcome of specifiable social
processes.On the whole, however, academicsociology has not adopted
this reading of Marx. Merton, for example, having noted that Marxist
analysis allows the different spheres of mental production varying
degreesof independencefrom the economicbase,decidesthat Marx and
Engels regard scienceas having a greater degreeof independencethan
any other realm of thought. The following passageis quoted as crucial
evidencefor this interpretation:
With the changeof the economicfoundationthe entireimmensesuper-

structure is more or less rapidly transformed. In considering such
transformationsthe distinction should always be made betweenthe
material transformationof the economic conditions of production
which can be determinedwith the precision oj natural science,and the
legal, political, religious, aestheticor philosophic-inshort, ideologi-
cal forms in which men becomeconsciousof this conflict and fight it
out. (Marx, 1904, p. 12, italics addedby Merton)
But this passageis not viewed on its own. It is interpretedin the light of
what Merton regardsas the overall trend in Marx's treatmentof science.
, ... one line of development of Marxism, from the early German
Ideology to the latter writings of Engels, consists in a progressive
definition (and delimitation) of the extent to which the relations of
production do in fact condition knowledge and forms of thought'
(Merton, 1973,p. 14). Merton'sfinal reading,therefore,is that although
the focus of attentionof natural sciencemay be socially determined,this
is true neitherof its conceptualapparatusnor of its substantiveconclu-
sions.Marx and Engelsareseenas grantingsciencea statusquite distinct
from that of ideology.
The CustomarySociologicalView ofScience 9
It might be objected at this point that Merton is unlikely to be an
accurateinterpreter of Marx, given that his own theoretical frame of
reference,that of functional analysis,is so different. Whetherthis is true
or not, Merton's interpretationof Marx, which is also his own view, is
important if only becauseMerton's work has exertedsuch a pervasive
influence on the sociological analysis of science.Moreover, it is worth
noting that many authorsstrongly committed to a Marxist analysis of
sciencehavereachedsimilar conclusions.To illustrate this point, let me
discussbriefly a recentessayby Rose and Rose. Theseauthorscertainly
pay more attention than does Merton to Marx's idea that science
generatesits own ideology. They emphasisethat 'scientism'or 'positiv-
ism' has becomeso dominantin present-dayindustrial societiesthat any
knowledge-claim which falls outside its scope is widely regarded as
necessarilyvacuous.They write that:
... sciencebecomesan ideology and scientists the ideologists. How

doesthis work? As the material world controls the limits of an inter-
pretationof the scientistin his own work, the answerlies, as Marx and
Engelssaw, outsidethe preciseresearcharea,wherethe scientist,freed
from such constraints,talks (typically in the name of science) pure
ideology. In the nameof science,invoking neutrality, techniqueand
expertise,the scientistsupportsthe ruling strata... (1976, pp. 8 -9)
In this passage,the authorsmake a distinction betweenthe technical

knowledge-claimsascientistmakeswithin his own researchnetwork and
the claims he makesin other social contexts.They suggestthat the former
are normally controlled by the natureof the physical world. It must be
recognised,of course, that scientists will sometimesbe influenced by
social pressuresto proposeunjustified knowledge-claims.But, as long as
there is no interference from outside with the technical criteria of
adequacyapplied by the researchcommunity, these socially generated
claims will be judgedto be inadequateby other specialistsand they will
be rejected.Thus within a specific researchareathose knowledge-claims
which come to be acceptedas valid can be seento be non-ideological.
They provide an accurateaccount of certain features of the physical
world and their content is, accordingly, independentof participants'
social relationshipsand vested interests. However, in other social con-
texts, the situationis quite different. Non-scientists,and indeed special-
ists in other areas,seldom have the technical competenceto assessthe
adequacyof a particularscientist'sclaims. It follows that he will be able
to use specialisedknowledge so as to furnish an apparent technical
rationalefor policies which expresshis own social interestsas well as the
interestsof other groups on whose behalf he is acting.
Claims madeby scientistsin the wider social context, then, will often
be ideological; but this will be obscuredby their technicalcontentand by
scientists'ability to invoke the 'objective facts of the natural world' as
leading inevitably to certain economic,political and social conclusions.
This is the kind of analysis which is generatedby the Roses' Marxist
approach.It is clearly different in emphasisfrom that of Merton and
Durkheim, and directs our attention to important questionswhich are
largely ignoredby the latter. (I will return to someof theseissuesin the
final chapter.)Yet the Roses'readingof Marx appearsto coincide with
that of Merton on one crucial issue. For them as for Merton, Marxist
analysisrecognisesthat scientists'knowledge-claimswithin their precise
researchareasare non-ideological.
Despite important analytical differences, therefore, the writers we
haveexaminedso far, with the possibleexceptionof Hessen,are agreed
on at least the following points: first, that scienceflourishes in large-
scale, industrial (capitalist) societies and that within such societies,
scientistscreatedistinct communitieswhich regulate the production of
certified knowledge; secondly, that although the rate of growth, the
focus of attentionand the use madeof scientific knowledgeare in large
measuresocially determined,its content is independentof social influ-
ences;and thirdly, that scientific researchcommunitiesare likely to have
special social characteristicswhich reduce the impact on members'
technicalwork of suchdistorting factors as bias, prejudiceand irration-
ality, and which are, therefore,crucial in enablingscientiststo generate
objectiveknowledge.It would even be possibleto bring Hesseninto the
fold if we were to allow a distinction to be made betweenthe situation
obtaining within capitalist society and that characteristicof socialist
societies. For, as was noted above, Hessen retains the notion that
ineluctable laws of nature are available within the Marxist-Leninist
framework. His main thesismay, therefore,be interpretedas a claim that
Newtonian science was partly pseudo-science,distorted by the social
relationshipsof capitalismand due to be replacedby the truly scientific
and determinateformulations forthcoming within socialist society (see
1931, pp. 211-12).
In the next section I shall examine the work of two more major
contributorsto the sociologyof knowledge.This will enableme to begin
to make explicit someof the philosophicalassumptionswhich underlie
the tendencyto regardscienceas a specialkind of sociologicalproblem.
MORE RECENT VARIANTS: MANNHEIM AND STARK

KarI.Mannheimis usually regardedas a centralfigure in the development
of the sociologyof knowledge(seeCurtis and Petras,1970). His work is
complicatedand, although his position on crucial issues undoubtedly
changedas his thoughtevolved,he neverprovided a clear final statement
of his framework of analysis.I will make no attempt, therefore,to give
an overall account of his interpretation of the social creation of
knowledge.I will insteadfocus narrowly on his treatmentof scienceas a
subject for sociologicalstudy.
Mannheim'ssociologyof knowledgeincludesa numberof ideastaken
directly from Marxism; for example, a belief in the importance of
economicinterestsand classgroupings,and in the ideological character
of much social thought. In the courseof his work, Mannheim tried to
extendthe Marxist notion of the 'existentialbase'to cover generations,
sects and occupationalgroups; he also supplementedthe concept of
'ideology' with an associated conceptof 'utopia'; and he provided
historical documentationof the connectionsbetweenthought and social
factors by meansof severalempirical studies.But the Marxist strain in
his work was combinedwith elementstaken from the Germanacademic
tradition of neo-Kantianthought (Mannheim, 1952, p. 5). One of the
main ideas which he adopted from this tradition was that a radical
distinction had to be made betweenthe methods and conceptsof the
natural sciences,on the one hand, and those of .the social sciencesand
historical thought, on the other. This has been widely discussedin
relation to the writings of Dilthey and others(Outhwaite, 1975). I shall
simply mention,therefore,a few of the points emphasisedby Mannheim.
In the first place, the phenomenaof the material world and the
relationshipsbetweenthem are seenas being invariant (Mannheim, 1936,
p. 116). Mannheim regularly refers to the natural world, and to the
conceptsappropriateto its study, as being 'timeless and static'. Valid
knowledge about such objective phenomena,he maintains, can be
obtainedonly by detached,impartial observation,by reliance on sense
data and by accuratemeasurement(Mannheim, 1952, pp. 4-16; 1936,
pp. 168-9). Becausethe empirical relationshipsof the natural world are
unchangingand universal, the criteria of truth by which knowledge-
claims areto be judgedarealso permanentand uniform (1936, p. 168). It
follows that naturalsciencedevelopsin a relatively straightline, as errors
are eliminatedand a growing number of truths are discerned.In short,
scientific knowledgeevolves through the gradual accumulationof per-
manentlyvalid conclusionsabout a stablephysical world.
Cultural products, however, cannot be investigated properly by
methodsof detachedobservationor by means of static concepts.For
correct categorisationand understandingof cultural phenomenaneces-
sarily involves the interpretationof participants'meanings;and mean-
ings cannot be simply observedlike objectsin the external world. Each
historical period and eachsocial group has its own distinctive valuesand
meanings.Each analyst begins from his own culturally specific frame-
work of meanings.Accordingly, no product of human culture can be
analysedadequatelyfrom a timeless perspective.The interpretationof
meaningsis essentiallydynamic. It must deal with the uniquefeaturesof
each cultural epoch and must be undertakenanew by the representa-
tives of every succeedinghistorical period (Mannheim, 1952, p. 61).
Moreover, therecan be no such thing as detached,uniform observation
of cultural products.Their meaningmust be acquired,instead,by means
of involvement and sympatheticunderstanding(Mannheim, 1936, p.
170).
In accordancewith this characterisationof the sciences and the
contrastingcultural disciplines,Mannheimcontinually treats the advan-
ced physical sciences as a special case from the perspectiveof the
sociology of knowledge.
Are the existential factors in the social processmerely of peripheral

significance... or do they penetrateinto the 'perspective'of concrete
particular assertions?... The historical and social genesisof an idea
would only be irrelevant to its ultimate validity if the temporal and
social conditions of its emergencehad no effect on its content and
form. If this were the case,any two periodsin the history of human
knowledgewould only be distinguishedfrom one anotherby the fact
that in the earlier period certainthings were still unknownand certain
errors still existed which, through later knowledge were completely
corrected.This simple relationship... may to a large extent be appro-
priate for the exactsciences... (1936, p. 271)
In the light of several statementsof this kind and in view of the

importance for the whole of Mannheim's thought of the distinction
betweennatural scienceand socio-historicalthought, virtually all inter-
pretershaveregardedhim as being entirely consistentand unambiguous
in treating scientific knowledge as beyond the scope of sociological
analysis(Mannheim,1952, p. 29; Merton, 1973, p. 21; Bloor, 1976, p. 8).
However, if we read Mannheim with particular care, we will observe
that, on at least a few occasions,he seemsto waver on this point. For
example, he follows the quotation given immediately above with the
following qualification: 'althoughindeedtoday the notion of the stabil-
ity of the categoricalstructureof the exactsciencesis, comparedwith the
logic of classicalphysics, considerablyshaken' (1936, p. 271). In this
passage,Mannheimseemsto be questioninghis own characterisationof
scientific knowledgeas timelessand immutable. His uncertaintyon this
point becomesmost noticeablewhen he is dealing with epistemological
issuesandhe finally offers a resolutionof the problemin epistemological
terms.
The basic epistemologicalproblem faced by Mannheim, as by Durk-
heim, is that of relativity. The sociology of knowledgeassertsthat all
'social thought', all thought outside the exact sciences,is relative to a
particularsocial position or undertakenfrom a particular perspectiveor
formulated in accordancewith certain social interests. Thus in this
sphere,thereappearto be no generallyapplicablecriteria for jUdging the
validity of any specific assertion.But clearly the sociologyof knowledge
is itself part of this intellectual domain. It seemsto follow, therefore,
that thereis no way of assessingthe validity of its own claims-including
the central claim that all social knowledge is existentially determined.
Mannheim, of course, wishes to reject this conclusion and he tries to
show that the assertionsof the cultural sciences,although different in
kind from thoseof the exactsciences,canstill furnish true knowledge.In
seekingto establishthis point Mannheimis almost, but not quite, led to
revise his epistemologicalassumptionsabout the natural sciences.
Mannheim does not abandonscienceentirely as a subject for socio-
logical investigation and in a few pages of Ideology and Utopia he
interprets the rise of science broadly along Marxist lines (1936, pp.
165-9): He argues that the methodology adopted by the advanced
scienceswas a by-product of the Weltanschauungof the ascendant
bourgeoisie.The world view of this class, which he describesas 'demo-
cratic cosmopolitanism',deniedthe valueof personal,qualitative '.know-
ledge'. Only formulations which were universa~ly valid and necessary
were allowed to standas genuineknowledge.
Similarly, every kind of knowledgewhich only certainspecific histori-

cal-social groups could acquire was distrusted. Only that kind of
knowledgewas wanted which was free from all the influences of the
subject'sWeltanschauung.What was not noticedwas that the world of
the purely quantifiableand analysablewas itself only discoverableon
the basis of a definite Weltanschauung.Similarly, it was not noticed
that a Weltanschauungis not of necessitya sourceof error, but often
gives accessto spheresof knowledgeotherwiseclosed.(1936, p. 168)
As the bourgeoisieachieved a position of social and political pre-

eminence,so scientific knowledgeand its associatedepistemologycame
to pervadeand to dominate intellectual life. Consequently,Mannheim
maintains, virtually all knowledge-claimshave come to be measured
againstthe particular epistemologyderived from the dominant form of
scientific knowledge.
The particularity of the theory of knowledge holding sway today is

now clearly demonstrableby the fact that the natural scienceshave
beenselectedas the ideal to which all knowledgeshould aspire. It is
only becausenatural science,especiallyin its quantifiable phases,is
largely detachablefrom the historical-socialperspectiveof the investi-
gator that the ideal of true knowledge was so construed that all
attemptsto attaina type of knowledgeaiming at the comprehensionof
quality areconsideredas methodsof inferior value. (1936, pp. 290-1)
In responseto this situation, Mannheimtries to formulate an alternative

epistemology which is appropriate for qualitative, existentially
determinedknowledgeand which is in accordancewith the conclusions
of his own version of the sociology of knowledge.
In Mannheim'sview the sociology of knowledge has shown that the
'positivist epistemology'of bourgeois society is itself partial and the
product of a particular, limited Weltanschauung.This epistemologyis,
therefore, inadequatein the sense that it fails to recognise its own
limitations and its own dependenceon historically specific assumptions.
Accordingly, its application must be confined in future to the special
realm of knowledge about the physical world, to which alone it is
appropriate,and it must be supplementedby a broader epistemology
which recognisesthe partial characterof all humanperspectives.It has
becomepossiblefor us today (Mannheimwas writing in the 1930s),in a
way which was impossible before the advent of the sociology of
knowledge,to treat situationallydetachedknowledgeas a marginal and
specialcaseof the situationally conditioned. From this epistemological
position one assumes'the inherently relational structure of human
knowledgeGust as the essentiallyperspectivisticnature of visually per-
ceived objects is admitted without question)... It is not intended to
assertthat objectsdo not exist or that relianceupon observationis useless
and futile but rather that the answerswe get to the questionswe put to
the subjectmatterare, in certain cases,in the natureof things, possible
only within the limits of the observer'sperspective'(1936, p. 300). This
doesnot meanthat we are abandoningthe notion of 'objectivity' or the
possibility of establishing'facts'. Ratherit meansthat our conceptionof
objectivity has to change.
Thereis no needto deny that peoplecan often reachwhat they take to
be 'objective' conclusionsabout particular phenomena, i.e. conclusions
which are verifiable by the application of establishedprocedures.Yet
epistemologicallytheseobjectiveconclusionsmust be regardedas incom-
plete, as the product of a specific perspectiveand as open to revision in
new socialsituationswhereother perspectivesare broughtto bear. In the
caseof different observersworking within a commonframe of reference,
objectivity must be conceived as the application of agreed criteria of
adequacy to particular knowledge-claims (1936, pp. 300-1). When
participantshavedifferent perspectives,however,objectivity can only be
attainedin 'a more roundaboutfashion'. Mannheim continuesto treat
objectivity as being indistinguishablefrom intellectual agreement,but
arguesthat the latter will be possibleonly in so far as the results of each
perspectivearetranslatedinto the other andreconciled, usuallyat a more
general level. This notion.of divergent frames of reference becomes
central to his new epistemology.(Mannheim writes of two alternative
versionsof his theory of knowledge, but both versions dependon this
principle of resolving the differencesof specific perspectiveswithin a
more comprehensiveformulation.) Where some choice has to be made
betweenperspectives,pre-eminenceis given to that 'which gives evidence
of the greatestcomprehensiveness and the greatestfruitfulness in dealing
with empirical materials'(1936, p 301). Thus the old static epistemologi-
cal conceptionof true statementscorrespondingto the realities of ~
directly observable world has been abandonedfor most realms of
thought-andperhapsfor thought in its entirety.
We will have to reckon with situationaldeterminationas an inherent

factor in knowledge,as well as with the theory of relationismand the
theory of the changingbasisof thought... we must reject the notion
that there is a 'sphereof truth in itself' as a disruptive and unjustifi-
able hypothesis.It is instructiveto note that the natural sciencesseem
to be, in many respects, in a closely analogoussituation . .. (1936,
p. 305, emphasisadded)
This last sentencebrings us back to the 'marginal, special case' of

science.It is followed in Mannheim'stext by a single page discussing
certain developmentsin modern physics which were then quite recent.
The point of the discussionis to show that the establishedcertaintiesof
classicalphysicsappearedat that time to be giving way to a much more
'relativist' framework of ideas. Mannheim notes that in quantum
mechanicsit had come to be regardedas impossible to conceive of
measurements independentlyof the actions and techniquesinvolved in
measuring.He points out that empirical relationshipsat the sub-atomic
level were thought to be inherently indeterminateand that traditional
notions about particles having a specific location and a definite and
ascertainabletrajectoryof movementhad now beenabandoned.And, of
course,he mentionsEinstein'stheoryof relativity and the way in which it
brought the position of the observerinto the very equationsof physics
(1936, pp. 305-6).
Mannheimsuggeststhat this trend of thought in natural scienceis, in
its 'unformulatedrelationism', surprisingly similar to his own. At this
point in the argumentit seemspossible that Mannheim will be led to
reject his previous neo-Kantian portrayal of scientific knowledge as
composedof universaland static truths. It seemspossiblethat he will go
on to claim that knowledgeof the physicalworld, like that of the social,
dependson the kinds of questionsthat we pose, on the purposesof the
knowers and on their socially derived perspectives.There would be no
epistemologicalinconsistencyfor Mannheim in adopting this position.
For his relational epistemologyprovides for the 'objectivity' of socially
derivedknowledge-claims;and it is no lessconvincingto claim that active
commitmentto a partial perspectivewill reveal truths about the physical
world than it is to make the same claim regarding the social world.
Furthermore,this line of reasoningwould have enabled Mannheimto
avoid treatingone major areaof knowledgeas a partial exceptionto his
general epistemologicalprinciples and it would have enabled him to
undertakea sociologicalanalysisof that recentand puzzling transforma-
tion in modernphysics.
Yet, despitethesepotential advantages,Mannheim draws back from
the conclusionthat scientific knowledgeis in any way socially contingent.
He doesnot take the decisivestepof claiming that his alternativeepiste-
mology fits the physical sciencesin exactly the sameway as it does the
historical disciplines. He is careful to state that the two realms are no
more than analogous.Their appropriateepistemologiesare parallel but
distinct. The relationismof physicalscienceis best seenas a specialcase
of the generalprinciple of relationism.What he seemsto meanby this is
that, whereasthe knowledgeavailableto observersin the socio-historical
sphereis necessarilyrelatedto their social position, cultural background,
group interests,and so on, the knowledgeattainableby observersof the
physical world is necessarilyconstrainedonly by their position in time
and space.Thus Mannheim reachesa final position from which both
types of knowledgeare seenas inherently limited and revisable; but he
maintainsthe distinction betweenthe two spheresby claiming that the
limitations or constraintsessentialto eachintellectual domain are quite
different in character.
To summarise,we can say that Mannheimhasdonethreethings in his
discussionof epistemology.He has tried to restrict the scope of the
'positivist epistemology'to the sphereof naturalscience.He has tried to
outline an alternative, relational epistemology for socio-historical,
existentially determinedthought. And he has raised the possibility that
the old epistemologyis not even entirely adequatefor the advanced
physical sciences.But he has resolved his doubts about the status of
scientific knowledgewithout seriously challengingthe orthodox episte-
mology and, consequently,without openingthe door for a fully-fledged
sociologyof science.One reasonwhy Mannheimwent no further in this
directionmay well havebeenthat he was entirely dependenton the work
of historiansof sciencefor his own views about scientific development.
He had to wait, therefore, for suitablestudiesin the history of ideasto
emerge;and thesestudiesdid not comeuntil long after his death. But in
addition there is the fact that so much of Mannheim's thought was
formulatedin terms of the epistemologicaldistinction betweenscientific
and socio-historicalknowledge.To haveattempteda seriousrevision of
the epistemologyassociatedwith natural sciencewould have necessitated
a wholesalerevision of his own sociologicalcorpus.
I have suggested,then, that some passagesin Mannheim'swritings
contain the germ of a new philosophical view of science, one more
favourable than the standardview to the possibility of a sociological
approachto scientific knowledge.But thesepassageshavegenerallybeen
missed or they have been interpeted so as to make them entirely
consistentwith Mannheim'smore frequently expressedidea that science
constitutesa specialepistemological,and hencesociological, case.Since
Mannheim, this latter view has becomefirmly entrenchedin the socio-
logy of knowledge. This is well exemplified in the writings of Werner
Stark, which appearedtwenty years or so after the publication of
Ideology and Utopia. In Stark'saccountof the sociology of knowledge,
there is no longer any questioningof the statusof scientific knowledge.
Mannheim'sepistemologicalreservationshavebeenforgotten and argu-
ments which remained largely implicit in Mannheim are marshalled
clearly and explicitly by Stark to exclude science from sociological
consideration.
Stark begins by stating what he regards as a necessaryassumption
aboutthe physicalworld, namely,that it hasa fundamentalpermanence.
Thereis inherentin the naturalworld, he suggests,a determinatestateof
affairs which scientists can observe and represent with increasing
accuracy and completeness.Because physical scientists can achieve
an abiding correspondencebetweenan invariant nature and their own
formulations, they are able to establish their distinctive degree of
intellectualconsensus.Scientific consensusis a product of the objectivity
of scientific knowledge. This situation, however, does not apply in
all so-called sciences.Like Mannheim, Stark believes that there is a
fundamentaldifferencebetweenthoseconcernedwith the physicalworld
andthoseconcernedwith society. 'The facts of societyaremade,andever
re-made,by us, whereasthe facts of natureare not. They are data in a
much more stringent meaningof the term' (Stark, 1958, p. 165).
The secondmajor contentionin Stark'sanalysisseemsinitially to be
slightly at odds with the first. For, having maintainedthat there is a
straightforward,unproblematiccorrespondencebetweentrue scientific
knowledgeand invariant natural phenomena,he suggeststhat another
reasonwhy knowledgeof the physicalworld is cumulativeand reliable is
that suchknowledgeis always formulatedwithin a perspectiveorganised
in terms of technicalefficiency.
. . .whereasman has more than onceshifted his vantagepoint for the

considerationof social facts so that thesefacts appearto him in ever
new, and often surprising, outlines, he has always kept to the same
spot for surveying the facts of nature... so that these latter facts
have always offered to him the self-same surface. He has merely
learnedto look ·more closely... Whetherhe likes it or not, he must,
under all circumstances,pursue, among others, the economic and
technologicalvalues,the valueof science.(1958, p. 166)
The dependenceof scientific knowledgeon a particular perspectiveon,

or attitudetowards,the naturalworld appearsat first sight to weakenthe
belief in the certainty and definitive character of that knowledge.
Scientific knowledge seems to have become to some extent socially
contingent. But Stark resolves this problem by postulating that the
centralvalueguiding man'sattemptsto understandthe natural world has
necessarilyalways beenthe same. Not only have men always sought to
understandnaturein orderto exertcontrol over natural processesbut, in
Stark's view, there is no alternative stanceavailable to them. Thus the
factual content of science is beyond the scope of sociological inquiry
becauseit is universal.It is the productof the cumulativeapplicationof a
uniform perspectiveto a determinatenatural world.
This view of scientific knowledgehas numerousfurther implications,
someof which are illustrated byStark. One obvious implication is that,
as it is essentiallythe samecorpusof knowledgewhich has beenbuilt up
over time, there can be no change of meaning in the factual basis of
science.True scientific knowledge can be formulated in only one way.
The body of knowledge becomesincreasingly comprehensivewith the
passageof time, but the genuine factual basis is neither revised nor
reformulated.A secondimplication is that scientific discovery is differ-
ent from that in other realms of cultural activity. In science,discovery
consistsnot so much in creatingnew meanings,as it does in philosophy
and the arts, as in recognising realities which, in some sense,already
exist. For example: ' ... the astronomermerely attempts to grasp the
data, the pre-existent data: he is entirely controlled by them ... the
scientistallows himself to be impressedby the objective truth of reality,
the man of culture expressesthe values in which he believes' (1958,
p. 167).
Stark realises,of course,that the dataof sciencecan be conceptualised
and interpretedat variouslevels of generality,and that certainlevels and
conceptionsmay be linked only indirectly with empirical observations.
The higher levels of analysis, however, he regards as a metaphysical
appendageto real scientific knowledge. The metaphysicsof physics, he
accepts,may well be influenced,even determined,by social factors. But
the metaphysicalaccompanimentscan be separatedout from positive
science.'For sciencealwaysasks: what is, while in the questionsraisedby
metaphysicsthere always occurs the further, and disparatequestion of
why?' (1958, p. 175): He gives as an example of positive, factual
knowledge, the proposition that heavy bodies fall at an accelerating
speed.Another example is taken from Darwinian theory, namely, the
proposition that 'life is an ongoing struggle for survival in which the
relatively weak are progressivelyeliminatedandonly the relatively strong
survive' (1958, p. 170). This is a particularlyinterestingexample, because
the Darwinian thesis can be seen as deriving at least partly from
Malthus'sanalysisof social dynamics.Stark argues,however, that even
this is not a casewhere genuine scientific knowledge has been signifi-
cantly moulded by social factors.
Socialdevelopmentsdo not determinethe contentof scientific develop-

ments,simply becausetheydo not determinenaturalfacts; but they may
well open the eyes of the scientists to natural facts which, though
pre-existentand always there, had not beendiscoveredbefore. This is
what happenedin the caseof Darwinism... (1958, p. 171)
Whatever its origins, the Darwinian proposition is, in Stark's view,

absolutelytrue. It hasbeenshownto be a summary statement of the facts
of natureand is, therefore,unrevisable.The facts of natureare beyon.d
man'scontrol. Consequently,the contentof any propositionembodying
such facts cannotbe determinedby social factors.
Stark maintainsthat there is 'causefor rejoicing' in the fact that the
main movements in the sociology of knowledge have agreed in all
essentialswith his characterisationof science (1958, p. 167), and he
quotessuch diverseauthorsas Marx, Lukacs, Mannheim,Alfred Weber
and Merton to illustrate this agreement.It is clear from the discussion
abovethat therehas beenless certaintyamongsociologistsof knowledge
about the natureof sciencethan Stark seemsto imply. Nevertheless,I
think that he is substantiallycorrect if he is taken to mean that most
writers in this field haveoperatedwithin the limits of a single, standard
philosophyof scientific knowledge. It is largely becausesociologistsof
knowledgehavebeenunableto offer a seriousalternativeto the standard
epistemologicalview of science that they have been propelled into a
position from which scientific knowledgeand the intellectualactivities of
scientistshave to be treated with special deference(for a discussionof
'inductivism' in the history of science,see Agassi, 1963). In the next
section, I will outline briefly the 'standardview of science' which has
beenimplicit in the discussionso far. In the following sectionI will show
how it has influencedthosestudieswhich sociologistshave made of the
socialworld of science.
THE STANDARD VIEW OF SCIENTIFIC KNOWLEDGE

Most sociologistsof knowledgehave adoptedsomeversion or other of
what Scheffler(1967) has called 'the standardview of science'.I do not
meanby this that they haveall endorsedevery statementthat I shall make
below in trying to summarisethe standardview. But I do meanto suggest
that, althoughdifferent analystshaveusedthe standardview in different
ways and with varying emphasis,sociologists'thought about scienceas a
social phenomenonhas usually been formulated within this framework
of assumptions.The readerwill find that all the main points containedin
the following paragraphshave alreadybeenillustrated in the discussion
above.
From the perspectiveof the standardview, the natural world is to be
regardedas real and objective. Its characteristicscannot be determined
by the preferencesor intentions of its observers.These characteristics
can, however, be more or less faithfully represented.Science is that
intellectual enterpriseconcernedwith providing an accurate accountof
the objects,processesandrelationshipsoccurringin the world of natural
phenomena.To the extent that scientific knowledgeis valid, it reveals
and encapsulatesin its systematicstatementsthe true characterof this
world. As Galileo puts it: 'the conclusionsof naturalscienceare true and
necessary,andthe judgmentof man has nothing to do with them' (1953,
p. 63). Although the natural world is, in a certain sense,undergoing
continuouschangeand movement,there exist underlying and unchang-
ing uniformities. Thesebasic empirical regularities can be expressedas
universaland permanentlaws of nature,which tell us what is alwaysand
everywhere the case. Unbiased, detached observation furnishes the
evidence on which these laws are built. The creation of scientific
knowledge 'begins with the plain and unembroideredevidenceof the
senses,with innocent,unprejudicedobservation... and builds upon it a
great mansionof natural law' (Medawar, 1969, p. 147). Indeed, obser-
vationallawsareno more than generalpropositionssummarisinga body
of reliable factual evidence.The validity of the factual foundation of
scientific knowledgecan be guaranteedwith a high degreeof confidence
becausesciencehasevolvedstringentcriteria, for example,in connection
with experimentalprocedures,by meansof which empirical knowledge-
claims are evaluated and their accurate representationof empirical
phenomenais ensured.Thus acceptedscientific knowledge, becauseit
hassatisfiedtheseimpersonal,technicalcriteria of adequacy,is indepen-
dent of thosesubjectivefactors, such as personalprejudice, emotional
involvement and self-interest,which might otherwise distort scientists'
perceptionof the externalworld.
Although the body of scientific knowledge is basically empirical, it
doescontainhigh-level generalisationswhich are not observationallaws
and which in somecasescannotbe directly derivedfrom or testedagainst
observations.These more abstract and more speculativepropositions
play an important role in scientific thought by explaining observed
regularities,by co-ordinatingseparateobservationallaws into coherent
intellectual frameworks and sometimesby revealing observablepheno-
menawhich werepreviouslyunknown. In certaincases,the development
of new observationaltechniquesleadsto the direct confirmationof these
abstractspeculationsand they eventuallybecomeindistinguishablefrom
ordinary observationallaws (O'Neil, 1969). But it is not necessaryto
conceiveof theoreticallaws as actually representingthe realities of the
natural world. It is not surprisingthereforethat, like statementsabout
the 'ether', they are frequently abandonedby scientists when their
usefulnesshas come to an end.
A fundamentaldistinction must be made, therefore, betweenobser-
vational laws and theoretical laws (Nagel, 1961, ch. 5). The latter are
revisableanddispensable,but the former are not. Whilst the former deal
with observablefacts, the latter often deal with unobservableentities.
The CustomarySociologicalView of Science 21
Neverthelessevery effort is made to verify or to test theoretical laws.
When a theoretical law generatesinferenceswhich are not upheld by
observation,it is either revisedin accordancewith the new evidenceor it
is renouncedin favour of an alternative hypothesis.In casesof uncer-
tainty, various hypotheseswill be testeduntil one is found which fits the
full rangeof data. This hypothesisthen becomesa candidatetheoretical
law. Although there is constantchangeand revision at the theoreti<;al
level, this is not incompatiblewith cumulativedevelopmentat the factual
level. The establishedfacts subsumedunder an abandonedtheoretical
law are typically passedon to its successorwhich, in addition, will have
brought within its scope a number of newly certified facts about the
natural world. 'Thus it is that science can be cumulative at the
observationalor experimentallevel, despiteits lack of cumulativenessat
the theoreticallevel. Throughoutthe apparentflux of changingscientific
beliefs, then, there is a solid growth of knowledge which represents
progressin empirical understanding'(Scheffler, 1967, p. 9).
The basic, observationallaws of scienceare consideredto be true,
primary and certain, becausethey are built into the fabric of the natural
world. Discoveringa law is like discoveringAmerica, in the sensethat
both werealreadytherewaiting to be revealed(MacKinnon, 1972, p. 16).
Oncean observationallaw has beendiscoveredit appliesuniversally and
it commandsuniversal assent. There may be some slight room for
cultural variation with respect to theoretical speculations, for their
contentis not wholly determinedby observationaldata. But the greater
portion of scientific knowledge, directly rooted as it is in empirical
evidence, is necessarilyindependentof the society or the specialised
group which first made it available. The social origin of scientific
knowledgeis almost completelyirrelevantto its content,for the latter is
determinedby the natureof the physical world itself.
THE SOCIOLOGY OF SCIENCE

Occasionallythis standardepistemologyhas beenmade fully explicit in
the course of sociological investigation of science. This is done most
clearly by De Ore, who statesin his introduction to the sociology of
science'that a real world exists independentlyof our knowledgeof it;
that this real world is to an extent knowable through a process of
approximation; and that knowledge is true to the degree to which it
approximatesor is isomorphicto the structureof reality' (1955, p. 37).
The centralimplication of theseassumptions,De Ore concludes,is that
sociologyshould be concerned,not with the actual cognitive contentof
science, not with certified knowledge as such, but with the social
conditionswhich make possiblethe attainmentof objective knowledge.
Although few sociologistsof sciencehave been as explicit as De Ore
about their philosophical presuppositions,they have all until recently
adopted his approach to the empirical study of science. The whole
tradition of sociological work on sciencewithin which De Ore writes,
beginningwith Merton'spioneeringresearchin the 1930sand continuing
off and on for thirty yearsor so, hassystematicallyavoidedexamination
of the substanceof scientific thought(Mulkay, 1969). It has offered us 'a
sociologythat deals with the allegedly fixed normativecommitmentsof
scientistsbut which paysscantattentionto the social significanceof their
patentlychangingcognitivecommitments'(King, 1971, p. 15). Although
this quotationconveyssomethingimportant about the main tradition in
the sociology of science, it would be wrong to infer from it that
sociologists have concerned themselves solely with investigating the
normative structureof science.Thus Merton, for instance, by far the
most influential figure in this field, has studiedthe allocationof rewards
in scienceas well asconsideringthe economic,technologicaland military
factors which facilitated the emergenceand growth of modern science.
Nevertheless,a consistent and continuing theme in the sociology of
sciencehas been the depiction of that 'complex of values and norms
which is held to be binding on the man of science' (Merton, 1973,
pp. 268-9), and which is seen as being crucially involved in the
generationof certified knowledge.It is in sociologists'accountsof the
supposed'ethos of science'that epistemologicalpresumptionsbecome
most evident and exert their clearest influence on the content of
sociologicalanalysis.
The natureof the 'scientific ethos' was first sketchedout by Merton,
as part of his thesisthat seventeenth-century Puritanismhad contributed
significantly to the birth of modern sciencein England. Merton argued
that the Puritancomplexof values led to the 'largely unwitting further-
ance of modern science' (1970, p. 136). Puritans emphasisedcultural
values such as utility, rationality, empiricism, individualism, anti-
traditionalismand this-worldly ascetism.This set of inter-relatedvalues
and norms was thought to parallel thosecharacteristicof science(1970,
p. 137). Consequently,the marked increasein 'scientific' activity which
occurredduring the seventeenthcenturycould be seen,at least partly, as
an unanticipatedconsequence of the increasingdominanceof the Puritan
movement. As a result of their religious values, those Puritans who
engagedin philosophical debate tended to concern themselveswith
empirical issuesand to addresstheseissuesin a rational, methodicaland
impersonalmanner. (Socio-economicfactors were seenas being impor-
tant in focusingintereston certainkinds of empirical problems.)The fact
that Puritans were over-representedamongstthe adherentsof the new
natural philosophy and among the founding members of the Royal
Society was strong evidence of there being a connection between
Puritanismand the establishmentof the modern scientific community.
At no point in his analysis does Merton try to establish a direct
connection between Puritan values and the intellectual products of
scientific endeavour.Indeedhe statesquite clearly in a subsequentpaper
that the substantivefindings of scienceare beyond the scope of his,
purely sociological,interests(1973, p. 268). His aim is, instead,the more
limited one of showing that thesevalues, in so far as they have become
institutionalisedin the scientific community, are essentialrequirements
for the regular production and acceptanceof properly confirmed and
logically consistentstatementsof empirical regularities.
The institutional goal of science is the extension of certified know-

ledge. The technical methodsemployed toward this end provide the
relevantdefinition of knowledge: empirically confirmed and logically
consistentstatementsof regularities(which are, in effect, predictions).
The institutional imperatives (mores) derive from the goal and the
methods.The entire structure of technical and moral norms imple-
ments the final objective. The technical norm of empirical evidence,
adequateand reliable, is a prerequisitefor sustainedtrue prediction;
the technicalnorm of logical consistency,a prerequisitefor systematic
and valid prediction. The mores of sciencepossessa methodologic
rationalebut they are binding, not only becausethey are procedurally
efficient, but becausethey are believedright and good. They are moral
as well as technicalprescriptions.Four setsof institutional imperatives
-universalism, communism, disinterestedness,organised skepti-
cism-aretaken to comprise the ethos of modern science. (Merton,
1973, p. 270)
Scientists' belief in the goodnessof these mores has its historical

origins in the religious commitmentsof the founding membersof their
professional community. But these mores are also methodologically
essentialfor the systematiccreation of valid knowledge. Thus, as the
scientific communityhasover the yearsseveredits links with the religious
sphere,scientistshave ceasedto justify their values in religious terms.
Instead, having come to recognisethe methodologicalimport of these
values, they have tended to justify them to the wider society as the
cultural basis of scientific truth and as the 'pure' sourceof practically
effective knowledge(Merton, 1970, p. xxii).
SinceMerton'soriginal formulation of the institutional imperativesof
sciencenumerousadditionshavebeensuggested:for example,norms of
originality, humility, independence,emotional neutrality and impartial-
ity have been proposed (Barber, 1952; Storer, 1966; Merton, 1973;
Mitroff, 1974). In addition, a few critical discussionshave appearedin
recent years (Lemaine and Matalon, 1969; Barnes and Dolby, 1970;
Weingart, 1974). But I do not intend to examineeitherthe supplementary
or the critical literature here. The point I wish to emphasiseis that this
portrayal of the ethos of scienceis powerful, and its power is demon-
stratedby its continuedvitality over more than three decades,becauseit
uses the standardview of scienceas a taken-for-grantedinterpretative
resource.For example, if the conclusionsof scienceare simply state-
ments of observableregularities,accuratewithin the technical limits of
the time, then it seemsto follow necessarilythat the particular personal
or socialcharacteristicsof thoseproposingsuchstatementsare irrelevant
to scientists' judgementof their validity. 'Objectivity precludesparti-
cularism'(Merton, 1973, p. 270). If knowledge-claimswerejudgedon the
basisof particularisticcriteria, then claims would be acceptedthat did not
correspondwith the objectiveworld. Giventhat valid scientific knowledge
is objective, it follows that scientists must regularly use impersonal,
universalisticcriteria in the courseof their professionalactivities. The
samekind of reasoningcan be appliedto other elementsof the scientific
ethos. Organisedscepticismand intellectual independenceare required
becausescientific knowledgemust not be taken on trust. All presupposi-
tions andall knowledge-claims,including one'sown, mustbe continually
scrutinisedfor logical consistencyandfor empiricalaccuracy.No person's
claims should be taken as valid becauseof his position in the scientific
community. If theseprescriptionswere ignored, inaccuratepropositions
would inevitably enter the corpus of certified knowledge. Similarly the
communalownershipof knowledgeis required. Without free and open
communicationof findings it would be impossiblefor scientiststo subject
all knowledge-claimsto the same critical appraisal or to apply their
universalisticcriteria of scientific adequacyconsistently.It is not being
proposed, of course, that scientists always conform in full to these
principles. Deviant actionssuchas fraud, secrecyand intellectual preju-
dice do occur. But, it is argued, they occur infrequently, for otherwise
naturalsciencewould not havethe validity which we know it to have. To
the extent that science does regularly produce valid and practically
effective knowledge, it appearsthat these principles must have been
operative.
Among the normative principles of which the 'scientific ethos' is
composed,the most important is that of universalism.This principle is
thoughtto be implementedin manydifferent ways in science.It seemsto
require,'for instance,that members'standingin the scientific community
should be basedon merit rather than any ascriptive criteria and that
scientific careersshouldbe open to all thosewith ability. But universal-
ism finds its most fundamentalexpressionin the assessment of the results
of scientific research. To say that scientists judge knowledge-claims
universalisticallyis not merelyto saythat scientists,like other specialists,
make use of technical criteria of adequacy;for the technical criteria of
different intellectualtraditions or different groups may be incompatible
and they mayor may not serveto establishobjective empirical regulari-
ties. Thus universalismmeansthat 'truth-claims, whatevertheir source,
areto be subjectedto pre-establishedimpersonalcriteria: consonantwith
observationand with previously confirmed knowledge' (Merton, 1973,
p. 270); or, to put it another way, universalismmeansthat the diverse
strandsin the developmentof scientific thought are guided by 'more or
less common criteria and rules of evidence which transcend other
differencesamongthe contendingintellectual traditions' (Merton, 1975,
p. 51). On the whole, the various other elements in the normative
structureare seen to contribute to the institutional goal of scienceby
ensuringthat thesepre-establishedcriteria of objectivity are rigorously
appliedto all knowledge-claimsbeforethe latter are acceptedas certified
knowledge.
This characterisationof the norms of science, intimately bound up
with the standardepistemology, has strongly influenced sociologists'
views of the overall structureof the scientific community as well as its
relations with the wider society (Mulkay, 1977a). Let me give one
exampleof how the standardepistemology,the establishedconceptionof
scientific norms and other aspectsof sociological analysis are inter-
related.The assumptionthat scientistsoperatewith universalisticcriteria
of adequacywhich are sociologically unproblematichas led to a mark-
edly functionalist interpretationof the systemof social stratification in
science (Cole and Cole, 1973). The decisive evidence on which this
interpretationis basedis that high rank in the researchcommunity is
empirically associatedwith the production of high-quality results and
low rank with the absenceof suchresults; and that, as long as 'quality of
results'is held constant,there seemsto be no strong connectionbetween
achieved rank and other social variables. Now, if we assume that
scientistsjudge their colleagues'work in terms of pre-establishedstable
criteria, we are inclined to see this evidenceas demonstratingthat the
scientific communityclosely resemblesa meritocracy.(A lot will depend,
of course,on the amount of variancein social rank which is accounted
for by variationsin the productionof high-quality work.) It will appear
that scientists undertakeresearchand pass on their findings to their
colleagues; that these findings are assessedon the basis of pre-
established,impersonalcriteria of adequacyand value; and that scien-
tists are rewarded with rank in proportion to their contribution to
knowledge. Thus scienceis one areaof social life where the functional
analysisof social stratification seemsto work; and it is made to work
becausesociologists'conceptionof scientific knowledgeallows them to
assume,with no need of empirical evidence on this point, that clear,
identical evaluativecriteria are availableto all members.
This interpretation,then, dependsindirectly on the standardview of
science.Once we begin to doubt whether knowledge-claimsare assessed
in this clear-cutfashionand to wonder whethertheir adequacyor quality
is perhapssocially negotiable,then it becomespossibleto seethe system
of social ranking ratherdifferently. For example,if the criteria employed
are themselvessocially createdand legitimated in the courseof the very
processof assessingknowledge-claims,the associationbetweenhigh rank
26 Scienceandthe SociologyofKnowledge
and the production of high-quality results could have an entirely
different sociologicalmeaning.For, having acknowledgedthat scientists
tend to establishtheir rank throughthe mediumof researchfindings, and
that therewill thereforebe an empirical associationbetweenquality and
rank, we would still have to ask: in what ways and by what meansare
adequacyand value attributed to specific results? Are there perhaps
systematicsocialdifferencesin participants'ability to establishthat their
work is of high quality? Thus, althoughit is clear that female scientists'
relatively low rank is associatedwith low quality of work (as recognised
by other, predominantlymale, scientists),we would no longer be forced
to seethis as the result of 'objective' differencesin the findings produced
by male and female researchers(Cole and Cole, 1973). It has become
possible to conceive that women and the members of other social
categoriesin scienceare systematicallypreventedfrom (or favoured in)
establishingthat their work is of high quality. By moving away from the
standardview and the associatednotion of 'universalism', that is, by
assumingthat cognitive criteria in science may be flexible and their
application to particular cases problematic, it becomes possible to
investigatewhetherthe social allocation of 'quality', and therebysocial
rank, is affectedby structural differenceswithin the scientific commun-
ity. In short, in so far as we modify the traditional epistemologicalbelief
that contributionsto knowledgecan be assessedobjectively and unam-
biguously, to that extent we are able to conceive of a wider range of
interpretativepossibilities, not only with respectto the social construc-
tion of scientific knowledgebut also in relation to social ranking, and
other social phenomena,in science. This will become clearer in later
chapters.
Clearly the major barrier preventingsociologistsfrom exploring these
possibilities is likely to be epistemologicalin character(Whitley, 1972).
One would not hesitate to consider such possibilities with respect to
'inferior' forms of knowledge. The difficulty with respectto scienceis
that we are departing from a well-entrenchedepistemology. We are
assumingthat scientists'accountsof the natural world are not to be
taken simply as reflectionsof an objective reality, nor as determinedby
invariant and transcendentrules of evidence.We are treating the notion
of 'consonancewith observation',for example,as being sociologically
problematic(Collins, 1975). It is not surprising, therefore,that sociolo-
gists did not ventureto posequestionsof the kind suggestedaboveuntil
the ground had bee·n preparedfor them by a series of debatesamong
philosophersand historians,in the courseof which the customaryview
of sciencewas seriouslychallenged.
2
Revisionsof the Standard View
In this chapter I intend to describe some recent contributions to

philosophy and historiographywhich have significant implications for
the sociologicalanalysisof science.I will makeno attemptto distinguish
the work of philosophersfrom that of historiansof science:most of the
thinkers whose work will be consideredhere bestridesuch disciplinary
boundariesand I will refer to them simply as 'philosophers'.Nor will I
try to providea systematicexpositionof eachmajor thinker'sideas.This
has beendone more than adequatelyelsewhere(Easlea,1973; Giddens,
1978),and I will assumethat my readeris not entirely unfamiliar with the
writings of such authors as Kuhn and Popper. Wherever possible,
therefore,I will draw upon the work of philosopherswhosewritings are
less well known, suchas Hanson,Ravetzand Hesse;partly becausethey
deserveto be more widely read, but also becauseI wish to show that the
changesin the philosophyof sciencewith which I shall be concernedare
not confined to one or two well-known but unrepresentativethinkers.
However, my argumentin this chapterwill not be organisedaroundthe
views of particularwriters nor in terms of the historical developmentof
ideasaboutscience.It will focus insteadupon the centralassumptionsof
the standardview, as they were outlined in Chapter1. I will take four of
its main contentionsin turn and examinehow far they needto be revised
in the light of this recentbody of literature.
THE UNIFORMITY OF NATURE

Both Mannheimand Stark begantheir attemptsto establishthe distinct-
ive characterof scientific knowledgewith a statementof what has been
called the 'principle of the uniformity of nature'. They maintainedthat
the phenomenaand relationshipsof the materialworld differ from those
of the social world in being invariant and stable. In their view this is
substantiatedby the fact that the basicconclusionsof physicalscience,its
laws of nature, are always and everywhere the same. Although the
principle of uniformity was regardedby thesesociologistsof knowledge
asa 'necessaryand crucial assumption'in any attemptto understandthe
nature of science,it has attractedlittle attention in the debateswhich
providethe materialfor this chapter.Fortunately,however,the principle
has beenexaminedrathercarefully by Hanson(1969). He showsthat its
usein the sociologyof knowledgehas beenbasedon a misunderstanding
and that it is a shaky foundation on which to build one's sociological
endeavours.
Hanson asks the question: how do we obtain the knowledge that
nature is regular, uniform and constant? How do we know that the
principle of uniformity is true? Thereareonly two ways in which its truth
could be established,he suggests:eitherby formal or by empirical means.
Clearly the principle of uniformity is not intendedas a mere formalism.
It is meant to convey something factual about the material world.
Consequently,the principle of uniformity must be establishedempiri-
cally. But if this is so, we are faced with a vicious circularity. For if the
principle is true, it must be presupposedin every empirical procedure-
including that by which we hopeto prove the truth of the principle itself.
If the principle is not assumed,we cannot establish its validity by
generalisingfrom empirical evidence.Unless the principle is assumedto
operate, there is no way in which we can infer it from particular
observations.
Hence, to gain knowledgeof the truth of theseprinciples [Hanson is

discussingthe principle of induction as well as the principle of unifor-
mity] by experimentand observationis to presupposein the searchthe
very existenceof that for which we are in search.If the principles are
true, we cannotlearn of it empirically, for the essenceof the principles
is that their truth is presupposedin every empirical enquiry. (1969,
p.408)
Hanson goes on to show with various specific examples that the

principle of uniformity is vacuousin the sensethat it makesno definite
assertion at all. Let me give just one of his illustrations. Consider
Newton'sinverse-squarelaw of gravitation and its derivative that freely
falling bodies acceleratetowards the Earth at 32.2 ft/sec. 2. Both these
propositionsmight be called 'laws of nature'.Hansonconsiderswhether
it makessenseto ask if theselaws could, under certain circumstances,be
different. For example, could the law of gravitation be different on
Mars, thereby overthrowing the principle of uniformity? Clearly the
formula expressingthe rate of free fall would be different on Mars,
becausethis formula is derived from the inverse-squarelaw in accord-
ancewith the specialconditionsassociatedwith the Earthalone.Thus the
principle of uniformity is in no way threatenedby a different rate of free
fall on Mars. The scope of this particular law of acceleration is
necessarilyrestrictedexclusively to the Earth. In this senseit is not a
fully-fledged law of nature and our questionis inappropriate.But what
of the law of gravitation proper?Could we say 'The law of gravitation
may take a different form on Mars' as we can say 'Freely falling bodies
accelerateat a different rate on Mars'? Hanson maintains, rightly I
Revisionsof the StandardView 29
think, that we could not. For if the 'law of gravitation' did not apply on
Mars, it could no longer be regardedas a law of nature. Any conclusive
observations,on Mars or elsewhere,which we believedto be inconsistent
with the law, would not reveal an area of non-uniformity; they would
show insteadthat our theory was wrong and our 'law' not a law at all.
'Thus, thata scientistexpresseshis law of gravitationin an identical form
on all occasionsproves nothing at all about the uniformity of nature.
What it may be said to prove is that we accordthe title "law of nature"
to nothing which is not expressedin an identical form on all occasions'
(1969, p. 353).
It seems,then, that modernphilosophicalanalysisrevealsthe principle
of uniformity to be no more than a rather misleading formulation of
what is meantby the term 'law of nature'.The principle of uniformity is
not an aspectof the natural world, but rather an aspect of scientists'
methodsfor constructingtheir accountsof that world. It cannotbe used,
therefore,asgroundsfor treatingthe generalisationsof naturalscienceas
definitive representationsof a stableand uniform physical reality.
FACT AND THEORY

Belief in the inherentstability and uniformity of the physical world has
often beenlinked to a particular view of the relationship betweenfact
and theory in science.From this orthodox position it is assumedthat
certain objects and processesexist in the physical world, that certain
events occur consistentlyand that certain stable relationships persist:
these objects, processes,events and relationships constitute the facts
which sciencehas to describeaccuratelyand explain convincingly. (For
the purposes of my argument, there is no need systematically to
distinguish particular facts and specific observationsfrom generalrela-
tionships betweenobservablephenomenaand the empirical generalisa-
tions which expresssuch relationships.)These facts are seen as being
theoretically neutral. They can, therefore, be expressedin a language
which is independentof theory and formulated in a way which simply
representsthe observablerealities of the physical world. Once firmly
established,facts remain unaffectedby interpretativeadvances.Indeed,
as long as there have beenno observationalerrors, they can undergono
change of content or meaning and they can be used, therefore, to
discriminateobjectively betweentheoreticalalternatives.
It is accepted,of course,that successfultheoriesusually generatenew
observationsand new facts about the physical world. However, once
such facts havebeenproperly confirmed, they are thought to acquirean
intellectual autonomywhich enablesthem to remain unaffectedeven by
the wholesalerefutationof the analyticalschemewhich gavethem birth.
. . .an experimental law, unlike a theoretical statement, invariably

possesses a determinateempirical contentwhich in principle can always
be controlledby observationalevidence... even when an experimental
law is explained by a given theory and is thus incorporatedinto the
framework of the latter'sideas... two characteristicscontinueto hold
for the law. It retainsa meaningthat can be formulatedindependently
of the theory, and it is based on observationalevidence that may
enablethe law to survive the eventual demiseof the theory. (Nagel,
1961, pp. 83-6)
This two-tier view of scientific knowledge, taken for granted by most

sociologistsof knowledge, has given rise to a number of intractable
philosophicaldifficulties. In seekingto resolvethesedifficulties, philos-
ophershavegraduallyevolveda new accountof the relationshipbetween
fact and theory which has major implications for the scopeof sociologi-
cal analysis.
It has sometimesbeenmaintainedthat when we conceiveof scientific
knowledgeas composedof separatetheoreticaland factual propositions
we are making use of a distinction between observableobjects and
unobservableor theoreticalobjects.Factualpropositions,it is suggested,
deal with the relations betweenobservableobjects, which are explained
by statementsinvolving such unobservableobjects as electrons,quarks
and genes.The nature of observableobjects is ascertainableby direct
experience and, assuming that proper experimental precautions are
taken, can be establishedwith great confidence. But the characterof
theoretical objects is known only indirectly and must, therefore, be
treatedas inherently speculative.Theoretical 'objects' are perhapsno
more thanconvenientfictions. Electrons,for example,areto be regarded
as hypothetical constructionsderived from our experienceswith real
objects like cathode ray tubes and galvanometers.Accordingly, the
notion of 'electron'can be abandonedif it proves unfruitful in the long
run or incompatiblewith newly observedfacts; and the propositionsin
which it features can be recast. In contrast, we can hardly deny the
existenceof suchobjectsas cathoderay tubesor galvanometers;and our
direct knowledgeof theseobjects,althoughit can always be mademore
precise,is not open to doubt or repudiation to anything like the same
extent.
Although this argumentmakesa plausibleappealto commonsense,it
is certainly incomplete.For instance,it deals only with objectsand fails
to cover the referentsof scientific conceptsconcernedwith relationships
or processes.But, more important, there are also strong reasonsfor
regardingthe distinction betweenobservableand unobservableobjectsas
untenable.The centralassumptionis that we can distinguishunambigu-
ously between observing an object directly and merely inferring its
characteristicsfrom its effects.But seeingan object directly is thought to
involve photonsmoving from the objectin questionandimpinging on the
Revisionsoj the StandardView 31
retina of the observer.As soonas we think in theseterms, the notion of
'direct observation'startsto loseits clarity and, thereby,its usefulnessin
separatingfact from theory. The vaguenessof the distinctionis also high-
lighted by consideringhow little is the differencebetweenseeingdirectly
and observingwith a magnifyingglass,and betweenthe latter and usinga
small telescope(Smart, 1973). Yet we arerequiredby this thesisto regard
the objects revealedby the telescopeas different in kind from those
visible to the naked eye. Furthermore,we cannot avoid the problem
simply by classifying 'seeingthrough a telescope'as direct observation.
For we are then faced with such anomalies as the failure of those
scholasticswho peered throughGalileo's telescopeto see the objects
which were so obvious to the disciple of Copernicus.We will return to
suchissuesin the next section.For the momentwe needonly note that no
clear distinction has been made between observableand unobservable
objects; and consequently,the distinction cannot be used to strengthen
the two-tier or standardview of scientific knowledge.
. . .if observabilityis merelya matterof degree,then thereseemsto be

no plausibleway of drawing a sharpline on this basisbetweenobjects
which do and objectswhich do not exist. Under the influenceof these
considerations,most philosophers have given up the attempt to
distinguish observablesfrom unobservableson this basis and focus
insteadon the terminologicaldistinction. (Grandy, 1973, p. 3)
Let us turn, therefore, to the questionof whether a difference can be

establishedbetweentheoreticaland observationalterms.
The debateabout the nature of scientific terms has been concerned
particularly with the meaning of theoretical concepts. The central
problem has arisen as a consequenceof taking factual statementsas
unproblematic(except in the 'trivial' senseof dependingon careful,
accurateobservation)and as conceptuallydistinct from thoseof theory.
Thus, if statementsof fact are independentof scientific theoriesand can
be usedas a neutralcheck upon them, and if theoreticalpropositionsgo
beyond the establishedfacts, what referencein the real world can be
attributedto theoreticalclaims and the terms in which they are formu-
lated? For instance, if the factual statementsof geneticistsabout the
colours of successivegenerationsof sweet peasare quite separatefrom
theoretical propositions about genes, about which we can obtain no
direct evidence,can the latter claims be said to have scientific meaning?
Furthermore,if the statementsin which terms like 'gene' occur have no
meaning, if they are not synthetic propositions,how can we maintain
that they are true or false? And if theoreticalclaims are neither true nor
false, they can hardly be regardedas furnishing valid knowledge.
Thereis someoverlapbetweenthis conceptionof the problemand that
which seeks to distinguish between kinds of scientific objects. This
second formulation, however, points in a more promising direction.
Whilst it appearsthat theoretical terms do not correspond toa special
type of entity and that theoreticalconceptsseemto havea different kind
of meaningfrom those used to report observations,it cannot be denied
that the useof theoreticaltermsis peculiarly characteristicof scienceand
that they must thereforebe meaningful in someway.
One responseto this problem is to acceptthat scientific theorieshave
no directmeaningand to regardthem simply as formal systems.Theoreti-
cal termscometo be seen,therefore,merely as logical devicesfor deriving
new observationalstatementsfrom establishedfacts and as acquiring
'indirect meaning' through their linkage with factual statements
(Carnap, 1939). But this answer raises further difficulties. In the first
place, it is difficult to reconcile the relatively trivial role assignedto
theory by this interpretation with the fact that theoretical work is
regardedas of fundamental.importanceby practisingresearchersand the
fact that theoristsreceive by far the greatesthonour and respectamong
scientists(Hagstrom, 1965). Thus this attempt to describethe structure
of scientific knowledge, although it succeedsin portraying its factual
basisin a way which supportsthe standardview of science,doesso at the
cost of treating its theoretical componentas largely redundantand of
implying that scientists themselvesare generally mistaken about the
relative value of the contributionsto knowledge made by theorists and
observers.Clearly an accountof scientific knowledgewhich made sense
of scientists'own high regard for theory would be preferable.
A move in this direction was madewith the recognition that it was in
practice extremely difficult to distinguish observationalterms, whose
meaning was 'derived from experience', from speculative theoretical
terms. For instance,thereseemslittle point in insisting that thereare two
meaningsof the term 'mass',one observable(as in the observedmassof
a volume of gas) and one theoretical (as in the mass of its constituent
molecules,which are individually unobservableyet which in sum deter-
mine the observablemass).Accordingly, it has increasinglycome to be
acceptedthat the distinction is arbitrary and as inherently ambiguousas
that betweenobservableand unobservableobjects (Carnap, 1966). One
major reasonwhy it has proved so difficult to separateobservational
from theoreticaltermsis that termsseemto acquiretheir meaning,not as
isolatedunits which in the caseof observationalterms can be referredto
correspondingphysical entities, but as elementswithin wider linguistic
frameworks.
It is not correct to speak,as is often done, of 'the experientialmean-

ing' of a term or a sentencein isolation. In the languageof science,and
for similar reasonseven in pre-scientific discourse,a single statement
usually has no experiential implications. A single sentence in a
scientific theory doesnot, as a rule, entail any observationsentences;
consequences assertingtheoccurrenceof certainobservablephenomena
can be derived from it only by conjoining it with a set of other, sub-
sidiary, hypotheses.Of the latter, some will usually be observation
sentences,others will be previously acceptedtheoretical statements.
Thus, for example,the relativistic theoryof the deflectionof light rays
in the gravitationalfield of the sun entails assertionsaboutobservable
phenomenaonly if it is conjoinedwith a considerablebody of astrono-
mical and optical theory as well as a large number of specific state-
ments about the instruments used in those observationsof solar
eclipseswhich serveto test the hypothesisin question.(Hempel, 1965,
p. 112)
If this is so, if observationalterms have no experientialmeaningapart

from their location in a broaderconceptualand propositionalscheme,it
is simply impossible to identify a separateclass of factual statements
constitutingthe bedrock on which scientific knowledgeis built.
The way in which particularobservationaltermsacquiretheir meaning
from a cluster of associatedpropositionsand conceptshas been made
especially clearby Hesse(1974). She begins with the assumptionthat
every physicalsituationis indefinitely complex. Each new situation is in
detail different from every other. This leadsto a loss of information in
everyapplicationof observationalterms, which leavesroom for changes
in classificationto take place under certain circumstances.Many of the
observationalcategoriesused in science,as well as in everydaylife, are
learned in concrete empirical contexts, where a direct associationis
establishedbetweenselectedaspectsof the situation and a certain term.
But completefluency in the use of descriptiveterms is not obtainedby
meansof direct associationalone. Learning a languagealso involves
learningcertaingeneralisationsor 'laws' containingits terms. Theselaws
are always linked together in symbolic networks. Knowledge of these
laws and networks is required in order to identify proper occasionsfor
the useof a physicalpredicate;such knowledgeenablesthe user to apply
the terms 'correctly' in situationsother than those in which they were
learnedinitially.
For example,a personmay learn to observeand identify the planet
Venus, partly by meansof the 'law' that 'starstwinkle in the night sky
but planetsdo not'. This law canthenbe usedas a resourcefor observing
other planetsand may leadto the 'correct' identification of Mars, and so
on. Of course, in applying this law, a newcomer may make what is
regardedby a moreexperiencedobserveras a mistake.He may think that
he is observinga planet, when what he is 'really' seeingis a star. If the
moreexperiencedobserverwishesto correctthis mistake,he may merely
suggestthat if his companionlooks more carefully he will see that the
celestialobject in questionis actually twinkling. But if his companionis
not convinced thus easily, it may be necessaryfor the experienced
observer to draw extensively upon his knowledge of astronomical
parallax, upon theoriesof gravitation and optical transmission,and so
on, in order to showthat the object simply cannotbe anotherplanet; for
it it were, the whole scienceof celestial mechanicswould be overturned.
The logic of this procedureis to take the current knowledge-systemand
its laws as given and explicitly to decidewhat it is we are observingin the
light of the requirementsof this system. Thus, the meaning of the
observationalterm 'planet' is derived from its use within a network of
related terms and propositions,and is not establishedsimply by direct
reference to a series of isolated empirical instances which can be
identified independentlyof this cluster of interpretative resources.As
Hesseherselfemphasises,this accountof scientific knowledgeinvolves a
far-reachingreinterpretationof the theory/observationdistinction.
Thereare considerableimplicationsfor the standardview of sciencein
this abandonmentof the orthodox distinction between theoretical and
observationalterms. In the first place, if all terms obtain their meaning
throughtheir location ina framework of conceptsand propositions,then
it seemsthat no statementof fact is theoretically neutral. Scientistsdo
not have accessto independentfindings againstwhich to check theoreti-
cal alternatives.They can neverstep entirely outsidetheir own analytical
scheme,for to do so would deprivetheir conceptsand their propositions
of meaning. Thus all empirical statementsare 'theory-laden' (Ryle,
1949). Furthermoreit follows that, in so far as the analytical framework
alters, so doesthe meaningof observationstatements(whetherparticular
findings or empirical generalisations)formulated within its frame of
reference.
. . . no feature in the total landscapeof functioning of a descriptive

predicateis exempt from modification under pressurefrom its sur-
roundings.That any empirical law may be abandonedin the face of
counter-examples is trite, but it becomeslesstrite when the functioning
of every predicateis found to dependessentiallyon somelaws or other
and when it is also the casethat any 'correct'situationof application-
eventhat in terms oj which the term was originally introduced-may
becomeincorrectin orderto preservea systemof laws and other appli-
cations. It is in this sensethat I shall understandthe 'theory depend-
ence' or 'theory ladenness'of all descriptivepredicates.(Hesse,1974,
p. 11)
We appear,then, to have reacheda conclusionwhich rejectstwo basic

assumptionsof the standardview; that is, we have concludedthat the
factual claims of scienceare neither independentof theory nor stablein
meaning. Even when the symbols on the page of a scientific textbook
remainunchangedover a fairly long period, their meaningin the eyesof
theresearchcommunitymay well bein continualflux, as the interpretative
context of researchevolves. In addition, it follows that the meaning of
given factual statementswill often differ for different sections of the
scientific community; for instance, for researchersas against school-
teachersand for membersof separatespecialties,dependingon how far
thesesocial groupingsoperatewith divergentinterpretativeframeworks.
Thus not only is the factual 'basis' of science theory-dependentand
revisablein meaning,but it also appearsto be socially variable.
This revised perspectiveon the relations betweenfact and theory has
severalstrong interpretations(e.g. Feyerabend,1975), as well as numer-
ous weakerversions(e.g. Scheffler, 1963). Before I discussin more detail
the characteristicsof these strong and weak versions, let me say a few
more words about the ways in which linguistic frameworks may affect
the formulation of factual propositions.There is, as I have stressed,an
influential empiricist tradition which treats facts as things or eventsout
thereto be observedand appropriatelydescribed.Thus facts are thought
to exist, eventhoughwe may not havewords with which to expressthem.
In order to appreciatethe limitations of this view it is useful to ask, as
Hanson(1969) does:what do the facts 'out there' look (or soundor feel)
like? ICould one photographa fact? lit is evident that a photograph, no
matterhow dearly focused,cannotpresentus with facts until we begin to
selectfrom it certainelementsand to formulate them in linguistic form.
(The argumenthere parallels that above taken from Hesse.)The facts
representedin the photograph are those features which can be so
expressed(Strawson,1959); and what canbe so expresseddependson the
linguistic and other symbolic resourcesavailable. In other words, the
nature of our languagewill tend to favour certain kinds of statement,
whilst prohibiting others.(Someof the work in cognitive anthropologyis
worth considering here. See Frake, 1961, and Hoijer, 1964.) This
argumentis as applicableto the vocabulariesof scienceas to any other
languagesystem.
. . . our typesof notationin physicsmay occasionallyrenderus insensi-

tive to featuresof the material world ... I am not sayingthat thereare
aspectsof ... protozoa,and subatomicentities that elude description
in the languagesavailableto us. My point is only that it is not logically
impossiblethat there might be. And if this point is sound we can see
that it is not logically impossible that we might have come to think
about the physical world very differently from the way we actually
think ... Given the samephysicalworld we might have(logically might
have)cometo speakof it differently ... In otherwords, the logical and
grammaticaltraits of our severalscientific languages,notations and
symbol-dustersmay affect how we seethe world, or what we under-
standto be the facts aboutthe world. (Hanson,1969, pp. 182-3)
Hansonstressesthat it is impossibleto demonstratethis thesis other

than logically. For it follows from the thesisitself that one cannotprove
it empirically by presentinga fact which is beyond the scope of our
linguistic resources.It should also be noted that no claim is being
advancedthat linguistic frameworks alone determine the content of
factual statements.As Scheffler points out, scientistsusing an identical
conceptualscheme can perfectly well formulate contradictory hypo-
theses.Nevertheless,it is clear that linguistic frameworksin scienceare
usuallydevisedin conjunctionwith substantivemodelsor interpretations
(Bohm, 1965); and that the latter do influence the content of factual
statementsin a fairly direct manner.
It follows that although... the samecategorizationallows the expres-

sion of alternative hypotheses,these hypotheseswill nevertheless
confer alternativemeaningson the categorizationin question.If I can
indeedformulate the denial of my hypothesisin terms of the categori-
zation by which it is expressed,I cannot accept such denial without
altering the very meaningof this categorization,for such acceptance
effectsa changein my languagesystem.(Scheffler,1967,p. 46)
Although linguistic frameworks are necessaryfor the statementof

facts, they give no hold on the external world until they are used to
formulate some positive accountof certain facets of that world. Only
when this has been done can the observerattribute significance to his
observationsand therebystatethem as meaningful propositions.This is
nicely illustrated by Darwin's failure to recognise that one of his
experimentshad revealed what would now be seen as an important
empirical regularity.
Having crossedsnapdragonsand producedhybrid varieties, Darwin

found what he called 'prepotency'-and what Mendel called 'domin-
ance'-in the first generationoffspring. What is more he obtained
both parental types in the second generation of hybrids, actually
countedthe numberof eachkind, and found 88 of the prepotenttype,
37 of the other. This result is not significantly different from a
Mendelian 3: 1 ratio, but Darwin did not know how to attribute
meaningto it. (Glass,1953, p. 152)
Mendel himselfwas ableto transformsimilar findings into a statementof

empirical regularity.. at least partly becausehe had clear expectationsin
mind which enabledhim to regardthe approximatenumbersoccurringin
actual experiments as crude expressionsof an idealised theoretical
relationship(Fisher, 1936). Like Darwin, however, other well-informed
and competentspecialists,such as Focke, Hoffman and Nageli, passed
over Mendel'sfindings without seeingthat they revealed'facts of nature'
not alreadywell established.This becameacceptedonly someforty years
later, when Mendel's observationshad becomeexplicable in terms of
subsequentdevelopmentsin the theory of chromosomesand particulate
inheritance.
So far in this section we have seen that the traditional distinctions
betweenobservableand unobservableentities, and betweenfactual and
theoreticalstatementsare impossible to maintain without considerable
qualification. The most we can do is to distinguish roughly between
propositionswhich are closer to particularevidenceand those which are
usedin a moregeneralsense.But all factual claims, whethervery specific
statementssuchas Mendel'sdescriptionof the resultsof particular crops
of peas,or relatively generalpropositions,such as those expressingthe
ratios of inheritedcharacteristics,are theory-laden.Technical terms like
'electron', 'quasar', 'gene', and the more commonsenseterms and
propositionswhich scientistsregularly use (Elliot, 1974), acquire their
scientific meaning from the linguistic, theoretical (and perhapssocial)
context in which they are embedded.The meaningof scientific observa-
tional categoriesand factual claims must be conceivedin terms of the
positionthey occupyin a theory. As Darwin succinctlyexpressedit: every
fact is a fact for or againsta theory (Hanson, 1969, pp. 216-17). This
interpretationof the relationship between fact and theory in science,
unlike the standardview, has the advantageof being consistentwith the
great importancethat scientiststhemselvesattachto theoreticalwork. It
also hasmajor implicationsfor the sociologyof knowledge,as Nagel has
noted.
The import of every observationstatementis thereforedeterminedby

sometheory that is acceptedby the investigator,so that the adequacy
of a theorycannotbe judged in the light of theory-neutralobservation
statements.Accordingly, if these claims are sound, they apparently
lead to a far-reaching'relativism of knowledge',to a scepticismcon-
cerning the possibility of achieving warranted knowledge of nature
that is much more radical than the relativism associatedwith the views
of Karl Mannheimand other sociologistsof knowledge.(Nagel et al.,
1971, p. 18)
It is hardly surprising, therefore, that this account of scientific know-

ledge, certainly in its strongerforms, generatesphilosophicaldifficulties
similar to those faced by Mannheim.
First, in someversions,scientific knowledgecomesclose to appearing
empirically vacuous. The idea that factual reports are formulated in
terms of their associatedtheoretical or metaphysical presuppositions
seemsto imply that eachtheory can only be testedin its own terms, yet
that within its own terms it is immune from refutation. Any particular
observationwill have no meaninguntil it has beentheoreticallyinterpre-
ted. However, the meaning supplied by the theory can only serve to
confirm that theory. It seemsonly too easy to dismiss any observation
which appearsto departfrom expectationby asserting,perhapsin a more
refined form, that: 'If it departsfrom expectation,then it is either an
incorrectobservationor it is an observationof the wrong type.' Within
this view of science,althoughan analyticalframeworkmay generatesuch
severeinternal inconsistenciesthat its adherentsadopt some alternative
scheme,it can never be unequivocally refuted by means of empirical
evidence(Kuhn, 1962; Lakatos,1970).
In addition, there is the related problem of 'incommensurability'.If
the meaningof scientific terms and propositionsdependson the whole
belief systemwithin which they are presented,then it is difficult to see
how any two theoriescan be regardedas rivals or their factual claims as
incompatible.Scientistsworking within different schemeswill be investi-
gating different worlds; and although their statementsmay sometimes
appearsuperficiallyto be identicalwhentakenin isolation, their meaning
within their divergentframesof referencewill not actually be the same.
Furthermore,genuinecommunicationwill be possibleonly amongthose
who share a common framework, within which given statementsare
understood in the same way. Kuhn (1970) has argued that these
interpretative frameworks are community-basedand that inability to
communicateeffectively is experiencedonly intermittently, whenevera
researchcommunity'S paradigm disintegrates under the pressure of
internally generatedinconsistencies.It is, however, just as plausible to
arguethat eachindividual scientistdevelopsa unique perspectiveon his
researcharea (Ravetz, 1971; Gilbert, 1976) and that, if the meaning of
particular assertionsdependson the whole framework of belief, each
scientist must be 'trappedin the web of his own meanings'(Scheffler,
1967, p. 46).
This kind of difficulty closely resemblesthat posed by Mannheim in
his attempt to devise a new epistemologyfor social knowledge. Like
Kuhn, Mannheim assumedthat agreementabout facts can usually be
reachedfairly easily by those operatingwithin a sharedframework of
meanings.But when divergent perspectivesare involved betweenwhich
somechoice must be made, there is in Mannheim'sview no alternative
but to try to translatethe 'opposing' perspectivesto a higher level of
commonmeaning,in such a way that the greatestfruitfulness in dealing
with empirical materialsis achieved.The new philosophyof sciencegoes
further than Mannheim,as Nagel notes, not only in arguing that man's
view of the physical as well as the social world dependson shared
meanings,but also in stressingthe difficulty of translating from one
networkof meaningsto anotherandof achievinga commonunderstand-
ing of what is to count as 'empirical material'.
This strong alternative to the standardview, although it has been
gaining ground since the 1950s, is still a minority position among
philosophers of science. Indeed, many of the latter have devoted
considerableeffort to amending, without entirely relinquishing, the
orthodox account of fact and theory in responseto the major points
raised by its more radical critics. It seems to me, however, that the
various moderate, compromiseviews which have resulted still differ
sufficiently from the standardview to warrant a reappraisalof the
sociology of scientific knowledge.
Certainof the ideasI have been discussinghave come to be generally
accepted,evenby thosewho do not wish to abandontotally the orthodox
position. In particular,it is accepted'that the senseand use of predicates
employedin the sciences,including those employedto report allegedly
observedmatters,is determinedby the generallaws and rules into which
these predicatesenter' (Nagel, 1971, pp. 19-20). It follows that all
factual statementsare corrigible in principle, that their connectionwith
the external world is problematicand mediatedby theoretical presup-
positions, and that their meaningis subject to changeas the analytical
contextitself develops.All this is accepted.What is challengedis that the
meaningof factual propositionsis 'determinedby the totality of laws and
rules of applicationbelongingto the corpus of assumptionsof a science
at a given time' (Nagel, 1971, p. 20). Scheffler makes the samepoint,
denying that meanings are 'so interlocked that a change in anyone
affectsall the rest within a given languagesystem'(1967, p. 59). In other
words, it is being suggestedthat the cohesion, the connectedness,of
scientific knowledge-systems hasbeenexaggerated;andalso that it is this
misapprehension which hasgiven rise to the philosophicallyunacceptable
assertion of incommensurabilityand of empirical vacuity. Once we
recognise, it is argued, that each conceptual framework does not
constitute a seamlessweb, then we can avoid being driven to these
indefensibleconclusions. Let me give two brief illustrations of what
philosopherslike Nagel and Schefflerhavein mind. I shall begin with an
exampletaken from the field of radio astronomy (Edge and Mulkay,
1976).
Pulsarsare relatively small celestial objects which emit a rapid and
regular radio pulse. They were discoveredquite unexpectedlyin 1967-8
(Woolgar, 1976a).The researchwhich led to their discoverybegansome
months earlier, with the aim of identifying quasars-objectswhich
resemblepulsars only in having very small diametersand in emitting
electro-magneticradiationat radio wavelengths.The basicidea on which
this researchproject dependedwas that radio waves from very small
sourcesfluctuated quickly and irregularly, unlike those from ordinary
radio galaxies. This 'scintillation' is interpretedas a distortion of the
radio waves coming from small sources as the waves pass through
'plasmaclouds' surroundingthe sun. It is clear that the observationsof
scintillating radio scources whichensuedwere theory-ladenin the sense
discussedabove. Meaning was assigned to them within a complex
network of assumptions;for example,assumptionsabout the natureof
raQio wavesand radio receivers,about the propertiesof 'small' celestial
objects, about the physics of the sun's atmosphere,and so on. But
pulsarswere first observedbecausethey departedfrom the observational
expectationsgeneratedby theseassumptions.For instance,the pulsesof
radio emissionwere remarkably regular and they scintillated at night,
when the effect of the sun on radio waves was minimal.
Oncetheseunusualfeatureshad beennoticed,the observersconcerned
explored several quite distinct interpretations of their fiHdings-
including the possibility that the pulseswere artificially createdby some
alien intelligence. But theseinterpretativeexplorationswere linked only
in the most indirect mannerto the original set of assumptionswhich had
given meaningto the searchfor quasars.In the radio astronomers'view,
the establishedideasabout radio waves, radio receivers,plasmaclouds,
and so on, werein no way threatenedby this unexpecteddiscovery. They
were sufficiently indeterminateto cope with the existenceof the unanti-
cipated objects, without requiring any obvious revision. In this sense,
they provided part of the backgroundof interpretativeresourceswithin
which pulsars had to be understood.However, these original assump-
tions were of no positiveusein accountingfor the unusualfeaturesof the
observedradio signals. And, in fact, the observerseventuallymadeuse
of a largely unnoticed paper on neutron stars which had been in the
literaturefor sometime, in order to provide a preliminary interpretation
which they regardedas suitablefor their findings.
This exampleseemsto show that a single group of scientistscan have
accessto various bodies of theoreticalresourceswhich are not closely
connected;and that factual statements,although undoubtedlytheory-
laden, do not derive their meaningexclusively from that framework of
assumptionswhich gave them birth. Similar conclusionsare reachedby
Nagel in the courseof an examinationof some of Newton's researches
into optics. Nagel notesthat Newton's observationalterms acquiredat
least part of their meaning from various laws in which they were
embeddedand which Newton took for granted. But, he argues, the
conceptionsusedby Newton in interpretinghis findings were in no way
assumedin the design and executionof his experiments.
[This] showsthat an experimentintendedto ascertainon what factors

the occurrenceof a certain phenomenondependscan be describedso
that the statementof the observationsmadeis neutralas betweenalter-
native theories which may be proposedto explain the phenomenon,
even though the descriptivestatementwill inde~d presupposevarious
theories, laws, and other backgroundinformation that are not in
disputein the given inquiry. (Nagel, 1971, p. 26)
Thus once we allow that the networks of assumptionsand concepts

usedin scientific researchare neithercompletelydeterminatein meaning
nor wholly unified, it becomespossibleto escapefrom the 'stultifying
circularity' accordingto which factual statementsare linked indissolubly
to a particular framework of preconceptions.It does seem possible for
scientists sometimesto devise independentchecks for specific factual
claims; althoughit is clear that this can only be done by treating other
setsof assumptionsas unproblematic,at leastfor the moment.Similarly,
it may well be that the problem of incommensurability has been
exaggerated.The examples given above, for instance, of Newtonian
optics and the interpretationof pulsars,seemto show that scientistscan
agree about the theoretical basis for a given set of empirical results,
without therebycommittingthemselvesnecessarilyto identical interpreta-
tions of those results. If this is so, then at least some degree of
communicationis possiblebetweenparticipantsoperatingwith divergent
analytical frameworks (Grandy, 1973). This is what Kuhn, in reflecting
on his critics, has come to call 'partial communication'(Lakatos and
Musgrave, 1970).
I do not intend to pursuetheseissuesfurther here. I simply want to
establishthat, in abandoningthe standardview, one doesnot necessarily
have to adopt an extreme interpretationof scientific thought as com-
posed of closed, self-perpetuatingmeaning-systems.Even the more
moderateanalysesof fact and theory that I havedescribedabovedepart
significantly from the standardview. Gone is the simple notion that
scienceis built upon a growing corpus of neutral facts. Gone also is the
idea that well-establishedfacts are unrevisableand that, consequently,
scientific knowledgeaccumulatesin a relatively straightforwardfashion.
Nevertheless,although the notion of 'fact' has been made more provi-
sional and althoughfacts must be seenin relation to specific intellectual
frameworks, it would be wrong to conceive of scientists generally as
treatingtheir observationalor theoreticalknowledgeas merely hypothe-
tical and in constantdanger of collapse. In fact, one important factor
contributing to the impressive intellectual development of modern
science hasbeenthe capacityof its adherentsto forget their background
assumptionsandto concentrateon using theseassumptionsto undertake
detailed empirical exploration. Kuhn is surely correct in stressingthat
modernscienceis unusuallyfree, comparedwith other areasof intellect-
ual endeavour, from debates about fundamentals. Most scientific
researchis carriedout in a contextin which a whole seriesof assumptions
are so firmly entrenchedthat their revision or refutation is virtually
unthinkable. For instance, in the discovery of quasars(not pulsars)
astronomerswere faced, implicitly, with a choicebetweeneither recogni-
sing the existenceof extraordinaryobjects 100 times more luminous than
previously known radio galaxies, for which no remotely satisfactory
theoreticalaccountwas available,or with revising views about 'redshifts'
and stellarspectrawhich providedthe basisfor whole areasof astronomy
and its techniques(Edge and Mulkay, 1976). It is clear that very few
astronomersseriously consideredthe latter course, even after years of
interpretativefailure.
As we have seen,it is only againstthe backgroundfurnished by such
well-entrenchedassumptionsthat scientistsare able to formulate mean-
ingful and detailedhypothesesabout the world. For eachresearcherand
for each research community, only a limited number of issues are
normally treatedas empirically open. We must not, however,fall into the
trap of trying to distinguish, by meansof inherent characteristics,those
analytical resourceswhich are entrenchedfrom thosewhich are taken as
contingent. To do this would be very similar to trying to distinguish
observationalstatementsunambiguouslyfrom theoreticalstatements.It
must not be forgotten that the meaningof a scientific proposition varies
with the intellectualcontext in which it is used. This is not merely to say
that somepropositionsbecomemore firmly entrenchedas a researcharea
matures,althoughthis doestend to happen.It is to suggest,in addition,
that specificscientific formulations can be employedin various different
ways at the samepoint in time. Thus Hansonwrites:
... law sentencesare used sometimesto expresscontingent proposi-

tions, sometimesrules, recommendations,prescriptions,regulations,
conventions, sometimes a priori propositions... and sometimes
formally analytic statements... Few have appreciatedthe variety of
usesto which law sentencescan be put at anyonetime, indeed even
in oneexperimentalreport. (1965,p. 98)
Hansongoeson to show in detail how an individual physicist can be seen

to employ the secondlaw of motion in a variety of ways as he moves
from settingup an experiment,to constructinga machine,to defining an
areaof application, and so on. His conclusion is not surprising, for it
follows directly from the idea that the meaningof a propositiondepends
on its connectionswith other formulations. It clearly allows for the
possibility of considerablevariation in the meaningof scientific proposi-
tions in relation to changesin social context, in so far as social and
intellectual context vary together.
In this sectionI have shown that the orthodoxaccountof the relation
betweenfact and theory, the heart of the standardview of science, is
untenable.In demonstratingthis I have usedimplicitly an interpretation
of scientific observationwhich is directly opposedto that of the standard
view. Let me try to show clearly that a correct descriptionof observation
in sciencelends support to the argumentabove.
OBSERVATION IN SCIENCE
From within the standardview of science,observationis thought to be
scientifically adequatein so far as such distorting influences as bias,
intellectual prejudice and emotional involvement have been removed.
Properscientific observationoccurswhen the observerallows himself to
be impressedby an objectivereality. This view is consistentwith the fact
that naturalphilosophers,at the birth of modernscience,eliminatedfrom
considerationsecondary,subjective qualities, such as taste, smell and
colour, and concentratedon 'objective, measurableattributes', such as
motion and magnitude(Burtt, 1924). Thus observationin sciencehas
been seen as a plain recording of the unembroideredevidenceof the
sensesand as being quite separatefrom the creation of meanings.Much
recentwork in the philosophyof sciencehasbeendevotedto formulating
an alternativeaccountof scientific observationwhich is more consistent
with the revisedconceptionof the natureof scientific theory. In devising
this account philosophershave drawn heavily on the conclusions of
experimentalpsychologistsabout human perceptionin general.
One fundamentalconclusionarising from work in psychologyis that
observationcan never be as passiveas the standardview requires. We
neversimply receiveand registerinputs from the externalworld. Instead
we act upon that world so as to createa series of discerniblebut ever-
changingcuesaboutits characteristics;and in the very act of perception,
the observerinterprets these cues in terms of the cultural resourceshe
brings to bear.For example,in artificial situationswhere we are required
to observesolely by meansof touch, we constructor infer the structure
and compositionof objectsplacedin our handsby performinga seriesof
operations,suchas pressing,turning and balancing.As a result of such
activemanipulation,we aregenerallyableto placeobjectsinto categories
and to producestatementsabout their overall shapeand generalattri-
buteswhich go well beyondour actualcontactwith them. When askedto
report what we have observed,we tend to describethe inferred structure
in terms of the conventionalcategoriesavailable to us. We do not, and
almost certainly could not, describeour complex series of hand move-
ments, nor the fragmentedtactile sensationsproducedby those move-
ments. Our piecemealexplorations of the objects occur only on the
fringes of awareness.We regard ourselves as having observed the
relatively invariant structure which we have inferred out of a limited
range of sensations.
The accountjust given of the senseof touch is perhapsnot especially
surprising. After all, human beings seldom observeby meansof touch
alone. But, more interestingly, our dominant visual senseappearsto
operatein a similar way. Vision alsoinvolves necessarilyan active role on
the part of the observer.It appearsthat what we seeis constructedout of
constantly changing sensationsproduced by a continuous series of
movementsof the eye and of the body (Bohm, 1965). For example,the
eyeball vibratesin such a way that the retinal imageis constantlyshifted
by a distanceroughly equalto that betweenadjacentcells on the retina of
the eye. Superimposedon this movement,the eye has a regular swing
which is followed everyso often by a suddenreturn approximatelyto the
original point of focus. It appearsthat movementof this kind is essential
to visual perception.For when experimentalarrangementsare devised
which cancelout the effects of the eye'smovement,subjectseventually
becomeunable to see at all, even though clear images of the external
world arestill projectedonto the retina. Thus, althoughwe normally see
a world filled by solid, permanentobjectsin three dimensions,we have
no continuousoptical registerof theseobjects. Our eyes actually record
an ever-changingsequenceof momentary,two-dimensionaland inverted
impressions,out of which we construct the stable visual entities of
everydayknowledge(Borgerand Seaborn,1966, p. 118).
Studiesof the kind carried out by experimentalpsychologistsseemto
showthat direct observation,whetherscientific or otherwise,involves us
in more than merely registering and reporting 'the unembroidered
evidenceof the senses'.The observerhas no alternativebut to embroider
the evidence of his senses,for he receives from them no stable nor
completerecordof objectsor processesin the physicalworld. Insteadhe
receives,as a consequence of his own action upon that world, a seriesof
continuallychangingcuesaboutits characteristics.With the help of these
cues, the observer is able to perform inferential work, often quite
complexevenin the everydayworld, which enableshim to decidewhat it
is that he has observed.
The perceivedpictureis thereforenot just an imageor reflectionof our

momentarysenseimpressions,but ratherit is the outcomeof an ever-
changingconstruction... Such a constructionfunctions in effect, as a
kind of 'hypothesis'compatiblewith the observedinvariant featuresof
the person'sover-all experiencewith the environment in question.
(Bohm, 1965, p. 203)
The sameconclusionis reachedby Bruner in the courseof his influential

studies of perception/observation.He states clearly that we must no
longer think of observationas providing us with a representationof the
real world.
What we generallymeanwhen we speakof representationor veridica-

lity is that perceptionis predictive in varying degrees.That is to say,
the objectthat we seecan also be felt and smelledand there will some-
how be a match or congruity betweenwhat we see, feel and smell ...
Or, in still different terms, the categoricalplacementof the object
leadsto appropriateconsequences in terms of later behaviourdirected
towardthe perceivedobject: It appearsas an apple,andindeedit keeps
the doctor away if consumedonce a day ... The meaningof a thing,
thus, is the placement of an object in a networkof hypotheticalinfer-
enceconcerningits other observableproperties,its effects, and so on
... veridicality is not so much a matter of representationas it is a
matterof what I shall call 'model building'. (1974, pp. 10-11)
In this quotation Bruner shifts (uneasily perhaps)betweena behav-

ioural and a linguistic definition of veridicality. In some contexts a
behaviouraldefinition seemsreasonablyappropriate;when one is con-
sidering perceptionby animals, for instance. In the case of animals,
observationleads to behaviourwithout the mediation of words. (There
are now, of course,a few experimentalapesof whom this is no longer
entirely true.) In human perception,however, the linguistic element is
prominent; at least partly becausehuman observationoften leads to
statementsabout what was observed,in accordancewith which action is
initiated. Scientific observation,in particular, is necessarilyexpressedin
terms of words or equivalent symbols. (I have stressedthe role of
linguistic categories,but for discussionof a 'visual language'in science
seeRudwick, 1976.)Scientific knowledgeconsist/)of propositionsabout
the world, formulated in the conventionalforms of a specific language
system. Observationwhich is not contained in some kind of research
report and not formulatedin terms of generalcategoriesis irrelevant to
the purposesof science(Ravetz, 1971). In contrasta representationpure
and simple, say, in the visual form of a photographor retinal image,
containsin itself no linguistic component.(The argumenthere parallels
that in the previoussection.)A retinal image makesno assertionsabout
the world. Yet what we see,what we observe,hasa crucial bearingon our
propositional knowledge. If seeing were a purely visual phenomenon,
unaffectedby the categoriesthat the observerhasat his disposal,nothing
that we saw with our eyeswould be relevantto what we know about the
world. 'So our visual sensationsmust be cast in the form of language
beforethey can even be consideredin terms of what we know to be true.
Until a visual sensationcan be so considered,it is not observation... '
(Hanson,1974, p. 127). Even if one wishes to claim that human beings
do sometimes register the external world without the mediation of
linguistic categories,even if one regards some human perception as
equivalentto that of other animals, it is difficult to maintain that this
kind of observationis relevantto science.Scientific knowledgedealswith
what Ravetz has called 'intellectually constructedclassesof things and
events' (1971), that is, with generalclassificationswhich are defined by
certain properties of their members.Particular things and events are
classified (conceptualised)on the basis of how far they display the
featuresof a particularclass.And the observationalreports which enter
the body of scientific knowledgearecouchedin termsof thesecategories.
Scientific observation,then, is fundamentallydependenton language.
For most purposeswe can think of observationas the act of locating
things and events by meansof categories;and such categoriesacquire
their meaningby implying that certainstatementswill be found to apply
to that which has beenobserved.Thus, if a physicist seesan object as an
X-ray tube, he thereby assumesthat specific propositionswill apply to
this object; for instance,that under certain conditions fluorescencewill
appeararoundthe anode at high voltages. Similarly, if an astronomer
identifies an object as a pulsar he assumesthat it will move round the
heavenson siderealtime, that the pulses of electro-magneticradiation
which it emits will be regular and of short duration, and so on. It is clear
that theseverbal assertionsare somehowimplied in the act of observa-
tion, for if they were found subsequentlynot to apply, it would follow
necessarilythat the observedobject was not really an X-ray tube or not
really a pulsar. This is what Bruner meanswhen he writes that observa-
tion involves placing an 'object' in a network of hypotheticalinference.
Thereis a clearcorrespondence herebetweenBruner'sconclusionandthe
generalpositionarguedin the previoussection.But the argumenthasnow
beenextendedexplicitly to cover observation.Observationinvolves the
applicationof categoriesto senseimpressions.Categories,however,as we
haveseen,only have meaningwithin a network of related conceptsand
propositions.Consequently,observationconsistsin the interpretationof
senseimpressionsin terms of a linguistic and theoretical framework.
This does not mean that scientists first obtain various kinds of
sensationsand then apply conceptsand interepretationsto them. Con-
cepts and theories are always present in the very act of scientific
observation(Harris, 1970). The famous passagefrom Duhem which
follows illustrateshow this is particularly clear in the advancedphysical
sciences,whereobservationis mediatedthrougha complexvocabularyof
symbolic resources.(It is appropriateto quote Duhem here, becausehe
hasgreatlyinfluencedsuchmodernthinkersas Hansonand Hesse.)
Go into this laboratory; draw near this table crowded with so much
apparatus:an electric battery, copper wire wrapped in silk, vessels
filled with mercury,coils, a small iron bar carrying a mirror. An obser-
ver plungesthe metallic stemof a rod, mountedwith rubber,into small
holes; the iron oscillatesand, by meansof the mirror tied to it, sendsa
beamof light over to a celluloid ruler, and the observerfollows the
movementof the light beam on it. There, no doubt, you have an
experiment; by means of the vibration of this spot of light, this
physicist minutely observesthe oscillations of the piece of iron. Ask
him now what he is doing. Is he going to answer: 'I am studying the
oscillationsof the piece of iron carrying this mirror?' No, he will tell
you that he is measuringthe electrical resistanceof a coil. If you are
astonishedand ask him what meaningthesewords have,andwhat rela-
tion they haveto the phenomenahe has perceivedand which you have
at the same time perceived, he will reply that your question would
requiresomevery long explanations,and he will recommendthat you
take a coursein electricity. (1962, p. 145)
It is clear that the trained observerand the untrained observerin this
instancedo not seethe samethings or the sameevents.
The interpretativeconstructionsthrough which we observethe world
generateexpectationsabout the cues to be received in various types of
physicalsetting. Theseexpectationsmake us either more or less sensitive
to different kinds of cues and they can be thought of as providing
interpretative sets which enable us to translate expected cues fairly
smoothly into firm observations(Bruner, 1974). Thus what we observe
dependsin large measureon what we 'know', and therefore expect,
aboutthe world aroundus. Observationis 'shot through with interpreta-
tion, expectation,and wish' (Scheffler, 1967, p. 22). The importanceof
this general point to scientific researchhas, of course, been argued
emphatically and documentedextensively by Kuhn (1962). Scientific
observationin Kuhn's view, far from being an unselectiveand unstruc-
turedencounterwith a seriesof unfamiliar flashes,soundsand bumps,is
a preciselycalculatedcreationof theseas flashes, ~ounds and bumpsof a
particularkind. Sciencediffers from commonsense knowledge,not in the
elimination of preconceptions,but in the precision with which some of
these preconceptions are formulated and the detail with which they are
used to guide observation. 'Sometimes,as in a wave-length measure-
ment, everythingbut the most esotericdetail of the result is known in
advance, and the typical latitude of expectation is only somewhat
greater'(Kuhn, 1962, p. 35).
A similar characterisationof this aspectof sciencewas developedsome
time ago by Duhem. Duhem pointed out that ordinary testimony, based
on the observationalproceduresof everydaylife, can achievea very high
level of reliability. 'In a certainstreetof the city and neara certain hour I
saw a white horse:that is what I affirm with certainty' (1962, p. 163). But
this comparativecertaintyis attainedonly by restricting one's reports to
relatively gross features of what was observed and by omitting the
complexdetail. In contrast,scientiststry to deal in a precisemannerwith
a complexity and minutiaeof detailswhich would defy description,if the
scientistdid not have at his servicethe clear and concisesymbolic means
of representationand measurementfurnished by mathematicaltheory
(Duhem, 1962, p. 164). Becausescientific researchis constructedand the
resultantobservationsare expressedin terms of precisesymbolic formu-
lations, scientistsare able to probethe physical world in intricate detail.
Without the resources of a mathematically formulated theoretical
language, an account of any routine experiment would fill a whole
volume with the most confused, the most involved and the least
comprehensiblerecital imaginable (see also Kuhn, 1963). Duhem's
generalconclusionwill surprisemany readerseven today, some seventy
yearsafter it was first written.
The uninitiated believe that the result of a scientific experiment is

distinguishedfrom ordinary observationby a higher degreeof cer-
tainty. They aremistaken,for the accountof an experimentin physics
does not have the immediatecertainty, relatively easy to check, that
ordinary, non-scientific testimony has. Though less certain than the
latter, physicalexperimentis aheadof it in the number and precision
of the details it causesus to know: therein lies its true and essential
superiority. (1962, p. 163)
This is a far cry from the traditional view that science has evolved
observationalprocedureswhich effectively eliminate uncertainty.
Scientific observations,then, are typically construedin terms of an
establishedandcomplicatedrepertoireof interpretativeformulations. In
order to obtain preciseobservationsand detailed distinctions, scientists
necessarilytake for granted a wide range of backgroundassumptions.
These are normally used as unproblematic resources for organising
observationand for giving it scientific meaning.As Kuhn puts it, most of
scientific researchconsists in fitting observationsinto the conceptual
boxes provided by professionaleducation.Yet Kuhn is surely wrong in
claiming that in the greatmajority of specific areasof study thereis only
oneclusterof concepts,only one paradigmor exemplar,available.As we
noted in the previous section, there is a tendency in the literature to
exaggeratethe degree of intellectual rigidity or cohesion in science
(Merton, 1975). It is, in fact, far from unusual to find that there are
several candidateschemesin use in any particular area and that the
researchscientist'scentraldilemmais preciselythat of choosingbetween
these schemes(Lakatos, 1970). Nevertheless,Kuhn is clearly right in
stressingthat scientific discoveryis often associatedwith the failure to
match actual observationswith symbolically generatedexpectations.
When this happens,when what we expect to observe patently fails to
occur, it seemslikely thatscientific observationcomesclosestto resembl-
ing the account embeddedin the standardview of science. In such
circumstances,to returnto Duhem'sexamplequotedabove, the scientist
may abandonthe notion that he is measuringthe electricalresistanceof a
coil and revert to describingin gross terms the oscillationsof a piece of
iron carrying a mirror.
. . .theseobservationalsituations have a point to them just because

they contrastwith our more usual casesof seeing. The languageof
shapes,color patches,oscillations, and pointer readingsis the lang-
uageappropriateto the unsettledexperimentalsituation, whereconfu-
sion and perhapseven conceptual muddlementdominate. And the
seeingthat figures in suchsituationsis the sort wherethe observerdoes
not know what he is seeing.He will not be satisfieduntil he doesknow,
until his observationscohereand are intelligible as againstthe general
backgroundof his alreadyacceptedand establishedknowledge.And it
is this latter kind of seeing that is the goal of observation.For it is
largely in terms of it, and seldom in terms of merely phenomenal
seeing,that new inquiry will proceed.(Hanson,1969, pp. 108-9)
It is important to stress, once more, that what Hanson here calls

'phenomenalseeing'is not a mereregisteringof the externalworld. Even
phenomenalobservation is an interpretative act, but one in which
generally availablecultural resourcesare utilised in place of specialised
technicalvocabulary.Thus anotherway of saying that the scientistdoes
not know what he is seeing is to say that the formulation of his
observationsin the crude terminology of everyday life is normally
deemed'to be inadequatein the social context of physical research.
However, such a formulation may well be entirely appropriatein the
optician's testing room, in the psychologist's laboratory, or when
theoreticallygeneratedexpectationsno longer appearto hold (see also
Chapter 4 below). It seems, therefore, that what is to count as an
accurateobservationof physical phenomenawill vary from one social
settingto another(Lewis, 1956, p. 52).
It appears,then, that observationof the physicalworld has at leastthe
following features,noneof which is consistentwith the standardview of
science.It is an active process,in the sensethat the observercreatesand
respondsto a dynamic sequenceof cues. It involves categorisation,in
that cues are used to place hypothesisedobjects and events in terms of
pre-establishedsets of concepts.It is inferential, that is, the observer
necessarilygeneralisesfrom a rangeof cueswhich is always 'incomplete'
in order to establishwhat it is he perceives.Observationis not separate
from interpretation; rather these are two facets of a single process.In
most scientific research,observationis intimately guided by and expres-
sed in terms of a complex repertoire of symbolic formulations. These
give rise to comparatively precise expectations,in relation to which
observationscan be assignedtheir scientific significance.The quantified
languageof scienceas well as its controlled experimental procedures
produce exceptionally detailed and diverse empirical evidence. But
neither the certainty of these observationsnor the correctnessof their
formulation is in any way guaranteed.Nor is thereany 'one right way' of
reportingthe resultsof a given observation.Judgementsof observational
adequacyseemto vary, like the meaningof propositions,accordingto
the interpretativeand social context.This themewill be taken up again in
the next section.
THE ASSESSMENTOF KNOWLEDGE-CLAIMS

The discussionso far of the principle of uniformity, the nature of
observation and the interdependenceof fact and theory has major
implications for every other componentof the standardview of science
50 Scienceandthe SociologyofKnowledge
(Easlea,1973). Once we have departedfrom the traditional interpreta-
tion of fact, theoryand observation,we must necessarilymove towardsa
new account of the processesof scientific discovery, the nature of
scientific consensus,the characterof scientific progress,and so on. I
shall not attempt to cover all of theseramifications in the rest of this
chapter. Instead, I shall concentrateon examining how far we can
continue to maintain that there exist common criteria and rules of
evidencefor assessingscientific knowledge-claims,which are applicable
irrespectiveof differencesin substantiveconcernor analyticalapproach.
This issueis worth exploring, not only becausethe assumptionthat there
are such criteria is an important part of the standardview, but also
becauseit will help me to indicate someof the sociological implications
of the newer philosophyof science.
In the philosophicaldebateaboutthe criteria usedto certify scientific
knowledge,most attention has been given to the principles involved in
validating theoreticalclaims. Popperis not unrepresentativein suggest-
ing that it is scientists'way of choosingbetweentheories which make
science rational (1963, p. 215). Despite philosophers' emphasis on
theory, it is clear that most practising scientistsare also required, and
probably more frequently, to make judgementsabout the accuracyor
reliability of particular observational reports. It may initially seem
surprising,therefore,that therehasbeenso little concernwith the criteria
by which specific data are judged, especially in view of the fact that
'agreementwith the data' has beenso frequently proposedas a funda-
mental principle for the assessment of theoreticalclaims (Frank, 1961).
However, one reasonfor concentratingrather more on theoreticalthan
on observational criteria is that observation is itself theory-laden.
Consequently,when general criteria are used to judge the accuracyof
particularobservationalreports,they are interpretedin terms of specific
theoreticalassumptionsand a specific analytical context.
Let us take, for instance,the criterion of 'replicability'. There can be
no doubt that scientists often use a notion of this kind in deciding
whether or not particular experimental claims should be accepted
(Ravetz, 1971). Empirical results which cannot be reproducedunder
specifiedconditionsareusually regardedas untrustworthy.However, the
applicationof sucha rule to specificinstancesis far from unproblematic;
for what countsas following a rule cannotbe ascertainedfrom inspection
of the rule itself (Wittgenstein,1953). What is to count as a 'replication'
dependson scientists'theoriesabout the phenomenaunder study and on
their view of the factorswhich may influencethe observationalsituation.
Consequently,as theoreticalframeworksevolve and experimentaltech-
niquesdevelop,so the way in which the generalcriterion of 'replicability'
is appliedin any given area necessarilyalters. This can be illustrated by
referring back to the earlier discussionof nineteenth-centurywork on
inheritance. Before the emergenceof Mendelian genetics, Darwin's
experimentalresults on snapdragonswere hardly better than random
numbers; and although there was no reason to doubt the reported
observationsof such a renownednaturalist (see below, however, vis-a-
vis Mendel), there was no independentway of checking their accuracy.
To have repeatedDarwin's experimentand obtaineddifferent numbers
would havehad no clearimplicationswith respectto the latter's findings.
For numbershaveto be interpretedbeforeonecanjudgewhat is to cou!1t
as an equivalent result. Thus, it was not until after the theoretical
developmentsat the end of the nineteenthcentury, and the accompany-
ing advancesin statisticaltechniquesof inference,that Darwin's findings
could be seen to be 'not significantly different from a Mendelian 3: 1
ratio'. The acceptanceof Mendel'swork by geneticists,and its theoreti-
cal interpretation, furnished them with criteria in terms of which
Darwin's findings could be seento be accurateand, indeed, to provide
further confirmation (or replication) of Mendel'sown results. Darwin's
observationalaccuracy(within acceptablelimits) was guaranteedby its
conformity with theoreticalexpectation.
The dependence of 'replicability' on theoreticalcontextis revealedeven
more clearly in the attemptby Fisher during the 1930s to 'reconstruct'
Mendel's original experimentsas exactly as possible. Although he was
able to confirm many of the latter's results, Fisher had some major
reservations.
A seriousandalmostinexplicablediscrepancyhas,however,appeared,
in that in one seriesof results the numbersobservedagreeexcellently
with the two to one ratio [this experimentrequireda 2: 1 and not a 3: 1
ratio], which Mendel himself expected,but differ significantly from
what shouldhavebeenexpectedhad his theory beencorrectedto allow
for the small size of his test progenies.To supposethat Mendel recog-
nized this theoretical complication, and adjusted the frequencies
supposedlyobservedto allow for it, would be to contravenethe weight
of the evidencesuppliedin detail by his paperas a whole. Although no
explanationcan be expectedto be satisfactory,it remainsa possibility
among others that Mendel was deceivedby someassistantwho knew
too well what was expected.This possibility is supportedby independ-
ent evidencethat the data of most, if not all, of the experimentshave
beenfalsified so as to agreeclosely with Mendel'sexpectations.(1936,
p.132)
Fisher'sremarks show clearly how the criteria for identifying accurate

observationshad changedas the corpusof geneticknowledgehad grown.
Although geneticistswere still working within a Mendelian framework,
they had developedcriteria for assessingexperimentalresults on small
populations(basedon statistical analysis of variation in small popula-
tions) which in retrospectmadesomeof Mendel's observationalclaims
appearextremelyunlikely. Accordingly Fishersuggeststhat someonehad
tamperedwith the resultsso as to ensurethat they satisfiedthe criteria of
accuracywhich would have been assumedto apply some seventy years
before.
This example also shows how an apparent failure to replicate in
accordancewith currentobservationalcriteria can be easily reinterpreted
within a given theoretical context in such a way that establishedviews
remain secure. By the time that Fisher's work was carried out, the
conclusionsdrawn from Mendel'sexperimentswere too well entrenched
to be easily challenged-evenby the demonstrationthat some of the
original resultswere suspect.What Fisher does, therefore,is to take the
Mendelianframework andits subsequentelaborationsfor grantedand to
use these interpetative resourcesin reconstructing Mendel's research
programme.He is disturbed to find that some of Mendel's supposed
findings appearto havebeenspurious.But Fisheris forced to acceptthis
conclusionbecause,in view of the current framework of analysis, he
knowswhat Mendel'sactual results must have been like. Consequently,
he is obliged to 'explain away' apparent observationswhich do not
conform to expectation;and he does this simply by assertingthat the
datamust havebeenwrongly recorded.Fisher'stentativeexplanationof
Mendel'ssuspiciouslypreciseresultsin terms of a hypotheticalassistant,
who knew too well what was expected,illustrates with what flexibility
scientistscan reinterpretobservationalmaterial in order to avoid incon-
sistencywith 'establishedknowledge'.
When we look at all closelyat the observationalcriterion of 'replicabil-
ity' we find that it is almost a mere formality. It has no contentuntil it is
put into practicein a particularscientific context. It can perhapsbest be
thought of as a recommendationthat scientistsought to try to produce
numerous'equivalent'observationsin connectionwith any given prob-
lem. But replication introducesno new criteria by which experimental
results can be judged. Nothing more is involved in replication than the
normal proceduresof interpreting findings in accordancewith other
evidenceandin the light of analyticalconceptions.If this illustration is at
all representative,it seems that the kind of observational criteria
discussedin the philosophicalliterature obtain their meaning from the
theoretical or interpretativecontext in which they are applied. Let us
turn, therefore,to examinethe principles by which theoriesare thought
to be validated.
Numerouscriteria appearto be cited by scientistsin assessingrelatively
generalisedor theoreticalknowledgeclaims; criteria such as agreement
with the evidence,simplicity, accuracy,scope,fruitfulness and elegance.
One of the main difficulties with theseprinciples is that they deal with
quite different dimensions. How, then, is it possible for scientists to
combinethem in a mannerwhich is not arbitrary and which providesthe
commoncriteria requiredby the standardview? For instance,how is the
requisiteof accuracyto be reconciledwith that of simplicity? How much
inaccuracywill we allow in order to achievea given level of simplicity,
and why? Such questionsare undoubtedly resolved in practice. Thus
'physicsis filled with laws that expressproportionality,such as Hooke's
law in elasticityor Ohm'slaw in electrodynamics.In all thesecases,there
is no doubt that a non-linear relationshipwould describethe facts in a
moreaccurateway, but one tries to get alongwith a linear law as much as
possible' (Frank, 1961, p. 14). In these instances,generalisationsare
statedin sucha way that accuracyand agreementwith the evidenceare to
someextentsacrificedin order to achievea convenientlevel of simplicity
andto facilitate certainkinds of mathematicalcomputation.But the bal-
ancebetweenthesedifferent dimensionswhich is achievedin any particu-
lar setof formulationsis clearlynot given by the physicalworld. It is rather
a conventionalarrangementwhich differs over time and from one group
of scientiststo another(Ravetz,1971; Lakatos,1976).This doesnot mean
that the waysin which criteria are combinedare entirely arbitrary. But it
doesmeanthat different procedures,and thereforedifferent knowledge-
claims, would be acceptablein different interpretativecontexts.
Therewas a time when, in physics,laws that could be expressedwith-

out using differential calculuswere preferred,and in the long struggle
betweenthe corpuscularand the wave theoriesof light, the argument
was rife that the corpusculartheory was mathematicallysimpler, while
the wavetheoryrequiredthe solution of boundaryproblemsof partial
differential equations,[deemedto be] a highly complex matter. We
note that evena purely mathematicalestimationof simplicity depends
upon the stateof culture of a certainperiod. (Frank, 1961, p. 14)
In short, the criteria used in evaluating theoretical claims, like those

applied to particular observationalreports, seemto vary in meaningin
accordancewith the context in which they are used. They cannot be
regarded,therefore,as providing a meansof assessingknowledge-claims
which is independentof specific analytical commitments.
This conclusionappliesjust as much to the criterion of 'consonance
with observation'as to the other more obviously conventionalcriteria.
There was a time when it was thought that scientific theories or
generalisationswereacceptedwhen they could be provedto be true, that
is, in agreementwith the observablefacts (Frank, 1961). Thus scientific
knowledge-claimswere seen as claims to have establishedthe truth
(Shepherdand Johnston, 1976). But this view has been shown to be
inadequatein the courseof the philosophicaldebateaboutinduction. It is
inadequatebecause,logically, it is impossibleto prove any general,law-
like proposition by means of evidence about particular instances.
Generalpropositionscannotbe inducedfrom a particularset of observa-
tions with certainty, becauseeach general proposition necessarilygoes
beyondthe casesalreadyobserved,to include othersabout which there is
at presentno evidence.In other words, the universal generalisationsof
sciencemust be seenas being imaginatively hypothesised,or extrapola-
ted, from essentiallyincomplete sets of observations.Moreover, such
universalpropositionscan never be subsequentlyproved, even by means
of the most extensiveand detailed series of successfulpredictions. For
there can be no complete guaranteethat the next test will not produce
such disagreementbetweenprediction and observationalresults that the
universality of the proposition will have to be abandoned.(This argu-
ment presupposes that observationsare not completelydeterminedby the
theory they are intendedto test.)
It has thereforebeen suggested,particularly by Popper (1959, 1963),
that the major criterion of theoreticaladequacyin scienceis and should
be the ability of a claim to withstand attempts at falsification. This
argument seems initially plausible because,in principle, although an
immense number of positive observationsis insufficient to prove a
generalisation,one negative observationseemsenough to disprove it.
However, this thesis would only hold in relation to isolated theoretical
propositionswhich could be comparedwith absolutely unproblematic
observations.As soon as we take into account that any theoretical
proposition is linked to and dependsupon others, that observationis
itself an interpretativeact and that sometheoreticalassumptionsmust be
madein order to establishthe meaningof an observation,the appealing
simplicity of the criterion of 'resistanceto falsification' is lost. We are
never in a position where we can measure an isolated and simple
theoreticalstatementagainst an unmediatednatural world. Instead we
formulate and comparecomplex networks of theoretical and observa-
tional statements,and we seek to establish as consistent an account
overall as possibleof the phenomenawith which we are concerned.We
never, therefore,proposea theory in such a way 'that Nature may shout
NO. Rather, we proposea maze of theories, and Nature may shout
INCONSiSTENT' (Lakatos, 1968, p. 162). The most that scientistscan
achieve,then, is a high level of consistencyamong the various compon-
ents of their analytical framework.
One clear implication of this philosophical debate is that scientific
knowledge is inherently inconclusive. For it is impossible to move
conclusivelyfrom reportsabout particular samplesof things and events
to statementsaboutgeneralclassesof things and events(Ravetz,1971). It
seemsto follow that, becauseformally valid demonstrationsyielding
certainty are unattainablein science, scientific knowledge-claimsare
assessednot for their truth but for their capacity to meet the require-
ments of a particular interpretativecontext. Such contexts in science
containat leastthe two following generalrequirements:first, consistency
with other knowledge-claims,which may vary in their level of generality
and in the degree to which they are firmly established; secondly,
conformity to conventional standardsof adequacy, for example, in
connection with quantitative precision, rigour of argument, range of
evidence, etc., which are regarded by the members of a research
'community' to be appropriate to a given class of problems. The
relevanceof prior knowledge-claimsto the assessmentof subsequent
claims has already been demonstratedsufficiently. We have also noted
examplesof variation in the conventionalcriteria of mathematicalrigour
and experimentalaccuracy. It is worth emphasisingthe limitations of
these two elementsas basesfor assessingnew knowledge-claims.Prior
claims can give us no certainty when we use them to screensubsequent
claims, becauseall scientific claims are inherently inconclusive.Criteria
of adequacyare no more certain, for they cannoteven be establishedby
normal scientific procedures,that is, by means of argument based on
controlled observation.This is partly becausethey are often difficult to
make explicit, being akin to the tacit knowledge which is passedon by
craftsmen through personal contact; a characteristic which helps to
explain why philosophershave not succeededin studying them in any
detail. It also meansthat they are not easily subject to critical public
evaluationvia the journals. Thus, whilst criteria of scientific adequacy
operateas resourcesfor jUdging knowledge-claims,their own 'adequacy'
can only be establishedin a most indirect and uncertainmanner(Ravetz,
1971 ).
It is important to note that an item can appear to satisfy current
criteria of adequacyas an observationalor theoreticalclaim and yet be
rejectedon groundsof inconsistencywith establishedknowledge.Consis-
tency of this latter kind is crucial in determining whether or not a
particular claim is rejected.Nevertheless,at least in mature fields, there
is always a close connectionbetweenthe body of establishedknowledge
and the criteria of adequacyused by its practitioners. The range and
complexity of such criteria operativein any field reflect the structureof
the interpretativeschemein use. Consequently,althoughit is possibleto
distinguish certain general classesof criteria, the membersof anyone
class are realised in different ways in each area of research. Ravetz
identifies two broad types of criteria of adequacy: those relating to
processesof inferenceand thoserelating to evidence.The latter are more
varied than the former:
... for they control not only the conditions of the production of data
and information but also the strength and fit of the evidence in its
particularcontext.It will frequently be necessaryfor someof the evid-
ence to be explained and defended explicitly; and these subsidiary
argumentsmust also meet criteria of adequacyappropriateto their
function. Thus the complexity of a solved problem is matchedby that
of the set of relevant criteria of adequacy;and that set will depend
closely on the field of inquiry. Hence it is impossibleto produce an
explicit list oj criteria oj adequacy applying to a wide class oj
problems.(Ravetz, 1971, pp. 154-5)
It follows from this analysisthat criteria of adequacywill changeand

evolve as the technical culture alters in which they are embedded.'A
scientific problem, unlike a textbook exercise,carries with it no guaran-
tee that there exists a "correct" solution against which those actually
achievedcan be tested' (Ravetz, 1971, p. 149). (Although in a situation
resemblingKuhnian normal science,the existenceof such solutions may
be assumedby participants.)Thus, any significant knowledge-claimis
likely to entail somerevision of criteria of adequacy,as well as involving
alteration to the current theoretical framework. Let me illustrate this
point with another example taken from radio astronomy (Edge and
Mulkay, 1976). When the first reports of celestial radio emissionwere
publishedin the 1930s, they were virtually ignored by physicistsand by
astronomers.There were sound reasonsfor this. In the first place there
was Planck's Law of black-body radiation, from which it could be
inferred that radio emissionfrom celestialobjects would not be detect-
able. Secondly, there was the apparentanomaly that the sun, despite
being a dominant sourceof electro-magneticradiation at visual wave-
lengths,was reportedto be a relatively weak sourceof such radiation at
radio wavelengths.Thirdly, the early findings in radio astronomyseemed
to show that therewereeffectively two separateuniverses:the traditional
one made up of optically observableobjects and a previously unsus-
pecteduniversecomposedof an almost completelydistinct collection of
objectsdetectable onlyby radio techniques.As a result of theseand other
apparentanomalies,optical astronomerson the whole were unwilling to
pay much attentionto the early knowledge-claimsof radio astronomers.
Theseclaims were too difficult to reconcilewith well-establishedoptical
knowledge. And in addition, they did not satisfy the standards of
precisionor rigour required by the community of optical astronomers.
Consequentlyit was possiblefor astronomersto dismiss thesestartling
assertions untilthe evidenceon which they were basedwas 'adequate',
that is, comparableto the results of optical work.
In the long run, of course,.optical astronomersdid change their
interpretationof manycelestialphenomenain responseto the findings of
radio astronomers.Theoristswerethe first to respondfavourably; partly
because,being lessinvolved in observation,they were less committed to
the criteria of adequacygoverningoptical observation.They were able to
recognise that this new kind of knowledge-claim had to be judged
initially by different, and lessdemanding,standards.Closecollaboration
in observational work, however, occurred very seldom until radio
astronomershad achievedstandardswhich approximatedto the estab-
lished criteria of optical adequacy.Theselatter criteria had beenevolved
within the long tradition of optical researchand were appropriatefor
current visual techniques. But they were quite inappropriate when
appliedto the crude initial observationsby radio methods,at least in the
view of radio astronomers.It is clear that the first knowledge-claimsto
be proposedby the latter were inconsistentwith both the established
astronomicalframework of interpretationand the associatedcriteria of
observationaladequacy.For their claims to be considered,not only was
it necessaryto accept that the body of optical knowledge might need
extensiverevision, but also that the very criteria by which an astronomi-
cal knowledge-claimcould be deemedworthy of investigationmight have
to be set aside.In casesof major claims like this, it is quite misleadingto
think in terms of assessmentby the application of clear and predeter-
mined criteria. Rather the knowledge-claim,if it signifies at all, implies
that the cultural resourcesavailable for purposes of assessmentare
themselvesin need of amendment.
Becausethe evaluationof scientific knowledge claims is highly com-
plex, involving a subtle balance of tacit criteria of adequacy with
estimatesof consistency,significanceand analyticalimport, it tendsto be
a somewhatprolongedprocess-except in caseswhere claims are univer-
sally ignored as trivial or unanimouslyrejectedas inconceivable.In the
course of this process,knowledge-claimsmove through a sequenceof
phases,within which their content and their meaning are continually
reinterpretedin accordancewith the demandsof different interpretative
and social contexts.
One social context through which almost all significant scientific
knowledge-claimspassis that of evaluationon behalf of a professional
journal. But publication in a journal in no way establishesthat a claim
has been acceptedby the scientific community. The assessmentcarried
out at this junctureis relatively routine. Seldom is any attempt made to
reproduceexperimentaldata; and complex argumentsare not usually
examinedin great detail. In short, the knowledge-claimis subject at this
stage to a fairly superficial and preliminary appraisal of adequacy,
consistencyand significance(Mulkay and Williams, 1971; Ravetz, 1971).
However, a much more stringent selectionprocedurecomesinto opera-
tion after the claim has becomepublicly available. For it is now treated
by other researchers as a symbolic resource.At this point, eachinterested
researcherasks, in effect: 'How far does this contribution help me to
produce "better" solutions to my own problem?' It seems that the
answer to this question is usually negative, for the great majority of
published reports receive little attention in the subsequentliterature.
Thus claims which are found wanting are seldom publicly refuted. They
are usually ignored instead.A relatively small proportion of knowledge-
claims, however, are noticed and used widely by other researchers.But
the elementswhich are regardedby othersas the most significant partsof
a claim are not necessarilythe sameas thoseintendedby its author; nor is
the claim necessarilyinterpretedin the sameway. As a knowledge-claim
becomesdivorced from its original context, it is subject to varying
reinterpretationsevenwhilst remainingwithin one areaof investigation.
When a solvedproblemhasbeenpresentedto the community,and new

work is done on its basis, then the objectsof investigationwill neces-
sarily change,sometimesonly slightly, but sometimesdrastically. In a
retrospect on the original problem, even after a brief period of
development,its argumentwill be seenas concerningobjectswhich no
longer exist. Thereis then the questionof whetherit can'betranslated
or recast so as to relate to the newer objects descendedfrom the
original ones,and still be an adequatefoundation for a conclusion.If
not, then the original conclusion is rejected as dealing with non-
objects, or as ascribing false propertiesto real objects. But if such a
translationor recastingis possible,then the original solved problem is
seento havecontainedsomeelementwhich is invariant with respectto
the changesin the objectsof investigation.(Ravetz,1971, p. 189)
It is this informal processof selectionthat directs the courseof scientific

developmentand that transforms those few formulations which with-
stand reinterpretation and which continue to generate 'successful'
conclusionsinto the major intellectualachievementsof science(Toulmin,
1961).
Those select few assertionswhich are found to be capableof wide-
ranging and fruitful elaborationare passedfrom one researcharea to
another and eventually reach the scientific textbook and the historical
record. Theseinfluential formulationsare undoubtedlypart of scientific
knowledge.There is, however, no clear distinction to be made between
thoseclaimswhich contributeto knowledgeand thosewhich do not, apart
from their interpretationand use in a particular context; for a claim
unambiguouslyrejected by one researchcommunity can be treated by
anotheras a valid and fundamentalassertion.Mendel'swork oncemore
standsas a well-known example.Furthermore,the meaningand content
of theseformulationsdoesnot remainstable.They undergowhat Ravetz
has called a 'processof standardization'(1971, pp. 199-208). Ravetz
arguesthat a claim mustbe standardisedif it is to becomea fairly general
componentof scientific knowledge.If an assertionremainsattachedto
its original problem, it will not outlive the solution of that problem. But
in so far as it is divorced from that problem, it must be statedin some
standardform which can be employedfor various new purposesand in
diversescientific contexts.In this process,however,someof the original
content is lost. Meaning is lost by translation in scienceas well as in
literature.Pointsof obscurityand conceptualdifficulties are overlooked.
The limitations of underlying assumptions are forgotten. And the
balanceand emphasisof the original formulation are alteredto meet the
needsof new areasof application. In addition, becausethe knowledge,
technical skills and standardsof adequacy of the various audiences
involved are likely to be quite diverse, the standardisedversion must be
considerablysimplified. This processreachesits climax in the passageof
certified knowledgefrom the researchcommunity to the school curricu-
lum and the popular text; but theseare merely the most obvious parts of
a continuoussequenceof reinterpretationandchangeof meaning(for an
empirical illustration, seeGilbert, 1976a).
. . . the contentof a standardizedfact may decay, almost without limit;

the degreeof sophisticationand of faithfulnessto its original which is
necessaryfor its adequateperformanceof its function will dependvery
stronglyon its use.Also, it can be seenthat a version of a standardized
fact which is goodenoughfor one function can be quite inadequatefor
another; and since any standardized fact performs a variety of
functions, it will naturally appearin a variety of versions. (Ravetz,
1971, p. 202)
The conclusionsof this section can be briefly summarisedas follows.

Scientific knowledge-claimsare not assessedby means of invariant,
universalcriteria. Although certain broad conceptionshave beenidenti-
fied in the philosophical literature as common bases for accepting or
rejecting claims, these conceptionsare necessarilyinterpretedby scien-
tists in terms of particular theoretical ideas and specific analytical
repertoires.The rules of evidence,criteria of consistency,and so on, in
scienceare not rigid. They are certainly flexible enoughto allow scientists
considerableleeway in interpreting evidence so as to support well-
entrenchedassumptions.Moreover, the standardsapplied in the selec-
tion of knowledge-claimsare not inherent in the phenomenaof the
physical world. All claims are judged partly according to conventional
criteria of adequacy,which vary over time and from one group or social
contextto another,and partly in terms of their consistencywith an ever-
changing interpretative framework. Any significant claim is likely to
entail somerevision of current criteria of adequacy,as well as implying
that the establishedcorpus of knowledge is inadequatein some way.
Consequently,the assessmentof such claims within a researchspecialty
tends to occur relatively slowly, and is often characterisedby marked
opposition, as members explore the implications of the claim. The
processof assessmentis, therefore, alsoa processof reinterpretation.
Similarly, reinterpretationor meaning-change occurs,in accordancewith
different criteria of adequacyand different analytical purposes,when-
everassertionsenteranotherresearchareaor movealtogetheroutsidethe
researchcommunity. In short, contrary to the standardview, it seems
that scientific knowledge is not stable in meaning, not independentof
social context and not certified by the application of generally agreed
proceduresof verification.
SOCIOLOGICAL IMPLICATIONS
In so far as the argumentspresentedin this chapter,and in the literature

on which it draws, are deemedto be valid, the epistemologicalbarrier to
a sociological analysis of scientific knowledge has been removed;
although many practical difficulties still remain. Sociologistsof know-
ledge treatedscienceas beyondthe scopeof their analysisbecauseit was
thought to be a special kind of knowledge. This it clearly is in some
respects,for instance,in the extent to which it has achievedintellectual
supremacyin the modern world. Thus when they stressedthat science
and the scientific community were special social phenomena,thinkers
like Durkheim and Mannheimwere undoubtedlycorrect. But their view
of scientific knowledge required them to place a strict limit on their
conceptionof the social characterof science.In the first place, they saw
that science played an important part in modern society becausethe
provision of objective knowledge had many social side-effects,not the
least of which was the displacementof religion noted by Durkheim.
Secondly,sciencewas sociologicallysignificant becausescientistsseemed
not only to havedeviseda researchmethodwhich progressivelyrevealed
the realities of the externalworld, but also to have evolved an appropri-
ate form of social organisationwhich kept their knowledgeexceptionally
free from distortion by social or personalinfluences. It was this social
aspectof sciencewhich particularly interestedMerton. But sociological
analysis had to stop there. Thus we find Mannheim eagerly accepting
that the findings of physicsare relative to the observer'sposition in time
and space,but drawing back from the possibility that they might in some
sensebe socially relative. Sociologicalanalysiscould not go this far, for
sciencewas not simply one social constructionamongothers;it was 'real,
certain, indubitableand demonstrableknowledge'of an objective world
(Popper, 1963, p. 93). Sociology could deal either with the social
conditionswhich helpedto reveal(or to hide) the objective world or with
the social consequencesof objective knowledge. Sociology could say
nothingaboutthe form or contentof scientific knowledgeitself, because
the conclusionsof sciencewere thought to be determinedby the physical
and not the social world.
I have tried to show in this chapterthat there are good grounds for
rejecting this portrayal of science.In particular, the central assumption
that scientific knowledge is based on a direct representationof the
physicalworld has been criticised from severaldirections. For instance,
factual statementshave been shown to dependon speculativeassump-
tions. Observationhas beenshownto be guided by linguistic categories.
And the acceptanceof knowledge-claims hasbeen shown to involve
indeterminateand variablecriteria. Scientific knowledge,then, necessar-
ily offers an account of the physical world which is mediatedthrough
availablecultural resources;and theseresourcesare in no way definitive.
The indeterminacyof scientific criteria, the inconclusivecharacterof the
generalknowledge-claimsof science,the dependenceof such claims on
the availablesymbolic resourcesall indicatethat the physical world could
be analysedperfectly adequatelyby meansof languageand presupposi-
tions quite different from those employed in the modern scientific
community. There is, therefore, nothing in the physical world which
uniquelydeterminesthe conclusionsof that community.It is, of course,
·self-evidentthat the external world exerts constrainton the conclusions
of science.But this constraintoperatesthrough the meaningscreatedby
scientistsin their attemptsto interpretthat world. Thesemeanings,as we
have seen, are inherently inconclusive, continually revised and partly
dependenton the social context in which interpretationoccurs. If this
view, central to the new philosophyof science,is accepted,there is no
alternativebut to regard the productsof scienceas social constructions
like all othercultural products.Accordingly, there seemsevery reasonto
explorehow far and in what ways scientific knowledgeis conditionedby
its social milieu, how change of meaning is brought about and how
knowledge is used as a cultural resource in various kinds of social
interaction.It is, no doubt, possiblethat this revised view is fundamen-
tally wrong and its conclusionsmerely a result of defectivephilosophical
analysis.If this is so, the attempt to explore the social constructionof
scientific knowledgewill probablyfounder. But not to makethe attempt,
at a time when the earlier set of philosophicalideasappearsclearly to be
inadequate,is an indefensibleposition.
In orderto strengthenthis conclusion,let me refer back briefly to part
of the discussionin the previous chapter. There I describedthe crucial
distinction made by many sociologistsof knowledge betweenscientific
and social/historicalthought. The latter was regardedas amenableto
sociological analysis becauseit had the following characteristics.All
social thought, it was suggested,is related to a particular social context
and undertakenfrom a particular, historically unique perspective.The
knowledge gained is restricted by, and formulated in terms of, the
observer'snecessarilylimited interpretativeframework. Another way of
putting this is that the answersobtaineddependon the questionsposed
and on the questioner'spresuppositions.Particular assertionscan only
be understoodin the light of these backgroundassumptionsand any
attemptto assesstheir validity must utilise specific rather than generally
applicablecriteria. Owing to the diversity and evolution of interpretative
frameworks, meaningsare 'made and ever re-made' in the course of
social life and past contributions are continually reinterpreted.These,
then, are the main characteristicsattributed by sociologists to social
constructions,that is, to mentalproductshaving their origin in the social
as opposedto the physicalworld. They were thoughtto indicatethat only
socially variable knowledge of this kind was open to sociological
analysis,becauseonly heredid participants'socially derived perspectives
appear to penetrate into the meaning and evaluation of particular
assertions. However, the similarity between this characterisationof
social thought and the newer philosophical account of scienceis very
obvious. If the revised view of' science is accepted, the basis for the
traditional distinction betweenscientific and social thoughtis eliminated,
as is the exclusion of scientific knowledge from sociological interpreta-
tion (Barnes, 1974). Of course, we would hardly expect any other
conclusion, for one of the central claims of the revised view is that
scientific assertionsare socially created and not directly given by the
physical world as was previously supposed.(It is not implied that there
are no differences at all between the analysis of the physical and the
social worlds; only that at this level of generality the distinction
traditionally used in the sociology of knowledge no longer seems to
apply.)
The implications of this conclusionare considerable,for it enablesus
to ask a much wider range of questionsthan before about the social
natureof science.We can ask, for example,to what extent presupposi-
tions which are widespreadin modern society have been implicit in
scientific researchand mouldedits findings. We canexamineexactly how
scientistsdecideon the adequacyand significanceof knowledge-claims,
and whethertheir assessments are as disinterestedas has been customar-
ily supposed.We can study how the meaning of scientific assertionsis
reinterpretedin different social situations and whether, for instance,
suchreinterpretationcan serveas a sourceof social power. In general,we
can assesshow far detailedsociologicalinvestigationof the social life of
scienceprovidessupport for the revised philosophicalview; and we can
explore what differencethe revised view makesto our understandingof
the social relationshipsinvolved in the creation of scientific knowledge.
Theseare someof the issueswhich have recentlybegun to appearworth
investigating. I will pursue them a little further in the chapterswhich
follow, thereby taking a few tentative steps towards including science
fully within the scopeof the sociology of knowledge.
3
Cultural Interpretation in Science
In this chapterI will discusssomerecentsociological work on the social

and cognitivedimensionsof the scientific researchcommunity. My aim is
to sketchthe outlines of an analysisof the social productionof scientific
knowledge which is basedon detailed empirical evidenceand which is
consistentwith the new philosophyof science.I will also try to show that
the analysis which begins to emerge re-establisheslinks between the
sociologyof scienceand the sociology of knowledge. This themewill be
taken further in the final chapter.I makeno claim to provide a complete
review of the relevantliteraturein thesetwo chapters.Rather,I will select
for close attention those contributionswhich seemto me to be particu-
larly important in connectionwith my own line of argument.
In Chapter I I argued that when sociologists take for granted the
standardview of science,they are thereby inclined towards a particular
portrayal of the normative structureof the scientific researchcommun-
ity. In the next sectionI will try to show that this classicportrayal of 'the
scientific ethos' is inadequateand that an alternativeinterpretationcan
be formulated which is more consistentwith the available evidenceand
which lends support to the accountof scientific knowledgegiven in the
last chapter. (Much of the discussion in this section is taken from
Mulkay, 1976a.)
THE SOCIAL RHETORIC OF SCIENCE

As I tried to show in the openingchapter,the entire structureof moral
and technicalnorms in sciencehas beenconceivedas implementingwhat
is taken to be the ultimate goal of science,namely, the establishmentof
objectiveknowledgeof the physicalworld. Given the customaryassump-
tion that the scientific community has achieved knowledge which is
closely and increasingly 'isomorphic to the structureof reality', it has
beenseento follow necessarilythat the membersof that community have
beenpredominantlyopen-minded,disinterested,impartial, independent,
self-critical, and so on, in their intellectual endeavours.Assuming also
that behaviourof this kind will not occur spontaneouslythroughoutan
intellectual community, it has seemedpreferable to regard these attri-
butesas characteristicsof the communityas such,that is, as norms which
define the social expectationsto which scientistsare generallyobliged to
conform in the courseof their professionalactivities. (Readersunfamil-
iar with thesenormsshouldconsultMitroff, 1974.)
In orderto show that conformity to thesenorms is an essentialfeature
of modern science,those presentingthis argumenttend to describethe
negativeconsequences of deviant acts. The central idea is that actions
which contravenesuch norms will clearly distort any resulting know-
ledge-claims.For instance,if scientistsbecometoo committed to their
own ideas, that is, if they fail to abide by the norm of emotional
neutrality, they will be unable to perceivewhen their ideas are inconsis-
tent with reliable evidence.Similarly, if scientistsadopt personal,non-
universalisticcriteriain assessingknowledge-claims,their judgementswill
be biased and will tend to diverge from the objective realities of the
physicalworld. At the sametime, if secrecyand intellectual theftwere to
exist to any extent in scienceand the norm of communalityceasedto be
an effective guide for social action, then it seemslikely that the smooth
and impartial extensionof certified knowledgewould be disrupted. As
long as one remains within the epistemological framework of the
standardview of science,it is not difficult to find reasonswhy departure
from any of the normativeprinciples listed above will tend to interfere
with the creation of valid knowledgeabout the empirical world. Thus
from this perspectivethe normativestructureis the crucial featureof the
scientific community. The norms of scienceare seenas prescribingthat
scientistsshouldbe detached,uncommitted,impersonal,self-critical and
open-mindedin their attemptsto gatherand interpret objectiveevidence
about the natural world. It is assumedthat considerableconformity to
thesenorms is maintained;and the institutionalisationof thesenorms is
seen as accounting for that rapid accumulationof reliable knowledge
which has been the unique achievementof the modern scientific com-
munity.
The standardphilosophical view of science favours the assumption
that once certain major sourcesof distortion have been removed, it is
fairly easyto recogniseby meansof systematicobservationthe empirical
regularities of the external world. Accordingly, most of the normative
principles postulatedby sociologistshave been conceivedas minimising
the impact of potentialsourcesof distortion. This is one reasonwhy the
expositionof thesenorms tendsto take the negativeform outlined in the
preceding paragraph; the necessity of conformity to the norms is
demonstratedby showing that if scientistsare not disinterested,humble,
emotionallyneutral,intellectuallyindependent,etc., their perceptionand
their judgementof reality will suffer. Thus the normative structureof
scienceis seenas ensuring,as far as is humanlypossible,that the external
world is allowed 'to speakfor itself'. The newer philosophyof science,
however, provides much weaker grounds for inferring that scientists
must abideby this set of norms. From the newer philosophicalperspect-
ive the physicalworld is not so much revealedas socially andintellectually
Cultural Interpretationin Science 65
constructed.Take the principle of emotional neutrality, for example.
Recentphilosophicalanalysisstresses,not simply that completeneutral-
ity is impossible,but that considerablecOr.1mitmentis necessarybefore
even the simplest kind of observationalwork can begin. Moreover, as
researchersprobetheir chosenphenomenain greaterdepth and detail, so
their reliance on unexaminedassumptionsis likely to increase.At the
same time, it follows from the newer philosophy of science that
normative principles will acquire some of their meaning from the
intellectualcontext in which they are implemented,that is, in part from
members'specific scientific commitments.What is to count as emotional
neutrality, impartiality or disinterestedness may well vary in line with
scientists' researchskills and interpretative frameworks. For example,
scientist A may object that scientist B did not act in a disinterested
fashion when he failed to recogniseor cite A's work. But B may respond
by claiming that he ignored that work, not out of any intent to gain an
advantageover A, but becauseit was fundamentallymisconceivedand
could only serve to confuse and mislead other researchersless well
informed than himself.
To someunknown degree,therefore, the meaningto participantsof
thesenormativeprinciples may dependon intellectualcommitmentsand
may be socially variable within science.This is likely to be particularly
true of the principle of universalism, so central to the established
sociological interpetation,accordingto which scientistsare expectedto
judge knowledge-claims by means of impersonal, pre-established
criteria. For if, as appearsoften to be the case, available criteria are
unclear, or not easily applicable to particular instances,or different
personsare using different criteria, it seems impossiblefor scientiststo
employ this principle in practice; even though, once reasonableconsen-
sus has beenachieved,scientistsmay be able to formulate ex postfacto
the criteria which they have finally agreedare appropriateto a particular
body of knowledge.In other words, the sociologicalnotion of 'universal-
ism' presupposes that technicalcriteria are generallyavailablein science,
in sucha way that firm, impersonaljudgementscanbe madewith respect
to most knowledge-claimsand, thereby,with respectto the rewardsand
facilities which scientists deserve. The newer philosophy of science,
however, byemphasisingthat the establishmentof scientific knowledge
is a creativeprocessin which prior standardsare frequently modified and
new social meaningscreated,makesthis presuppositionquestionable.
The generalpoint that I wish to makeis that the revised philosophical
accountof science,unlike the standardview, doesnot lead obviously and
directly to the customarycharacterisationof the scientific ethosaccepted
for many yearsby sociologists.From the newer philosophicalperspective
thereis no value in identifying a set of social norms which is designedto
minimise distortion; partly becausescience is not conceived as an
enterpriseconcernedwith the definitive representationof reality, but also
becauseit would be assumedthat the meaning of normative principles
would vary in accordancewith changesin interpretativecontext. Accord-
ingly, if we wishedto formulate a set of scientific norms within the new
framework of philosophicalassumptions,these norms would undoubt-
edly be significantly different in form and contentfrom thoseusedin the
past. I do not, however,intendto do this. For thereare good groundsfor
revising much more radically the whole idea of normative regulationin
science.By following this line of thought we will arrive at an accountof
the cultural resourcesof sciencewhich is not only consistentwith the
philosophicalposition outlined above, but which will also help to bring
sciencefirmly within the scopeof the sociology of knowledge.
In recentyears,therehasbeensomecriticism of the original analysisof
the scientific ethos.One reasonfor suchcriticism is that detailedstudy by
historiansand sociologistshas shown that in practice scientistsdeviate
from some at least of thesesupposednorms with a frequency which is
remarkableif we presume that the latter are firmly institutionalised.
Another reasonis that none of the empirical studiesdesignedto discover
how far samplesof scientistsexpressagreementwith verbal formulations
of the norms has producedevidenceof any strong generalcommitment
(Mulkay, 1969; but seealso Storer, 1973). One responseto findings such
as these is to argue that the central normative element in science is
furnished, not by this set of social norms, but by the scientific frame-
works and technicalproceduresin terms of which the researchcommun-
ity is internally differentiated.But this is not the only possibleresponse.
For we can argue that the original set of social norms was not so much
wrong as incomplete.Merton, for example,has tried to account for the
very considerabledeviation from thesenorms by introducing thenotion
of 'counter-norm'(1973, ch. 18). Sciencehe suggests,like other social
institutions,doesnot employ a single set of compatiblenorms, but rather
a seriesof conflicting pairs of norms. This responsehas been explored
much more fully by Mitroff (1974), in the courseof a detailed study of
'moon scientists'.
One of the important merits of Mitroff's study is that it provides a
great deal of first-hand empirical material. In particular, it contains a
large number of quotationsfrom practising scientists.This means, not
only that his own argumentis exceptionallywell documented,but also
that it is possibleto an unusualdegreefor the readerto extendMitroff's
own interpretation of his data. Mitroff shows, first of all, that the
scientistsin his sampledo sometimesuse variantsof the norms described
above, as standardsfor judging the actions of their fellows and as
prescriptions about how researchersought to behave. But the over-
whelming import of his evidence is that, in addition, there exists in
sciencean exactly oppositeset of formulations and that conformity to
thesealternativeformulations can also be interpreted,by participantsas
well as by observers,as being essentialto the furtheranceof science.
Let me give just a few examples.
Mitroff suggeststhat the norm of emotionalneutrality is counteredby
a norm of emotionalcommitment.Thus manyof the scientistsstudiedby
him said that strong, even 'unreasonable',commitment to one's ideas
was necessaryin science,becausewithout it researcherswould be unable
to bring to fruition lengthy and laborious projects or to withstand the
disappointmentswhich inevitably attendthe explorationof a recalcitrant
empirical world. Similarly, the norm of universalism appearsto be
balancedby a norm of particularism. Scientistsfrequently regard it as
perfectly acceptableto judge knowledge-claimson the basis of personal
criteria. Insteadof subjectingall researchreports in their topic area to
impersonalscrutiny, scientistsregularly select out of the literature the
findings of thosecolleagueswhosework, for one reasonor another,they
havecometo regardas reliable. In other words, scientistsoften regardit
as properto judgethe man ratherthan the knowledge-claim.Onceagain,
the counter-normcan be seenas functional, on the groundsthat it saves
researchers'time and effort, speeds up the rate at which research
develops,yet at the same time ensuresthat greater weight is given in
generalto the judgementof those scientistswho are perceivedby their
colleaguesto be 'more able' or 'more experienced'.
Let me give onemoreexample.Mitroff producesevidenceto showthat
the ideal of common ownershipof knowledgeis balancedby a norm in
favour of secrecy.He also suggeststhat secrecy,far from hindering the
advanceof science,actually contributesto this objectivein severalways.
In the first place, by keeping their results secret,researchersare able to
avoid disruptive priority disputes.Secondly,attemptsby others to steal
or appropriatea scientist'swork serveto confirm the significanceof that
work and to motivate him to continue his efforts. Thirdly, by keeping
their findings from others, scientists are able to make sure that their
resultsare reliable without jeopardisingtheir own priority and, thereby,
without undermining their enthusiasmfor further research. Various
supplementaryarguments of this kind supporting each 'norm' are
available to scientistsas well as to the sociological analyst.
Mitroff's centralargument,then, is that there is not one set of norms
in science but at least two sets. The first set has been more or less
accuratelyidentified by thoseworking in the functionalist tradition. But
to describe the ethos of science in terms of this first set alone is to
producean accountof sciencewhich is quite misleading;for each of the
initial set of norms is matchedby an opposingprinciple which justifies
and prescribesactions in completeopposition. Thus common accessto
informationis not an unrestrictedideal in science;it is balancedby rules
in favour of secrecy. Intellectual detachment is often regarded as
important by scientists; but no more so than strong commitment.
Rational reflection is seenas essential;but so are irrationality and free-
rangingimagination;and so on. In Mitroff's view, the empirical evidence
which he has producedrequiresus to conceiveof the scientific commun-
ity as governedby these two major sets of norms and to interpret the
dynamicsof this community in terms of the complex interplay between
thesenormative structures.
Therecan be little doubt that this evidencepreventsus from accepting
the initial set of norms alone as the normative structureof science(see
also Blissett, 1972). However, I wish to argue that there are no
compelling reasonsfor regarding either set of formulations or the two
setscombinedas providing the rules governingsocial life in: science.This
becomesclear as soon as we look more closely at the kind of evidence
which is being adduced. Mitroff rightly criticises the procedure of
extracting norms of sciencefrom the 'highly select writings of the rare,
great scientists'. He suggeststhat we should derive 'the institutional
normsof science'not only from the idealisedattitudesof great scientists,
but also from the messy behaviour and complicated attitudes found
throughout the scientific community at large (1974, p. 15). He then
proceedsto formulate his counter-normsby selectingout certain descrip-
tive and prescriptivestatementsmade by participantswhich appear to
contradictthe original set of norms. It is clear, therefore,that both sets
of formulations are usedby scientiststo describeand to judge their own
actionsand thoseof their colleaguesand to prescribecorrect professional
behaviour.But the mereuse by participantsof theseverbal formulations
does not demonstratethat they are the 'institutional norms' of science.
Social norms can be regardedas institutionalised when deviance is
penalisedand when conformity is regularly rewarded.Clearly the kind of
analysis we have been consideringso far assumesthat norms and/or
counter-normsare institutionalisedin this sense;for otherwiseit would
be difficult to seethem as making essentialcontributionsto the extension
of certified knowledgeand to the progressof science.It is assumed,as
Storer puts it, that thereis an efficient 'internal police system'in science
(1966, p. 85). However, when we examinethe considerableliterature on
the allocationof professionalrewardsand the dynamicsof social control
in science, we find little indication that receipt of such rewards is in
practice conditional on scientists' having conformed to the supposed
norms or to the putativecounter-normsin the courseof their research.
The allocation of institutional rewardsin scienceis closely associated
with the system of formal communication. Scientists convey to their
colleaguesinformation which they believe to be interesting and reliable
by meansof the professionaljournals. Although thereis also a consider-
able informal exchangeof information, scientistsare able to establisha
conclusive claim to the credit for a particular contribution only by
publishing it formally under their own names(Merton, 1973). In return
for information which is judged to be of value, scientists receive
professionalrecognition in various forms and are therebyable to build
up a personalreputation,which canin turn be usedto obtain other scarce
Cultural Interpretation in Science 69
resourcessuch as students, research funds and academic promotion
(Gilbert, 1977a). Perhapsthe most important feature of this system, in
the presentcontext, is that the main medium of formal communication,
the research paper, is written in a strict conventional style which
concentratesattentionon technicalissues.Accordingly, referencesto the
opinions, interestsor characterof the author are rigorously excluded.
The report is typically written in the passive, so that allusions to the
actionsandchoicesof the authordo not occur. The effect of such devices
is to produce an aura of anonymity, so that the research becomes
'anyone's'research(Gilbert, 1976a).
Thereare, then, well-establishednorms governing the style of formal
communicationin science. But these must not be mistaken for norms
which regulate thesocial dynamicsof researchin general.As Medawar
has pointed out, the impersonalconventionsof the researchpaper not
only 'concealbut actively misrepresent'(1969, p. 169) the complex and
diverseprocessesinvolved in the productionand legitimation of scientific
findings. This divergencebetweenthe formal proceduresof communica-
tion and the actual social relationshipsinvolved in researchexists partly
becausethe rules governing the formulation of researchreports make it
virtually impossiblefor scientiststo passmoral judgements,on the basis
of published findings alone, about the author of a report. Thus their
responseto a publishedreport and their allocation of recognition to its
author, in the absenceof other information aboutthe author, cannotbe
influencedby the latter's conformity whilst carrying out his researchto
any particularset of social norms. In short, the conventionalform of the
scientific paper, which eliminates any referenceto authors' conduct in
carrying out their research,actually works to prevent scientists from
allocating rewards in accordancewith contributors' conformity to or
departurefrom any particular code of ethics. Of course, as Mitroff
demonstrates,scientistswho are in informal contact do regularly make
moraljudgementsof their fellows alongat leasttwo major dimensions.As
we shall seelater, theseinformal judgementsmay well affect the way in
which scientistsrespondto others'resultsand may accordinglyinfluence
the allocation of rewards. However, in the course of informal inter-
action, participantscan choosefreely from either set of norms. Thereis,
therefore,no reasonto expectthat theseinformal processeswill produce
generalconformity to either one of these opposing normative orders.
Thus, it is not surprising to find that detailed empirical study has
revealedno clear connectionbetweenconformity to any particularset of
social norms and the receipt of professionalrewards (Cole and Cole,
1973).
It appears,then, that conformity to most of the supposednorms and
counter-normsof scienceis largely irrelevant to the institutional proces-
seswherebyrewardsaredistributed.Researchers are simply rewardedfor
communicatinginformation which their colleaguesdeem to be useful in
70 Scienceand the Sociologyof Knowledge
the pursuit of their own studies.There are no institutional mechanisms
for rewardingin any direct way conformity to either set of social norms;
nor is it possibleto show that the provision of acceptableinformation
presupposesthe implementationof either set, for, as Mitroff demon-
strates,both contradictory sets can be interpretedas being thus presup-
posed.But if the 'norms'and 'counter-norms'describedin the sociologi-
cal literatureare not elementsin an institutionalisednormativestructure,
how are we to interpret the evidencepresentedby Merton, Mitroff and
others?One answerto this questionhas alreadybeensuggested,that is,
they are undoubtedlyrelatively standardisedverbal formulations which
are usedby participantsto describethe actionsof scientists,to assessor
evaluatesuch actionsand to prescribeacceptableor permissiblekinds of
social action. But standardisedformulationsof an evaluativekind never
govern social interactionin any straightforwardfashion. This point has
been made strongly by Gouldner:
... moral rules are not given automatic and mechanicalconformity

simply becausethey, in some sense, 'exist' ... conformity is not so
much given as negotiated. .. The rule thus servesas a vehicle through
which . .. tensionis expressed... thereis usually more than one rule in
a moral code that can be claimed to be relevant to a decision and in
terms of which it may be legitimated. A central factor influencing
one's choice of a specific rule to govern a decision is its expected
consequencesfor the functional autonomy of the part ... What one
conceivesto be moral, tends to vary with one's interests. (1971,pp.
217-18)
The relevanceto scienceof the general argumentsummarisedin this

quotation can be illustrated by referring once again to the discovery of
pulsars.When the first paper on pulsars was published in 1968 by the
radio astronomygroup at Cambridge,there were numerousaccusations
of secrecyfrom membersof groupsin competitionwith the astronomers
at Cambridge.It was said that the Cambridgegroup had unduly delayed
publication; that they publishedinsufficient data to allow other groups
to undertakesupplementaryresearch;that they should have passedon
their results before publication to close colleagues in neighbouring
laboratories;that their secrecypreventedthem from obtaining valuable
advicefrom others;and that their actiontendedto impedethe advanceof
science. Members of the Cambridge group, however, were able to
provide various principles justifying their actions. In the first place, it
was claimedthat it was perfectly legitimategenerallyto avoid passingon
information which could lead to anticipation by others. Secondly,
secrecywas justified on the groundsthat it gaveresearchers time to check
their results and to publish high-quality work, thereby ensuring the
smooth developmentof scientific knowledge.Thirdly, it was said to be
legitimateto makesurethat importantresultsimprovedthe reputationof
one'sown group and its ability to obtain researchfunds. Fourthly, it was
also saidthat scientistshad the right to protect thefirst achievementof a
young researchstudentor the right of observersto have the first attempt
at interpretingtheir own findings. Fifthly, it was said that stepshad to be
taken, in the particular case of pulsars, to prevent the press from
misrepresentingthis remarkablediscovery. As we would expect in view
of this confusingvariety of diffuse and overlappingrules, somepartici-
pants denied that there had been any undue delay in making public the
first observationsof pulsars.
In the original sociologicalstudy from which this material is taken, the
conclusionsabout the operating principles relevant to the communica-
tion of researchresults are summarisedas follows.
Theredoesnot appear,however,to be a generalcommitmentto these

principles; nor are there clear proceduralrules governing the commu-
nication of results. As a result, misunderstandingand resentmentare
sometimesproducedby what are variously regardedas secrecyor as
legitimate control over the circulation of scientific information. (Edge
and Mulkay, 1976, p. 250)
It is worth noting in connectionwith the argumentabovethat conformity

to social norms is irrelevant to the receipt of rewards, that despitethe
heateddifferencesof opinion at the time of the pulsardiscoveryaboutthe
proprietyof the actionsof the Cambridgegroup, six yearslater two of its
membersreceiveda Nobel Prize basedin large part on this discovery.
In science,then, we have a complex moral languagewhich appearsto
focus on certain recurrentthemesor issues; for instance,on procedures
of communication,the place of rationality, the importanceof impartial-
ity and of commitment, and so on. But if the example described
immediately above is representativeand if the preceding argument is
correct, then no particular solutions to the problems raised by these
issuesfor participantsare firmly institutionalised.Instead,the standard-
ised verbal formulationsto be found in the scientific communityprovide
a repertoireor vocabularywhich scientistscan use flexibly to categorise
professional actions differently in various social contexts. A major
influence upon scientists'choice of one verbal formulation rather than
another,as Gouldnernotes,is likely to be their interestsor objectives.It
can be assumedthat, for a given scientist or group of scientists,these
interests will vary from one social context to another. Thus, in the
examplegiven above, when researcherswere frustrated by the apparent
reluctanceof othersto makesignificant findings availableto them, they
tendedto selectprinciples favouring communality which justified their
condemnationof the others' behaviourand added weight to their own
exhortations.In contrast, those scientistswho had made the discovery
wereableto find principlesin favour of personalownershipof results. In
different circumstances,a person'sor a group's choice of rules can be
entirely reversed. Not only is it possible to vary one's choice of
formulationsas one attemptsto identify the evaluativecharacteristicsof
different acts,but it is also possibleto apply different formulationsto the
sameact as one's social context changes.For instance,it is possibleto
'justify' keepingcertainresultssecreton the groundsthat to do otherwise
would jeopardisea graduatestudent's first efforts; and then subse-
quently, when recognitionor prizes are being allocated,to maintainthat
in this contextthe student'scontributionshould be regardedas unimpor-
tant.
It must not be imaginedfor onemomentthat I am in any way accusing
scientistsof bad faith or of being less moral than other social groupings.
I am simply arguing that within the relatively distinct community
concernedwith scientific research,as indeedin most areasof social life,
interactioncannotbe adequatelydepictedas expressinganyoneor more
sets of institutionalisednormativeprinciples or operativerules deriving
from suchprinciples.It seemsmoreappropriateto portray the 'normsof
science', not as defining clear social obligations to which scientists
generallyconform, but as flexible vocabulariesemployedby participants
in their attemptsto negotiatesuitablemeaningsfor their own and others'
acts in various social contexts. The details of these social dynamics in
scienceare not yet well understood.For instance, it may be that the
ability to control others' accessto valued information may increase
scientists'capacityto gain acceptancefor their preferredcategorisations.
This may havebeentrue to someextentin the caseof pulsars.But until we
move away from the traditional accountof the scientific ethosit is hardly
possibleto conceivethat suchexerciseof powerplays any significant part
in the scientific community.Therehas,therefore,beenlittle investigation
of phenomenaof this kind. Nevertheless,although analysis of the
negotiation of social categoriesin scienceis still in its infancy (Law,
1976), what is clear is that it is highly misleadingto regard the diffuse
repertoireof standardisedverbal formulationsas the normativestructure
of scienceor to maintain that it contributesin any direct way to the
advanceof scientific knowledge.
This last point returns us to the discussionwith which this section
began. In abandoningthe customarycharacterisationof the scientific
ethos, we have removedany apparentinconsistencybetweenthe socio-
logical analysis of scientific norms and the philosophical analysis of
scientific knowledge.But this leavesan appreciablegap in the sociologi-
cal interpretation of science. For the traditional conception of the
scientific ethos, although it avoided any close examinationof scientific
knowledgeitself, did at leastprovidesomekind of generalaccountof the
latter's genesis: namely, as the necessaryby-product of widespread
conformity to the supposednorms of science.Clearly, then, as this view
of the social origins of scientific knowledgemust be rejectedalong with
the traditional idea of the scientific ethos, an alternative sociological
analysisof the productionof scientific knowledgeis needed.In the next
section, I will examine a number of recent case studies of scientific
developmentwhich draw attentionto featuresquite unlike thoseassumed
in orthodox sociological analysisand much closer to the conclusionsof
the new philosophy of science. I have no intention, however, 9f
attemptingto provide a complete, formal analysisof the social produc-
tion of scientific knowledge.
THE DYNAMICS OF KNOWLEDGE-PRODUCTION
When we examinethe overall growth of modernscienceone thing stands

out clearly, namely, that there has been a continuouscreation of new
areasof investigationand new realmsof knowledge(Price, 1963). In the
seventeenthcentury, the entire physical world fell within the scope of
'natural philosophy'; and one man could encompassin his studies the
full rangeof available knowledge. By the end of the nineteenthcentury
this had becomequite impossible. Knowledge of the natural world had
become much more extensive as well as more detailed and complex
(Mason, 1962); and the major scientific disciplines had crystallisedinto
more or less distinct intellectual domains,each of which was separately
establishedin the centresof higher learning with control over profes-
sional training and over accessto its own areaof knowledge.This process
of intellectual and social differentiation has continuedup to the present
day; so that now each discipline is further sub-divided into many
specialties.Each of thesespecialtiesis composed,in turn, of numerous
specificareasof research,most of which deal with phenomenaunknown
a generationbefore. It is true that someof the most well-known advances
in scientific thought have not involved differentiation so much as the
reconceptualisationof existing bodies of knowledge (Kuhn, 1962).
Nevertheless,to a very considerableextent, scientific knowledge has
developedby the identification and detailedinvestigationof phenomena
which have not been known to exist previously or which have not been
studied before in any depth. The typical pattern of growth, then, in
scienceis not the revolutionary overthrow of an entrenched orthodoxy,
but the creation and exploration of a new area of ignorance (Holton,
1973, ch. 12). Within many such areasthere occursa gradual movement
through three discernible,although overlapping,stages;that is, from an
initial phaseof exploration,througha stageof unification and into a final
period of decline (Mulkay, Gilbert and Woolgar, 1975). Although this
sequenceis by no meansinevitable (Law, 1976), it is characteristicof
most areas where consensus,and therefore certified knowledge, is
actually achieved.
If we are to understandhow scientific knowledgeis socially produced,
there are advantagesin concentratingparticularly on the early stagesin
the recurrentmovementtowardsintellectualconsensusin science.Collins
elucidatesthis point with the help of an analogy.
When we consider the grounds of knowledge, we do it within an

environmentfilled with objectsof knowledgewhich are alreadyestab-
lished. To speakfiguratively, it is as though epistemologistsare con-
cernedwith the characteristicsof ships(knowledge)in bottles(validity)
while living in a world where all ships are alreadyin bottles with the
glue dried and the stringscut. A ship within a bottle is a natural object
in this world, andbecausethereis no way to reversethe process,it is not
easy to accept that the ship was ever just a bundle of sticks. Most
perceptions of the grounds of knowledge are structured in ways
derivedfrom this perspective.(1975, p. 205)
Collins goeson to arguethat in the studyof sciencewe canto someextent

escapefrom this predicament,becausewe can identify fairly easily those
scientistsworking on a commonproblem and we can thereforeexamine
contemporaryscientific developmentsas they proceed.I intend now to
look at some of Collins's empirical material, as well as other recent
sociologicalwork, in order to seewhat is revealedby close study of the
social productionof scientific knowledge.It will becomeevident that the
conclusions reached in these studies help to extend the argument
developedin the precedingsection.
Collins's first study (1974) is of a social network of scientistscon-
cernedwith building a special kind of laser called a 'TEA laser'. This
network was exploring a new area in the sensethat, although various
kinds of lasers had been built before so that general notions were
available of what lasersin generalought to do, this particular form of
laserhad only recently beenmadeoperational.In 1970 an effective TEA
laserwas reportedin the literaturefor the first time and numerousgroups
beganto try to producetheir own versions.At the time of the study some
had succeededand somehad not. Collins concentrateson the transmis-
sion of information amongthesegroups. Hiscentral finding is that even
thosescientistswho had actually produceda working laser were unable
to make fully explicit the knowledgewhich was required. In many cases,
of course, scientistswere simply unwilling to formulate their technical
knowledgeopenly. They preferred not to pass on all that they knew,
becauseto have done so would have reducedtheir competitive advan-
tage. But there were also cases where participants whose lasers were
operativeseemedto be unableto conveytheir knowledgeto others.Thus
one scientist would help another to build a machine which seemedto
both partiesto be identicalwith his own successfullaser, only to find that
inexplicably it failed to work. Furthermore, no group managed to
producean operatinglaser on the basisof formal publishedinformation
alone. Successalways depended on direct personal contact. These
contactsoften had to be repeatedmany times until, unpredictably,the
laser could be deemedto be operating properly (Collins and Harrison,
1975). Collins suggeststhat personalcontact was essentialbecauseonly
through direct interaction could scientists communicatethe tacit and
unformalisedknowledgeon which their work depended.
It seems,therefore,that in this areaat this early stageit was impossible
for participantsto know whethera scientist had absorbedthe required
knowledgeexcept through his successfulinteraction with other special-
ists. Members' scientific competenceand compliance with technical
standardscould not be judgedfully by the applicationof pre-established,
formal criteria; but only by ex post facto negotiation with those other
participantswho werealreadyregardedas competentjudges.An interest-
ing questionhereis whetherthis network'srelianceon informal processes
and tacit knowledge would have been less visible, without necessarily
being less important, at a later stagewhen all i~terested groups had an
operativelaser,and what looked like a set of 'principles of construction'
had been ostensiblyagreed.It may well be that at this later stageboth
participantsand analyst would have been much more likely to portray
the construction of TEA lasers as due to the implementation of an
unambiguousseries of specifiableinstructions. But this is speculation.
Collins emphasisesonly that it seemsinappropriateto describepartici-
pants at the time of his study in the traditional sociological manner as
carriers of independent,impersonalknowledge. 'The point is that the
unit of knowledge cannot be abstracted from the "carrier". The
scientist, his culture and skill are an integral part of what is known'
(1974, p. 183). The tacit and personal character of much scientific
knowledge and the social negotiation of what is to count as valid
knowledge are examined further in Collins's study of research into
gravitationalwaves (1975).
It is only recently that the phenomenonof gravitational waves has
receivedmore than passingscientific attention. Up to about 1969 only
one scientisthad tried to detectthis form of cosmic radiation. However,
after this man publisheda formal claim to have observedgravitational
waves, numerousother scientistsquickly enteredthe field and devised
their own observationalequipment.Therewas fairly completeagreement
amongparticipantsin this field on certain generalpoints. Most of them
agreed,for example,that gravitationalwavesare predictedby Einstein's
generaltheory and that certain catastrophicastrophysicaleventsshould
releasesuch waves. There was also broad agreementabout the kind of
observationalapparatusrequired and about the classical physics which
underliestheseexperimentaltools. In addition, all participantsappeared
to endorsesimilar generalcriteria of validity; for example, they agreed
on the needfor controlled experimentand for theoreticalinterpretation
which was consistentwith the experimentalevidence.This seems,then,
to have beennormal science,that is, an attemptto fill in the details of a
well-established paradigm by means of conventional observational
methods.Yet, Collins stresses,the work involved was in no way routine,
for most participants were fundamentally uncertain about how the
existenceof gravitationalwavescould be demonstratedand the network
was deeply divided about the meaning of its members' experimental
results.Thus neither establishedknowledgenor formal criteria provided
any unambiguousmeans of distinguishing valid from invalid claims.
Both thesekinds of cultural resourceswere interpretedin diverse ways
when applied to the new caseof gravitationalwaves.
Collins is particularly concernedto counterwhat might seemto be the
obvious interpretationof the activities of scientistsin this field, namely,
that they wereengagedip straightforwardreplicationor refutation of the
original claimant'sresults.The difficulty with this interpretationis that,
in practice, replication or disproof can only be achievedwhen there is
considerableagreementaboutthe meaningof observations,the adequacy
of experimental procedures, and so on. For replication to occur,
participantsmust be able to decidewhat are to count as equivalentand
reliable observations.Much of the researchon TEA lasersis fairly close
to the usual conceptionof replication because,in this case,there was a
criterion against which successor failure could be measuredand there
were 'competentjudges' with whom agreeddefinitions of 'success'and
'failure' could be reached;althoughevenhere the criterion was difficult
to specify and its application was a matter for negotiation. But neither
agreed criteria nor competent judges were available in the case of
gravitational waves.
Collins producesa variety of evidenceto demonstratethat there was
no commonassessment of the experimentalproceduresor results of any
of the membersof this network. What onescientistviewed as interesting,
another dismissed as unimpressiveand a third violently rejected as
outright fraud. Accordingly, nobody was at all interestedin attempting
to repeat in detail the original experiment which had provided the
impetus for further researchin the area. There was no point either in
duplicating or in failing to obtain a result which had no scientific
meaning.Thus participantswere more directly concernedwith devising
somenew kind of observationwhich they thought had a better chanceof
being recognisedas a 'competentmeasureof gravitational waves'.
Although most participantsclearly thought of themselvesas engaged
in checking in some way the original knowledge-claim as well as
subsequentfindings on gravitational waves, Collins suggeststhat they
did this indirectly by enteringinto 'negotiationsabout the meaningof a
competentexperiment'(1975, p. 216). Becausethere was such a variety
of scientific opinion in the field, members' primary concern was to
establishwhat shouldcount as a 'working gravity wave detector'.If they
were to succeedin doing this, Collins points out, they would provide an
effectiveinterpretationof the phenomenonof gravity waves.This can be
illustratedby the simple exampleof experimentson heat. Becausethere is
general agreementabout the nature of heat, a great variety of super-
ficially different activities areseenby scientistsas beingequivalentin that
they are all 'competent temperature measuringexperiments'.In treating
different measuresas equivalent,for example,dipping a glasstube filled
with mercuryor two metalslinked by a voltameterinto a liquid, scientists
aredrawing on their knowledgeaboutthe characteristicsof heat; and the
completeset of equivalentactivities goessomeway towardsdefining the
nature of heat. Furthermore,if some experimentsand measureswhich
are now excludedfrom the class of 'competentexperiments'were to be
included,it would follow that the acceptedpropertiesof heat would have
changed(Collins, 1975, p. 217). In the caseof temperaturemeasurement,
and to a lesserextent laser building, the technicalculture is sufficiently
well establishedfor participantsto distinguish competentfrom incom-
petent work and valid measuresfrom invalid m~asures with (compara-
tive) ease.But this was not so in the caseof researchon gravity waves.
Thus as researchersin this latter field negotiatedagreementabout which
experimentswere to be regardedas competentand equivalent,they were
defining the nature of their problematic empirical phenomenonand
creating a distinctive region of scientific culture.
Collins pays little attention to clarifying in detail the form that
negotiation takes in this field. But the picture seemsto be roughly as
follows. Participantsbring to the study of gravitational waves certain
generalassumptionswhich they appearto share.As we have seen,these
include such varied interpretativeresourcesas Einsteinian theory and
belief in experimentalmethods.Theseresourcesare brought to bear on
the largely unexplored phenomenonof gravitational waves; but they
yield no unique or unambiguousinterpretation. Participants do have
enough in common to concentrateon a limited range of empirical
variableswhich they are able to seeas related, as being membersof the
sameset. But Collins showsthat thesevariablesare employedselectively
and interpreteddifferently in argumentsfavouring different knowledge-
claims. What one participant takes as given, another takes as proble-
matic and even as underminingthe adequacyof the other'sclaim.
Thus eachexperimenter(or researchgroup) usesthe availabletechnical
culture and his own expertise in a flexible manner to reveal the
inadequaciesof others' findings and to supporthis own claims. Further-
more, those involved seem not to distinguish clearly betweentechnical
and non-technicalcriteria of evaluation(see also Bourdieu, 1975). For
judgementsaboutthe merit of knowledge-claimsare said by participants
to depend also on personal considerationssuch as one's faith in an
experimenter'scapabilitiesand honesty,one's views about his personal-
ity and intelligence, his reputation, social location and psychological
approach,one'saccessto 'inside information', and so on.
Of course,as we saw in the previous section, it is unlikely that these
verbal formulations determine scientists' technical assessmentsin any
strongsense.It is more probablethat they are a way of strengtheningthe
presentationof an evaluation already conceived. However, the same
point seems to apply equally to the use of scientific and technical
formulations (see also Bloor, 1976). Both the social and the technical
culture of science appear to provide members with flexible symbolic
resourceswhich can be, and are, combined to devise a considerable
variety of interpretativepositionsin connectionwith a commonresearch
problem. Scientistsdo havea social and technicalculture which provides
somethinglike a sharedcontext in this area of research.Thus certain
unorthodoxnotions, such as the existenceof a 'fifth physical force' or
the operationof psych~c forces, although sometimesused informally to
explain particularlydifficult results, were seldomallowed into the public
forum. They were not part of the acceptedrepertoire. But the precise
meaningof the orthodoxcultural repertoirehasto be establishedanew in
this emergentfield by processesof symbolic interpretationand negotia-
tion. As Collins puts it, quoting from McHugh: any consensuswhich
ensues'is conceivableonly as a socially organisedupshot of contingent
coursesof linguistic, conceptualand social behaviour' (1971, p. 329).
One of the important merits of Collins's work is its use of contem-
porary data. He providesus with an unusualview of scienceby recording
what happensinformally, beforea firm scientific interpretationhas been
acceptedand before an establishedbody of knowledge has come to
appearto be the only rational possibility (see also Kemp, 1977). Several
other recent studies, despite being retrospective rather than contem-
porary, havereachedsimilar conclusions.Suchretrospectivestudieshave
one major advantageover Collins's work, namely, they can more easily
follow the processof negotiation through its full sequence.They can,
therefore, examine how consensusis achieved, maintained and some-
times abandoned.Gilbert's study (1976b) of radar meteor research,for
instance,showshow consensusemergedwith respectto a major issue(the
origin of sporadic meteors),even though central participants both for
and against this consensusseemedat the time to regard the available
scientific evidenceas incomplete or inconclusive. In other words, the
attainmentof consensusin this case was not due solely to intellectual
considerations.In anotherretrospectivestudy, Pinch (1976) looks back
at quantummechanicsin the 1950s and is able to show how physicists
were able to defendtheir establishedframework of interpretationagainst
a threatening knowledge-claim by a highly selective employment of
theoreticalresourcesin different social contexts.This case,he suggests,
supports 'the view that scientific theories themselvesare multi-dimen-
sional and that what constitutesa theory in scienceis a variable and will
mean different things to different groups of scientists'.
A particularly well-documentedstudy of this kind, which I will now
examine in some detail, is that by Wynne (1976) of Barkla and the J
phenomenon.Wynne's objective is to revise the prevailing account of
'the J phenomenonaffair' and to correctthe interpretationcurrent in the
'oral folklore' of the scientific community. This latter interpretation
depictsthe disputeabout the J phenomenonas being resolvedin favour
of Barkla's opponentsby the applicationof impersonalrules, which are
independent of social context and which distinguish unequivocally
betweentruth and error. Wynne argues,in contrast,that the 'scientific
reasons'given in the literature for rejecting Barkla's claims are better
conceived as a rhetoric seeking to justify a rejection based on other
considerations.These other considerations,he suggests,include social
factors and point towards'a different versionof scientific rationality and
knowledge' (1976, p. 308).
There are two distinct phasesin the history of the J phenomenon.In
the first phase, between 1912 and 1923, the eminent physicist Barkla
proposed a theory to explain a new set of X.-ray emissions, the 'J
radiations',emanatingfrom the electronsof a specific 'shell' or seriesin
the atom. This theory was formulated in terms of the 'classical'
interpretationof X-ray scatteringand initially it was widely accepted.By
the early 1920s,however, observationalanomalieshad begunto appear.
In particular, physicists using the increasingly popular spectrometer,
which had been widely adopted becauseof its precision and the vast
amount of experimentalwork which it opened up, failed to confirm
Barkla'sresults. This was actually to be expectedfrom Barkla'spoint of
view; for the beams which he wished to measurewere of such low
intensity that, he reasoned,they would not be detectableby means of
thesenew observationaltechniques.Mainly for this reason,Barkla had
taken the unorthodoxdecision to use relatively 'old-fashioned'absorp-
tion methodswhich, although they were difficult to manipulateexperi-
mentally, had the crucial advantageof being more sensitive to low-
intensity emissions. Wynne maintains that Barkla's unusual, but
reasonedtechnical strategywas never taken fully into considerationin
assessmentsof his work. Instead physicists simply used their own
(inappropriate)'technicalnorms' in an unself-criticalmannerto question
the competenceof Barkla's researchand the validity of his knowledge-
claims.
Before long, theoreticalobjectionswere added to these experimental
difficulties. For in 1922 Compton proposed a new theory which,
supported by unexpected empirical evidence and some precise and
sophisticatedcalculations, shatteredthe classical theory of scattering.
One major implication was that, if Compton's theory were accepted,
therewas no room theoreticallyfor a J seriesin the atom. As Compton's
interpretationcame to attract widespreadsupport and as opposition to
the J seriesgrew, Barkla undertookmore absorptionexperimentsand in
1923 he abandonedthe J series theory as a valid explanation of his
results. But he did not adopt Compton's theoretical analysis. Barkla
continuedto believethat there was an observationalphenomenon,the J
phenomenon,which still required proper theoreticalinterpretation.
Barkla'srepudiationof his J seriestheory was not an entirely negative
move, for he had seen'the vague but tantalizing outline of things much
more revolutionary' (1976, p. 314). Accordingly, over the next decade,
he sought to develop an alternative theoretical model which differed
from his own initial theory and from the dominant interpretative
framework. During this secondphasehis work departedmore and more
from the researchtradition of mainstreamphysics. He continued to
reject the use of the spectrometerand someof the scientific assumptions
associatedwith that technique. He maintained that the fashionable
researchtechnology forced nature to fit preconceptionsbuilt into the
design of the major instrument and that the discovery of fundamental
results was sacrificed for highly precise routine measurements.He
stressedthat heterogeneousX-ray beams did not always behave as a
simple linear sum of individual wavelengthcomponents(a basic premise
of orthodoxspectrometeranalysis),but rather as an 'organic whole'. In
accordancewith this concernfor a more holistic approachhe employeda
scientific idiom derived in large measurefrom nineteenth-centurynatur-
philosophie; and he continued to use an out-of-date and unpopular
hydrodynamicmetaphor,which expressedhis belief in the continuity of
nature but which was unsuitablefor a physics community increasingly
committed to quantum theory and the view that radiation is always
emitted in discretequantities.
During this phaseof researchinto the J phenomenon,a considerable
number of paperswas published by Barkla and his students. But this
work had no impact on the developmentof orthodoxphysics, exceptthat
several reviews were produced in which Barkla's claims were finally
discredited.It is usually held that in thesereviews Barkla'swork is either
shownto be explicablein terms of Compton'sby then orthodox theory
or to be so unsystematicand irregular as to call for straightforward
rejectionas incompetentresearch.The centralpoint of Wynne'sanalysis
is that thesereviews do nothing of the kind and that, in fact, there is no
explicit refutation available of Barkla's work which can plausibly be
describedas an impartial evaluationbasedon establishedstandardsof
scientific rationality.
In the first place the most influential review, written by Dunbar,
appearedin 1928, some five years before Barkla finished publishing
results. It is, therefore, difficult to accept the popular opinion that
Dunbarfinally buried the J phenomenon.Secondly,all the reviews refer
to Barkla'smistakeover the J theoryas if this were in someway evidence
of his being mistakenagain about the J phenomenon.Wynne points out
that a 'strongclaim is thus being madehereagainstBarkla's credibility, a
claim that borrows little from any of the usual prescriptions for the
rational conduct of scientific disputes' (1976, p. 327). Barkla is also
criticised for being unableto identify the causalfactor which triggers off
the organic mode of action of heterogeneousX-rays as well as for
publishingincompleteaccountsof his experimentalarrangements.These
criticisms are usedin a highly selectivemanner.For, as we saw in the case
of laser research,it is literally impossibleto provide a completedescrip-
tion of experimental proceduresand this is not usually required of
experimentalists.Nor are observersalways expected to give a causal
explanationof their results before the latter are acceptedas competent.
In addition to theserelatively generalcharges,most reviewersprovide
a more detailedanalysisof some of Barkla's empirical claims. But their
conclusIonsat this substantivelevel are no moreconvincing. For instance
Dunbar, althoughhe noted that Barkla was attemptingto treat complex
radiationsas organic wholes by meansof appropriatetechniques,used
homogeneousbeams for his 'replications' and thereby excluded the
factor of heterogeneitywhich alone gave mean~ng to Barkla's experi-
ments. For Barkla, Dunbar'swork was not a replication of his own and
had no bearingon the validity of his findings. For Dunbarand for most
orthodox physicists, however, it was acceptedas a clear refutation.
Wynne also shows how Dunbar, in order to achieve his 'refutation' of
Barkla'sresults,was forced to reinterpretand repudiatesomeof his own
earlier work which he had previously presentedas being quite unequi-
vocal and free from the possibility of error. 'In marginal casessuch as
this, one seeswhat one is predisposedto see,and intepretationalcaution
is castto the winds in the tacit strategyof persuasion.But it is remarkable
how the meaningof the old observationswas changeddiametrically, so
as to supportthe new conclusions'(1976, p. 328). The logic of this kind
of argumentseemsto be that of interpretingboth empirical results and
general standardsof adequacy in accordancewith a predetermined
conclusion,in this casethe falsity of Barkla's claims.
Thereare numerousother examplesgiven by Wynne of Barkla'scritics
selectively interpreting data in a way that discredited their opponents'
views. In addition, he showshow these'empirical refutations'were given
added force by being embeddedin a web of critical comment which
systematicallyunderminedBarkla's scientific credibility. For example,
Barkla was accusedof wanting dogmaticallyto retain the classicaltheory
of X-ray scatteringand of being unwilling to give up his own J theory.
Wynneshowsthat both theseaccusationswere unjustified. One reviewer
even used a mistake made by two of Barkla's research students to
condemnhis entire researchschool, suggestingthat this erroneouspaper
was Barkla's only major support. The reviewer perhaps didnot notice
and certainly did not mention that Barkla never tried to use the paperin
support of the J phenomenon.It is clear, Wynne contends, that the
account usually given of the rejection of Barkla's work on the J
phenomenonis quite misleading.Barkla'sclaims were neverconclusively
disproved by clear demonstrationof their failure to meet impersonal
standardsof validity.
The publishedrefutationsof Barka'swork are markedby confusionas

to the nature of his exact position (even to the extent that views set
down in major journals were apparentlynot known); by standardsof
evidencewhich can only count as 'standards'to those already predis-
posedto accepttheir conclusions;by heavily implied sanctionsagainst
a scientific adversarywhich entail norms of a strictly non-rational
kind; and by extravagantinconsistencyas to why the phenomenon
should be rejected.Thesepublishedrefutationsshould, I propose,be
treatedmoreas symptoms,as rhetorical rationalizationsof a rejection
alreadysealedfor less.tangiblereasons.(Wynne, 1976, p. 335)
Having reachedthese conclusionswith respect to the case of the J

phenomenon,Wynne goeson to offer a somewhatspeculative,but none
the less interesting,generalanalysisalong the following lines. Scientific
consensus,and thus scientific knowledge, is not achievedby meansof
conclusiveproof and disproof. Scientistsare always faced with ultimate
uncertaintyand ambiguity. But intellectual commitmentsare necessary
and are regularly made. They are not achieved, however, by the
applicationof any set of pre-established,formal criteria. Adoption and
rejectionof researchprogrammesis a much more pragmaticprocessand
is greatly influencedby scientists'relatively local interests.For example,
the spectrometerbecamea dominantresearchtool largely becauseof its
practical advantages;it could be used routinely to provide precise
measurements over a wide rangeof areas.Yet, once firmly established,it
becamea major resourcefor rejectingknowledge-claimswhich embodied
different scientific assumptions.This means that 'orthodoxy lurches
along via apparentlyarbitrary factors locatedin social (including techni-
cal) practices; and that traditional predispositions,socialized values,
assumptionsand practices, and the like, play a crucial part in the
eventualroute which"true" scienceis deemedto have taken. That route
is only formally mappedout in hindsight'(1976, p. 336).
Wynne is here suggesting that the formal rationality of scientific
knowledge,which requiresthat such knowledgebe shown to conform to
invariant principles of validity, tends to be brought to completionafter
the event.Oneimportantaspectof this operationis the final repudiation
of any significant opposingviews. In this way, the appearanceof formal
rationality is strengthened.This, in turn, has two further consequences.
Within the researchcommunity, it fostersthe concentrationof intellect-
ual effort. In addition, it helps the scientific professionto maintain its
credibility in the wider society and to obtain economic and social
support.
This last statementdraws attention to the fact that the studies with
which we have been concernedso far deal only with those processes
occurring within relatively small and highly specialised research net-
works. Clearly thesesocial networksare partsof a broaderpatternand in
the final chapter I will examine some of the links between scientific
researchand the wider society. For the moment,however, I shall extend
thescopeof the discussiononly slightly, by looking at two studiesin which
and the rest of the researchcommunity. The first of thesestudiesis

by Collins and Pinch of parapsychology(1978). Theseauthorsstressthat
th~lt
some attention is given to the connectionsbetween specific specialties
the reality of such paranormalphenomenaas extrasensoryperceptionor

psychokinesisis irrelevant to their analysis. They are concernedsolely
with studying how scientistsattempt to establishor refute knowledge-
claims dealingwith that classof phenomenareferredto as 'paranormal'.
In particularthey arguethat, by exploring this case,we will be able to see
how scientific knowledgeis the contingentoutcomeof both social and
cognitive processesin the researchcommunity as a whole.
Claims by a recognisedscientistto havedemonstratedthe existenceof
paranormalphenomenawere first advancedin the 1930s. But general
agreementabout the meaning of basic observationsor the scientific
legitimacy of this kind of inquiry has still not been achievedtoday. One
reason for this is that the debate, or negotiation, about paranormal
phenomenahas not beenrestrictedto thosedoing detailedresearch,that
is, to parapsychologiststhemselves.Many scientistsworking primarily in
other areasseemto have regardedthe claims of parapsychologistsas a
threat to establishedknowledge and to the credibility of science.Thus
one pronouncedfeature of the debate has been the contest between
parapsychologists,whoseminimal claim has beenthat there are genuine
phenomenato be scientifically investigated,and orthodox scientistswho
have sought to underminethe parapsychologists'entire enterprise.
Collins and Pinch distinguish two 'forums' in which this debate
betweenorthodox and deviant scientistshas been carried out.
On the one hand there is what we will call the 'constitutive' forum,
which comprisesscientific theorisingand experimentand correspond-
ing publication and criticism in the learnedjournals and, perhaps,in
the formal conferencesetting.On the other hand, there is the forum in
which are set thoseactions which-accordingto old-fashionedphilo-
sophic orthodoxy-arenot supposedto affect the constitution of
'objective' knowledge. We will call this the 'contingent' forum, and
would expect to find there the content of popular and semi-popular
journals,discussionandgossip,fund raising and publicity seeking,the
setting up and joining of professionalorganisations,the corralling of
studentfollowers, and everythingthat scientistsdo in connectionwith
their work, but which is not found in the constitutiveforum. (1978)
Within the framework of what I havecalled the standardview of science,
which Collins and Pinch refer to as 'old-fashionedphilosophic ortho-
doxy', we would expect the successor failure of parapsychologyto be
determinedsolely in the constitutive forum. Parapsychologistswould
present the results of a number of carefully conducted experiments
through the formal channelsof communicationand theseresults would
be judgedby the application of impartial canonsof proof, consistency,
adequateevidence, and so on. Presumablywithin a relatively short
period of time, it would be possible for virtually all those involved,
whatever their initial inclination, to decide objectively whether or not
there were genuine phenomenato be investigated. Collins and Pinch
show, however,that the debatehas not beenremotely like this. They go
on to arguethat the issuewas not resolvablewithin the formal limits of
constitutivedebateand that both sideshavein fact continually employed
resourcesfrom the contingent forum in their attempts to establishas
authoritativetheir opposingdefinitions of paranormalphenomena.
Let us look first at the critics of parapsychology.In the first place, it is
clear that many of them preferred to keep the knowledge-claimsof
parapsychologists from being consideredat all in the constitutiveforum.
For example,papersfrom parapsychologists were regularly rejected,not
only when refereeswere evenly divided in their recommendations,but
even on those few occasionswhen there was a majority in favour of
publication. Moreover, when positive results were actually published,
journal editors were inclined to indicate in one way or anotherthat the
journal was not endorsing these findings. But, as we have noted, the
screeningof papersby refereesand editors is merely the first step in the
processof evaluation.Collins and Pinch stressthat formal presentation
in itself need never lead to acceptanceby determinedopponentsfor, as
we saw in the caseof gravitationalwaves,evaluativeresourcesare highly
flexible and can be employed to support quite divergent scientific
positions. Thus some participantssimply saw the empirical findings as
uninteresting.The so-calledresultswere 'empty correlations'which were
not worthy of further attention. For others,however,the decisivefactor
in rejecting the results of parapsychologicalresearch was that the
empirical resultswhich had beenobtainedhad beengiven no theoretical
analysis which was at all convincing. This was contested by many
parapsychologists,most of whom saw themselvesas working within an
appropriatetheoreticalframework.
So far, then, we have noted clear differencesof opinion about what
constitutes an empirical finding in this area and what counts as a
theoreticalinterpretation.In addition, it was possible,without doubting
the validity of someof the experimentalwork by the acceptedstandards,
to questionthe validity of the standardsthemselves.Thus onemathema-
tician reasonedroughly as follows: 'Experimentalresults interpretedby
classicalprobabilitytheoryappearto indicatethe existenceof paranormal
phenomena.Although the experiments seem to be satisfactory, the
conclusioncannotbe accepted.Therefore,thereis somethingwrong with
probability theory'. If this line of reasoningwere to be generallyaccepted
it would imply that establishedproceduresof inferenceover a wide area
of science, namely, those fields which depend on probability theory,
would have to be revised and the boundariesof certified knowledge
extensivelyre-drawn.Few scientists,however,thoughtthat therewas ar:ty
needto changetechniquesof statisticalinferencein responseto findings
which they saw as being blatantly false. A much more usual procedure
was merely to assertthat the existenceof paranormalphenomenawas
inconsistentwith incontestablescientific knowledgeand, therefore, that
any positive results, no matter how plausibly presented,must be the
product either of experimentalerror, fraud or self-deception.
At this point, Collins and Pinch suggest,we can see the critics of
parapsychologybeginningclearly to deploy the repertoireof the contin-
gent forum in the course of formal knowledge-constitutivedebate.
Thus the caseof parapsychologyis especiallyinformative in sociological
terms because,as a result of the new specialty'sdeviant and threatening
appearancein the view of the orthodox, the merging of the contingent
with the constitutive forum in the process of cognitive negotiation
becomesunusuallyevident. For instance,critics have frequently asserted
that parapsychologyis merely an irrational cult maintainedby faith, in
contrastwith genuinesciencewhich is basedon evidenceand demonstra-
tion. Parapsychologyis said to be merely another form of spiritualism
and belief in the occult. This characterisationenablescritics to account
convenientlyfor the fact that somephysicists,psychologists,and so on,
appear to have a different view of the available evidence and the
conclusionsto which it leads. Belief in the rationality and consensusof
science is made consistent with apparently fundamental intellectual
divergence,by viewing those with opposingideas as, in the last resort,
simply irrational.
Onceadvocatesof parapsychologyhavebeendefinedin this way, there
is no need to treat its knowledge-claimsseriously or to convince its
practitionersof their errors. A similar kind of argumentis that which
accuses,not just individual parapsychologists,but the whole specialty.
of fraud. From this perspectivethe critic may admit that the evidence
occasionallylooks compelling. But consistencyis maintainedbetween
this evidenceand what is taken to be establishedknowledgeby asserting
that the former is not and cannotbe real evidence.As Collins and Pinch
show, there is no defenceagainst this argument when it is pushedto
extremes.Critics can always show that results could have been fabri-
cated,either in the courseof normal interactionbetweenthe experimen-
ter and his subject or between the subject and outside helpers, for
example, by meansof hidden transmitters.
This tradition of giving credibility and persuasivenessto the fraud
hypothesis by the demonstration of its possibility has become a
standardprocedurein the critique of parapsychologicalexperiments
... The logic of the fraud hypothesisnot only appearsto removeany
needfor empirical testsfrom the scientific decision-makingprocessbut
can also be put forward without any empirical evidence that fraud
actually took place. (Collins and Pinch, 1978)
The tactics of the parapsychologistsappear to have been broadly

similar to those of their opponents.It seemsthat some parapsycholo-
gists, at least early in their research,had assumedthat the existenceof
paranormalphenomenacould be establishedexperimentally,that is, by
the formal publication of a definitive experiment.But as we have seen,
any experimentalresultsin this field can be rejectedwith the help of the
interpretative resources available to orthodox scientists. And para-
psychologists,as well as someof their opponents,havecometo recognise
this. Accordingly, parapsychologists haveemployeda variety of supple-
mentarymethodsfor establishingthe legitimacyof their claims, from the
relatively technical to the obviously social. At the technical level, they
have increasinglyadoptedthe complex experimentalmethodsof physi-
cists, biologistsand psychologists.They have also refined their statistical
proceduresso that, as early as 1937, they were able to gain official
endorsementby the Institute of MathematicalStatistics. Moreover, one
of the central foci of interest among parapsychologistshas been the
attempt to reconcile their findings with orthodox electromagneticand
quantum-mechanicaltheory. These efforts within the constitutive
context have been accompaniedby action in the contingent forum.
Collins and Pinch show that by the latter action parapsychologistshave
tried to influencethe receptiongiven to their knowledge-claims;they also
suggestthat contingentaction has to some extent helped parapsycholo-
gists to establishtelepathy, clairvoyanceand psychokinesisas genuine
scientific phenomena.
Collins and Pinch provide clear evidencethat parapsychologistshave
soughtsystematicallyto obtain university postsin their subject,to recruit
students,to acquire funds from legitimate sources,to publish in recog-
nisedjournals, and to join prestigiousscientific associationssuch as the
AAAS. In addition, it appearsthat, as parapsychologistshave increas-
ingly obtained these social objectives over the years, the legitimacy of
their intellectualendeavourshas becomemore widely accepted.This has
been particularly true in the USA, where the social organisation of
scienceprovides a relatively favourable environment for the establish-
ment of new scientific groupings(Ben-David, 1968). Of course,the mere
existenceof a relationship betweenincreasingscientific legitimacy and
organisationalexpansiontells us very little. Collins and Pinch claim that
parapsychologistshave gained increasing intellectual acceptanceby
obtaining most of the social characteristicsot other legitimate sCIentifIc
specialties.But clearly the processcould operatein the other direction,
with parapsychologybeing allowed to expand becauseits work was
becoming more acceptable.However, the authors do provide evidence
that social position may be used as a resourceto strengthenknowledge-
claims. For example, they describe instanceswhere parapsychologists'
'proper' scientific training, their affiliation with major researchestab-
lishments,and so on, were explicitly acceptedby the orthodoxas factors
which required them to take seriously knowledge-claimsabout para-
normal phenomena.But Collins and Pinch furnish no evidenceto show
that parapsychologistswere able to use their improving organisational
position directly to ensurethe acceptanceof any particular knowledge-
claim as valid. Thus their central conclusioncan best be interpretedas
follows: parapsychologists'action in the contingent forum helped to
persuadeoutsidersof the legitimacy of their overall researchenterprise;
and to create a social context in which particular knowledge-claims,
which would previouslyhavebeendismissedout of hand, wouldbe more
likely to receivecareful considerationand, in somecases,acceptance.
So far in this section I have consideredonly studies dealing with the
production of knowledgein new areasof scientific investigation. But a
similar combination of social and cognitive processescan be seen to
operatein the courseof 'revolutionary' situations.To illustrate this, let
me look briefly at a study by Frankel (1976) of developmentsin early
nineteenth-century Frenchresearchinto the classicaltopic of optics. This
case is particularly interesting becauseit shows how variations in the
social position of participants may foster the emergenceof radically
divergentscientific interpretationswithin a strongand fruitful consensus
and also how the social organisationof the researchcommunity may
affect the successof unorthodoxideas.
The central feature in French optical research during the period
1815-25 was a dramatic swing from the corpuscularto the wave theory
of light. Frankel arguesthat a Kuhnian analysis of this changeis not
entirely appropriate and by no means complete. From the Kuhnian
perspective,we would expectthis revolution to have followed a period of
crisis within the researchnetwork involved; this crisis arising out of an
awarenessof anomaliesdefined in terms of the current interpretative
framework and widely recognisedby participants. However, Frankel
showsthat in this instancetherewas no prior crisis and no recognitionof
major anomaliesby adherentsof the dominant paradigmuntil after the
revolutionarysequencehad alreadybegun.The precedingdecadehad, in
fact, beenone of unprecedentedsuccessand productiveelaborationfor
thoseemploying the corpusculartheory. At the beginning of the decade
of changethere was no awarenessof important problems which could
not be resolvedalong orthodox lines.
The major changein theoretical perspectivewhich did subsequently
occur could neither have begun nor have been brought to fruition
without the activities of a group of 'revolutionaries'.This was so because
almost all the influential figures in French physics were rigidly opposed
to non-corpuscularoptics and becausethesemen controlled the dissem-
ination of legitimateinformation in this field. The scientific revolution in
French optics was set in motion by actions carried out in both the
contingentand the constitutive forums by a small number of scientists
who were isolated, or estranged,from the centre of power and ortho-
doxy. Frankel identifies Arago and Fresnelas key figures in the revolt.
Fresnelwas at this time quite outsidethe researchcommunityand eagerly
seekinga chanceto make a scientific reputation. His sponsor,Arago,
was an establishedscientist,but he was alienatedfrom the ruling clique-
until after the successof the wave theory, that is, when he took over the
dominant position that Laplace had occupieda decadeearlier. Frankel
contendsthat scientistsin such marginal social positionsare more likely
to be receptiveto alternativeinterpretationsof resultswhich can always,
in principle, be seen in various different ways; and that they are more
likely to exploresuchalternativesbecausethey havemore to gain thereby
than the establishedrepresentatives of orthodoxy. It is certainly the case
that a large number of major scientific advanceshave originated with
researchersin marginal situations (Chubin, 1976). At the time that
Fresneland Arago beganto challengethe corpusculartheory there was
one alternativeschemeavailable which had undergonesome degreeof
scientific refinement and which could be used, therefore, to provide
immediately fairly sophisticatedanalysis. This was the wave theory,
which had recentlybeenemployedby Young in England but which had
until then madelittle impact on orthodoxy in either country.
It is clear that both Arago and Fresnelbecamecommittedto the wave
theory before any kind of 'proof' of its validity was available, even in
their own terms; and that from early on their central aim was conceived
as building up evidencein its favour. As their work proceeded,they came
to differ increasinglyfrom mainstreamscientistsin their recognition of
anomalies,in the theoretical resourcesthey brought to bear on these
anomaliesand in their detailedinterpretationof data. 'Thus diffraction,
for example,was seenas a centralproblemby Arago and Fresnel,while
to Biot it was a relatively minor concern. It is not the nature of the
anomalyalone, therefore,but the situationsof the scientistsinvolved in
the disputeas well, which determinesthe courseof events'(1976, p. 175).
Whilst the revolutionarieswere attemptingto redefinethe interpretative
frameworkof optical research,the moreorthodoxscientistscontinuedto
extend t-he corpusculartheory with undiminishedconfidence.At least
initially, they did this effectively. As Frankel stresses,the corpuscular
theory was not at this stagein a stateof intellectual crisis. Becausemost
participantscontinuedfor sometime to acceptthe corpusculartheory as
scientifically adequate,the case for social factors having a decisive
influence on the emergenceof an opposingperspectiveseemsespecially
strong.
If scientists' judgementsand interpretationsare significantly influ-
encedby social ~s well as cognitive factors, it follows that any deviant or
revolutionary group, if it is to succeed,must act at both the social and
cognitive levels. Frankel draws attention to three things which such a
minority group must do: it must produce 'solutions' to problems
regarded as important, it must pUblicise its views and it must win
recognition for its work. The degree to which these objectives can be
achieved, he suggests, depends significantly on variations in social
context. In order to clarify this point, Frankel makes some interesting
comparisonsbetweenthe position of Young in England and that of
Fresnel and Arago in France. The central point is that, as the same
intellectual resources were available in both countries, the varying
successof the wave theory in the two researchcommunitiesmust have
beendue primarily to social differences. For instance,it was relatively
easyfor Young to get his ideaspublishedand known in the open, laissez-
faire communityof British science.But he failed completelyto convince
anyone that his analysis was preferable to the corpusculartheory. In
contrast, Fresnel was able to win a hearing from the more closed and
centrally dominated community in France only through the 'political
leverage'of Arago. However, oncethis breakthroughhad beenachieved,
Fresnel was able to engage in dialogue with a group of 'competent
judges' and to succeedcomparativelyquickly in establishing that his
work could not be ignored.
The difference in the matter of recognition lay in the state of profes-

sionalization of the two communities. Recognition requires agreed-
upon standardsas to what constitutes an adequatesolution to a
problem. In the relatively amateurishstateof early nineteenthcentury
British physicstherewas no such agreement,so Young's critics could
disregard him for philosophical reasons without confronting his
mathematicalarguments.In France,the Laplaceschool and the Ecole
Polytechnique had combined to set certain standardswhich were
commonlyaccepted,even by warring partiesin a scientific revolution.
If onecould predict numericallythe resultsof experimentsfrom theor-
etical reasoningexpressedin the form of mathematicalequations,then
one was acknowledgedto have achieveda solution to the problem at
hand. (Not necessarilythe only solution, or the best solution, how-
ever.) Thus the judges in the diffraction contestcould award Fresnel
the prize on the basisof his mathematicsand his experiments,without
committingthemselvesto his theory. (Frankel, 1976, p. 176)
ConsequentlyArago, acting on Fresnel'sbehalf within the councils of

the Parisian scientific elite, was able to use established criteria of
90 Scienceand the Sociologyof Knowledge
adequacyas a resourcein the informal processof negotiation. He was
able to employ agreed standardsto define Fresnel's contributions as
significant and to convinceothersthat Fresnel'swork was consistentwith
those standards.
Frankel emphasisesthat, althoughthis recognition of the significance
of Fresnel'swork was the crucial first step towards the successof the
wave theory, it did not settle the issue. Most orthodox scientists
continued to deny the validity of Fresnel's interpretations and their
oppositionwas not endedsimply by increasinglyclear scientific demon-
strations in accordancewith agreed criteria. The standardssharedby
both sides were interpreted so as to support quite different scientific
analyses.Frankel claims that the rapid rise to dominanceof the wave
theory in Francewas not brought about by cognitive factors alone but
also by socio-political factors. In particular, the debate over the two
theoriesof light coincidedwith the victory of an anti-Laplacefaction in
the Frenchscientific communityas a whole. As a result, thereoccurreda
wide-rangingrevision of interpretativepositions in various fields.
From the early 1820s onwards, supporters of Laplace's view of
science,including the adherentsof the corpusculartheory of light, were
reducedto silence(Fox, 1974). Although many of them refusedto recant,
they were unable publicly to opposethe new orthodoxies. One conse-
quence of these social changes was a rapid and almost complete
acceptanceof the wave theory by the Parisian scientific community.
From their new positionsof power in teaching,researchand publishing,
adherentsof the wave theory were able to convert the next generationof
students in its entirety. It seems, therefore, that there was nothing
inevitable in this intellectual revolution. It did not occur as a necessary
consequenceof the logic of scientific ideas, but was set in motion and
brought to completion in part by contingentsocial factors. Where the
appropriatesocial context was lacking, as in England, the corpuscular
theory remaineddominantfor sometime to come. It is perhapsrelevant
to note that this was not the end of the 'corpuscularversuswave' debate,
which surfaced again in a new interpretative context in the present
<;entury. (See also the discussion of Holton's themata in the next
chapter.)
Let me conclude this section by summarisingvery briefly the main
conclusionsof the casestudiesdiscussedabove. I will refer in passingto
philosophicalsourceswhere similar points are made. In Collins's study
of the lasernetwork, we saw that scientific knowledgecannot always be
made explicit. It has a tacit componentwhich cannot be assessedby
meansof formal criteria (see Polanyi, 1958, and Ravetz, 1971). Such
tacit knowledge is transmitted most effectively through direct social
interaction,in the courseof which the adequacyof participants'know-
ledgehasto be informally negotiatedratherthan formally demonstrated.
In the study of gravitational waves we were shown that, even in a
researchcontext where the intellectual framework is well establishedand
theoretically elaborated, available knowledge and techniques can be
interpretedin quite diverseways (see Hesse,1974, and the discussionin
the precedingchapter about the variable meaning of concepts, etc.).
Beginning from significantly different interpretativepositions, partici-
pants appearedto be working out, by a processof 'claim and counter-
claim', the meaningof observationsand thereby,indirectly, establishing
the nature of the phenomenonwith which they were concerned(see
Feyerabend,1975). As in the case of laser research,this process of
negotiationwas not confined to the formal channelsof communication,
but dependedto a considerableextent on informal interaction which
enabledparticipantsto buttresstheir formal claims with a wide variety of
non-technicalevaluations.As we saw in the previous section, the use of
such non-technicalor social formulations is likely to be highly flexible
and context-dependent.
Wynne's study showed how consensusmay be achieved and main-
tained, not by clear disproof of alternativepositions, but in part by the
highly selectivepresentationof evidence(see Hanson, 1969, on selective
perception),by the 'misrepresentation'of opposingargumentsand by
attackson their author'sscientific integrity. Wynne suggeststhat scien-
tists necessarily make strong commitments to research programmes
which can never be formally justified in full (see Lakatos, 1970). Like
Collins, he stressesthe informal, unexplicated,intuitive elementsin the
creation,justification and maintenanceof scientific knowledge.Because
the commitmentson which the courseof scientific developmentdepends
cannotbe formalised, they are open to influence from a variety of social
factors. Wynneconcludesthat the formal rationality of scienceis at least
partly a form of justification constructedafter, and with the help of,
social and intellectualcommitments.Accordingly, he agreeswith Collins
that we will obtain a totally misleadingview of scienceif we infer its
social attributes from the formal characteristicsof the claims presented
in articles, reviews and textbooks. Formal knowledge claims have
meaningonly when they are interpretedby the membersof actual social
groupings.The way in which these interpretationsare realised depends
on the outcomeof contingentnegotiationsamong those members.
In the caseof parapsychologywe saw that well-establishedscientists
seemto be able to resist formal constitutiveknowledge-claimsindefini-
tely when theseclaims are basedon radically different assumptions(see
Kuhn, 1962; Polanyi, 1969). For the critics of parapsychology,the
centralassumptionthat paranormalphenomenado not exist was never in
question.Ratherit pervadedand gavemeaningto the whole armouryof
formal arguments which they employed (see also Bloor, 1976). It
ensuredthat, for thesecritics, every item of evidenceand every chain of
reasoningprovidedfurther groundsfor rejectionof the deviant views. In
this case,the informal reasoningand the use of contingentcriteria which
are normally excludedfrom open display in the formal arenaappeared
clearly in the courseof knowledge-constitutingdebate.The authorsshow
how parapsychologists,faced with the apparentimpossibility of proving
their case through formal discourse yet convinced of its validity,
explicitly adopted a supplementarystrategy in the contingent forum.
They argue that parapsychologistshave had some successwith this
strategy and have been able to use their improved position in the
scientific community as a resource for establishing their intellectual
claims. This themeis taken further in the final study by Frankel, who
arguesthat scientists'intellectualcommitments,their choice from avail-
able analytical resources, may be directly affected by their social
location; that the successof unorthodoxviews canbe crucially influenced
by the internal politics of science;and that the formal rationality of an
emergentparadigmmay be madeunproblematicfor a new generationof
scientistsby becomingentrenchedin the centresof scientific authority.
Thesecasestudiesare not free from interpretativeor methodological
difficulties. It would be surprising if they were; for studiesof this kind
havenot beenattempteduntil very recentlyandthe practicaldifficulties of
combiningan investigationof social relationshipswith detailed study of
intellectual developmentsare very great (Law, 1976; Woolgar, 1976b).
Their conclusions,therefore,shouldbe treatedas tentative.They cannot,
however,beignored.At the very least,the evidenceproducedin theseand
similar studiesdemandsfurther analysisand empirical investigation. In
the next sectionI will try to identify the centralimplication of thesepath-
breakingsociologicalstudies of the production of scientific knowledge
and to show that it is supportedby the revisedanalysisof social norms in
science.To put this anotherway, my aim will be to distinguishthe central
featureof the analyticalframeworkwhich, strengthened by its consistency
with the new philosophyof science,is emergingasan alternativeto the cus-
tomary sociologicalaccountof science.But before I do this, I want to
makeone important point. All the casestudiesdiscussedin this section
havedealtwith fields of empiricalscience.It shouldnot be assumed,how-
ever,that 'non-empirical'disciplinesdevelopquite differently or that their
knowledgehasa certaintyor clarity different in kind from that of empiri-
cal science.In otherwords, we cannotretain the categoryof 'specialcase'
traditionally usedby sociologistsof knowledgeby confining its scopeto
mathematicsandlogic. This hasbeenarguedmostforcibly by Bloor (1976,
chs 6 and 7), who shows that mathematicalformulations and logical
principles have no meaning until they are interpretedin terms of non-
formal, backgroundassumptions;that these assumptionsare socially
variable; that mathematicalreasoningis, therefore, context-dependent;
and that mathematicalproofs are producedby informal processof social
negotiation.Bloor's analysisis easily available and I will not discussit
further. I will, however, assumethat the conclusionswhich follow are
likely to apply as much to the formal as to the empirical sciences.
THE INTERPRETATION OF CUL TURAL RESOURCES
In the light of the material presentedin the two precedingsections, it

seemsthat one crucial fault in the orthodox sociological accountof the
production of scientific knowledge is the lack of any conception of
interpretationor negotiation(Law and French,1974: Bohme, 1975). This
omission,and the gap it leavesin sociological analysis,are evident with
respectto the treatmentof both social and cognitive/technicalresources.
From the orthodox perspective, it is assumed that sociologists can
identify that set of generalnormativeprinciples which in practiceguides
most activities in scienceand, indeed, which has to be institutionalised
within the researchcommunity in order to guaranteethat the great
majority of accepted knowledge-claims will be faithful to the real
physicalworld. Thesegeneralprinciplesare conceivedas providing clear
prescriptionsfor virtually all social actioninvolved in the productionand
certification of scientific knowledge.Their applicationby participantsto
particular acts is taken to be quite straightforward. Specific acts are
regarded as either conforming to a given rule or not, in a fairly
unambiguousmanner.I have tried to show, however,that for two main
reasonsthis conceptionof the normativestructureof scienceis unsatis-
factory. First, only part of the range of normative principles actually
usedby scientistsis recognisedas playing a significant role in science;and
the sub-setselectedfor attention emphasisesunduly those formulations
most obviously consistent with the standard philosophical view.
Secondly, and more important here, both general principles and more
specific operatingnorms have, as we have seen,always to be interpreted
in particularcases.In order to depict a given action, whethertheir own or
that of some other, as in accordancewith a prescription, participants
have to interpret that prescriptionin relation to supplementaryconsid-
erationsas well as those particular featuresof the context in which it is
being applied that are deemedto be relevant.
Accordingly, I have tried to show that we must revise the established
view of the relationship between social norms and the production of
scientific knowledge. The meaning of norms is always socially contin-
gent; that is, it dependson interpretationby actors in varying social
contexts.Becauseany specific norm can be madeconsistentwith a wide
rangeof apparentlydifferent actions, we cannot regard the production
of knowledgeas a simpleconsequence of conformity to any particularset
of normative formulations. I have suggestedinstead that it is more
appropriate to treat the norms of science as vocabularies which are
employedby membersin negotiatingmeaningsfor their own and their
colleagues' actions. Becausescientists have available a considerable
variety of formulations,eachof which can be applied to individual cases
in a flexible manner,any given act can always be interpretedin various
ways. The extent to which one interpretation rather than another
becomesacceptedby participantsis the outcomeof processesof social
interaction or negotiation; that is, as members exchangeviews and
attemptto convince,persuadeand influenceeachother, theseviews may
be modified, abandonedor reinforced. Although social negotiation in
sciencehas beenlittle studiedas yet, it seemslikely that its outcomeis
influenced by such factors as members'interests,their intellectual and
technical commitments,members'control over valued information and
researchfacilities and the strengthof their claim to scientific authority.
This argumentwith respectto social normsin scienceis supportedby a
closely parallel argumentwith respectto cognitive/technicalnorms. (I
use this last phraseto refer to the whole range of researchmethods,
techniques,criteria of adequacy,establishedbodiesof knowledge,and so
on, in so far as they are employedas resourcesfor judging knowledge-
claims and scientific competence.The similar phrase'technicalnorms',
which is sometimesused within the orthodox framework, usually has a
somewhatnarrowerconnotation.)From the perspectiveof the customary
sociological analysis, the supposedsocial norms are seen as necessary
becausethey ensurethat the technicalnormsof sciencewill be vigorously
applied in the selection of knowledge-claims(see Chapter 1). These
technical norms, in turn, are regardedas pre-established,impersonal
rules which directly regulatescientists'researchactions and intellectual
judgements. It is argued that, in this way, the 'entire structure of
technical and moral norms implements the final objective' (Merton,
1973, p. 270), that is, the accumulationof objective knowledge. The
technical resourcesof scienceare seenas being sociologically unproble-
matic. It is assumedthat their interpretation and application will be
uniform and independentof variations in social context within the
researchcommunity. (It is, of course,emphasisedthat, if the supposed
social norms are disrupted, the technical norms will ceaseto function
properly.) However, we have seen in the present chapter that this
assumptionis unjustified. Not only are social norms socially variable,
but cognitive/technicalnorms are also open to a considerablerange of
interpretationin any particular researcharea. In other words, general
evaluative criteria such as 'consonancewith establishedknowledge',
'consistencywith the evidence','competence','replicability', as well as
the content of specific bodies of knowledgeand technique,all require
interpretationin particularinstancesin much the sameway as do social
norms (Bohme, 1977). Indeed, it is difficult to envisagehow technical
resourcescould be employed differently from social resourcesin this
respect because,as every case study demonstrates,there is no clear
separationbetweenthe negotiationof social meaningand the assessment
of knowledge-claims.Both social and technicalformulations have to be
selectedand interpretedby participantsin particular instances;and both
kindsofresourcesare inextricably combinedin the sequenceof informal
interaction as well asformal demonstrationwherebyspecificknowledge-
claims come to be ratified. (Thus the distinction between social and
technicalresourcesmust not be reified. Cognitive/technicalformulations
are merely one kind of interpretativesocial resource.)
The sociologicalanalysisof science,then, has previouslyassumedthat
the productionof scientific knowledgecan be explainedby showingthat
generalconformity is maintainedto setsof formal rules (both social and
technical),the strict implementationof which guaranteesan undistorted
revelation of the real physical world. I have argued, in contrast, that
neitherof thesekinds of rule has a determinatemeaningfor participants
and that implementation therefore requires a continual process of
cultural reinterpretation.By means of this processscientists construct
their versionsof the physical world. The broad similarity betweenthis
revisedsociologicalposition and the new philosophyof scienceis clearly
evident at this point (see the discussion at the end of Chapter 2).
Sociologistsand philosophershaveconvergedon a conceptionof science
as an interpretativeenterprise,in the courseof which the nature of the
physical world is socially constructed.
It is also clear, I think, that this view brings sciencefully within the
scopeof the sociologyof knowledge.I do not mean by this that we can
begin to talk of the content of scientific knowledgeas 'determinedby
existential factors'. The kind of analysisimplied in such terminology is
clearly quite at variancewith the view that I havebeentrying to illustrate
and clarify. A rather bettergeneralformulation would be that scientific
knowledge is establishedby processesof negotiation, that is, by the
interpretationof cultural resourcesin the course of social interaction.
Cognitive/technicalresourcesare employedby scientistsin such negotia-
tion; but the eventualoutcomedependsalso on the availability of other
kinds of social resources.The conclusionsestablishedthrough scientific
negotiationare not, then, definitive accountsof the physical world. They
are rather claims which have been deemedto be adequateby a specific
group of actorsin a particularcultural and social context. Thereis, then,
at least a prima Jacie case in favour of the thesis that 'objects present
themselvesdifferently to scientistsin different social settings, and that
social resourcesenter into the structure of scientific assertions and
conclusions'(seeChapter1, p. 2).
In the present chapter, I have developedthis thesis with respect to
processesoccurringinside the researchcommunity. In the final chapter,I
will show that the sameapproachcan elucidatesome of the connections
betweenscientific knowledgeand the societyat large. This will enableme
to illustrate how some of the traditional questionsof the sociology of
knowledgemay be appliedto the caseof scienceand, perhaps,provision-
ally answered.
4
Scienceand the Wider Society
In general, sociologistsof knowledgehave paid particular attention to

the influence exertedby external social factors on the work of specific
groupsof cultural producers.For instance,Stark (1958) claims that the
existenceof radically divergent traditions of philosophical thought in
Germanyand Britain is primarily a consequenceof differencesin socio-
political environment.He suggestsalso thatthe transition from classicto
romanticart in Europeat the turn of the eighteenthcenturywas fostered
by a markedchangein the social position of the artist, which was in turn
a result of wider social developments.Theseare typical of the connec-
tions betweencultural productsand societywhich havebeenidentified by
sociologists.Not only the rate and direction of cultural development,but
also its content,are portrayedas directly dependenton external influen-
ces. In the caseof science,however,externalfactors havebeen regarded
as less powerful. It has come to be generallyagreedthat the speedand
direction of scientific developmentare considerablyaffected by social,
economicand technicalfactors originating outsidethe scientific research
community (Mendelsohn, 1964; Ben-David 1971). But most philoso-
phers, historiansandsociologistshavebeenunwilling to acceptthat such
externalfactorscan influencethe contentof scientific thought, that is, its
concepts,empirical findings and modesof interpretation(Young, 1973,
Lemaineet al., 1976).
The reasonsfor this are, I hope, clear by now. Scientific knowledge
has beenconceivedas an objectiverepresentationof the physical world.
The modernscientific communityhas beencreditedwith an ethos which
reduces social influences upon the production and reception of
knowledge-claimsto a minimum, therebyguaranteeingthe accumulation
of objective knowledge. Given these assumptions,direct connections
betweenthe wider society and the conclusionsof scienceare not to be
expected,exceptin thosefew cases,like that of Lysenko, where outside
intervention'distorts' scientists'results(Joravsky,1970). Consequently,
when sociologistshave sought to understandthe relationship between
scienceand the wider society,they haveconceivedtheir centraltask to be
that of demonstratingwhich kind of society is most amenableto the
institutionalisationof the 'scientific ethos'and most likely to supportan
autonomousresearchcommunity.
This has led to the thesis that democraticsocietiesfurnish the most
Scienceand the Wider Society 97
congenialsetting for scientific development,becausesuch societiesgive
academicscientiststhe freedom they require to record without bias the
facts of the naturalworld and becausescienceand democracysharethose
values on which the production of valid knowledge depends(Barber,
1952; Merton, 1957, p. 522; Hirsch, 1961; see also Polanyi, 1951).
However, the assumptionsbehind this line of reasoning have been
challenged in the two previous chapters. We have seen that it is
preferable to think of scientific knowledge as a contingent cultural
product,which cannotbe separatedfrom the social context in which it is
produced.We have also seenthat the supposedscientific ethosis merely
part of the cultural repertoireof science,and by no meansnecessarilythe
most important part of the creationof scientific knowledge.There is no
longer, therefore,any reasonto expect that scienceis best createdin a
social vacuum where institutionaliseddemocraticvalues allow disinter-
estedresearchersto formulate the 'one correct account of the physical
world'.
The revisions in the customary view of science which have been
presentedabove enable us to reconsiderthe possibility of there being
direct externalinfluenceson the contentof what scientistsconsiderto be
genuineknowledge.This issueis now empirically open and we can turn
to detailedstudiesof the developmentof scientific thoughtto seehow far
it is influenced,on the one hand, by the actionsand cultural productsof
non-scientistsand, on the other hand, by the actions and cultural
acquisitionsof scientiststhemselvesin non-scientificcontexts.We are no
longer forced to reject thesepossibilitiesout of hand as inconsistentwith
the cognitive and social character of science. Moreover, although
sociologistshave not yet explored thesepossibilities, somesocial histor-
ians of sciencehavebegunto do so (MacLeod, 1977). In the next section,
therefore, I will examine some recent work in the social history of
science.I will not discussthe full rangeof externalinfluenceson science.
I will concentrateinsteadon just a few analysesdealing with the content
of science,in order to establishclearlinks with the argumentpresentedin
previous chapters.
SCIENTISTS'USE OF 'EXTERNAL' CULTURAL RESOURCES

We can think of scientists as having access to two main cultural
reservoirs:that providedby 'the scientific community' and that provided
by the wider society. As the scientific community has grown larger, its
own resourceshavebecomemore extensive.Consequentlyit seemslikely
that, as Durkheim suggested,sciencehas becomeculturally more indep-
endent over time. Increasingly it is other scientific sub-cultures that
provide interpretative resources, with the products of physics and
mathematicsbeing exploited by specialties in chemistry and the life
sciences.'As sciencehas incorporatedinto itself more and more of the
cultural resourcesof the societiesin which it hasthrived, so it has become
moreinternally self-sufficient,with cross-fertilizationbetweenspecialties
replacing "fully external" inputs in the process of cultural change'
(Barnes,1974, p. 119).
Despitethis tendency,it is still possibleto find externalculture being
taken into science.For example,scientiststoday still make considerable
useof commonsense knowledgewhich is acquiredlargely in the courseof
non-scientific activities. Close observationof scientists at work shows
that they continuallymove betweena highly esotericterminologyand the
languageof everydaylife. Thus 'commonsense modesof perceptionand
operationare an integral and essentialfeature of recognizedscientific
practice'(Elliot, 1974, p. 25). Even in the most rarified areasof physics,
informal reasoningand debatemakeuseof a wide rangeof interpretative
notions brought in from ordinary discourse; and not only discourse
aboutphysical objectsbut also about social relationships.Thus particles
'attract' and 'repel' one another.They are 'captured'and they 'escape'.
They 'experience' 'forces'. They 'reject' and 'accept' 'signals'. They
'live' and 'decay',and so on (Holton, 1973, p. 106). It is, of course,clear
that theseterms acquire new meaningsas they are used in this unusual
context. Nevertheless,their meaning continues to resemblethat with
which scientistsareaccustomedin the courseof their everydaysocial life.
Physicistsadopt these terms becausethey are familiar and, therefore,
furnish ideas which can be applied by analogy to enable physicists to
reasonfrom the known to the unknown (Deutsch, 1959).
Little is known about such informal processesof reasoningin science
and the kinds of resourceswhich are employed.Only Holton (1973) has
attemptedany kind of systematicanalysisof historical examplesin order
to explorehow informal thinking contributesto scientific development.
Like the sociologistswhosework was discussedin the previous chapter,
he stressesthat informal processesare of fundamentalimportance in
scienceand that their significance has not been properly recognised,
largely becauseparticipantscover up the transitionfrom private specula-
tion to formal demonstration.When we examinein detail how scientists
actually reach their conclusions,as opposedto the way in which they
presentthem formally, we are 'overwhelmedby evidencesthat all too
often there is no regular procedure,no logical systemof discovery, no
simple continuousdevelopment.The processof discovery has been as
variedas the temperamentof the scientists'(1973, pp. 384-5).
Holton's central and well-documentedcontention is that informal
reasoningin sciencedependson basic presuppositionswhich inform and
guide scientists'work, sometimesleading them to 'wrong' conclusions
but also in many casesenablingthem to disregardcontrary evidencein
pursuit of what is subsequentlyseen to be the correct interpretation.
Each scientist, Holton argues, makes a commitment to a particular
approachtowardshis areaof study. He commitshimself, for instance,to
the notion of atomistic discretenessor continuity, to harmony or
conflict, to developmentor equilibrium, and so on. Holton stressesthat
these commitmentsare 'neither directly evolved from, nor resolvable
into, objective observationon the one hand, or logical, mathematical,
and other formal analytical ratiocination on the other hand' (1973, p.
57). They either precede formal interpretation or they are adopted
without formal proof to resolve interpretative problems. Indeed such
commitments,of which we have seen severalexamplesabove, are not
opento direct proof or disproof. Ratherthey define the point at which it
is no longer appropriateto ask further questions.They provide taken-
for-grantedassumptionswhich are used to generateinterpretationsand
thus to 'bridge over the gap of ignorance'.
Apart from the evident consistencybetween this analysis and that
developedin previouschapters,the point of particular relevancehere is
Holton's proposalthat this repertoireof basicthemesor presuppositions
spansthe boundariesbetweenthe scientific communityand the societyat
large. It is part of a common imaginativeinheritance.'What is interest-
ing is that on certainoccasions,during the transformationof conceptions
from the personalto the public realm, the scientist, perhapsunknow-
ingly, smugglesthe commitmentof his individual system,and that of his
society, into his supposedlyneutral, value-indifferentluggage' (1973, p.
101). Holton shows that in Greek science basic presuppositionswere
takenfairly openly from the social thought of the time. Sciencethen was
acceptedas an adjunct to moral philosophy. WhenHolton considersthe
emergenceof modern sciencehe finds that the natural philosophersof
Newton's time tried to avoid having to state or discussthe theological
and other 'non-scientific'notions which. in practice,contributedsignifi-
cantly to their analyses.This is one of the crucial differences between
modern scienceand its predecessors.The cultural and social roots of
knowledgehave been hidden away in modern science,on the mistaken
assumptionthat true knowledgeshould not involve relianceon unverifi-
able assumptions.
Holton (1973) carriesout severalcasestudiesto show that the cultural
connectionbetweenscienceand society today is not always as different
from that of Newton's, or even Plato's, time as we have been led to
believe.Thereis in practicea continualcultural exchangebetweenscience
and the wider society. Interpretative resourcesenter science mainly
throughinformal thinking, usually with only a very limited awarenessof
their external origins on the part of participants.They are refined and
modified in the courseof informal negotiation;and they are allowed into
the public annalsof scienceonly after appropriatereformulation. These
interpretativeresourcesare not generatedby the 'facts of nature', nor by
the social life of a segregatedresearchcommunity alone. They must be
understoodat least in part as productsof the social processesof society
at large.
If this analysisis correct, it should be possibleto producetwo broad
class.esof supportingevidence.In the first place, it should be possibleto
discern parallelsbetweenthe style of thought in certain areasof science
and that occurringin other areasof cultural production,such as painting
or philosophy-areaswhich are clearly influenced by the surrounding
social context. Holton has begun to do this, but his evidenceis as yet
quite tentative (see also Kroeber, 1944). The secondkind of evidenceis
that provided by detailed historical study of specific scientific develop-
ments.I intend nowto look at the emergenceof the Darwinian theory of
evolution where, becausea major scientific upheavalwas involved, the
documentationis extensiveand the movementof ideasrelatively easyto
discern.
Both sociologistsand historianshave generallytreatedthe contentof
the Darwinian theory of evolution and its acceptanceas being independ-
ent of the socialsettingin which it was broughtforth. We saw in Chapter
1 how Starkarguesthat Darwin'stheory is simply a summaryaccountof
observablefacts and is not, therefore, open to sociological analysis.
Historianssimilarly have distinguishedDarwin from other evolutionary
writers, such as Lamarck, Chambersand Spencer.It is acceptedthat the
speculationsof these authors, which seem now to have been mostly
inaccurate, were influenced by ideological and other non-scientific
factors. But Darwin is seen as the first to recogniseand describethe
actual mechanismof evolutionarychange.Consequently,his analysisis
presentedsimply as a detachedresponseto objective evidenceand 'is
treated in relative isolation from the social and intellectual context in
which he worked' (Young, 1971a, p. 443). Let me try to show just how
misleadingthis view is.
Five important elementsin Darwin's theory can be clearly identified.
The first of these was the belief that the facts of natural history,
comparativeanatomy,paleontology,and so on could be explainedbetter
by a conceptionof evolutionary developmentthan by the traditional
notion that specieswerestableand had beencreatedmore or lessin their
presentform. The secondelementwas the attemptto show that species
did in fact change over time. Becauseit was impossible to obtain
systematicevidenceof suchchangesas they occurred(presumably)in the
wild, Darwin turned to the close observationof domesticanimals and
plants. The third crucial element was the assumptionthat different
biological structures were functionally adaptedto different kinds of
environment.Fourthly, Darwin saw a direct parallel betweenthe process
of adaptationin domestic organismsand that taking place in natural
settings. The artificial selection of certain biological forms among
domestic animals and plants, in accordancewith the preferencesof
breeders,provideda model for understandingthe natural selectionwhich
occurred,in accordancewith the requirementsof survival, in the wild.
Lastly, Darwin acceptedthat the world of living things operatedin a
uniform fashion. There were, therefore, universal regularities in the
realm of biology as well as in the domains of astronomyand physics.
Thus Darwin regarded his account of the mechanism of 'natural
selection'as expressinga law of nature. Let me examinethe origins of
each of theseelements.
Therecanbe little doubt that theoriesof evolution sprangup in the late
eighteenthandearly nineteenthcenturiesin thosecountrieswherecapita,l-
ism was most advanced,that is, in Western Europe and particularly in
Britain. Sandow(1938) has shownclearly that the notion of gradualand
continuousbiological evolution occurredin thosesocietiesat that time as
a responseto the massiveaccumulationof new informationabout plants,
animals and fossils. This information had accruedas a by-product of
such developmentsas the worldwide exploration in search of markets
and the growth of the mining industry. Incidentally, then, the economic
and political expansionof capitalistEuropegeneratedevidencewhich led
to the formulation of new biological theories. The new data obtained
haphazardlyby menin pursuit of practicalaffairs often appearedto be at
variancewith establishedbiological views and, for somescholarsat least,
implied the need for a radically new interpretativeapproach.
By the fourth decade of the nineteenth century a new kind of
specialismhad emerged,which included Darwin, Huxley, Hooker and
Wallace. The membersof this specialismhad obtainedfirst-hand know-
ledgeof biological variation by taking advantageof the opportunitiesfor
doing field work offered to wealthy amateursby the official voyages
undertakento improve trade routesand to consolidatecolonial empires.
Virtually all of these men came to believe in the reality of biological
evolution. Thus Darwin'sacceptanceof the generalideaof evolution was
madepossibleby his social position; that is, by his belongingto a society
which had accessto a wider range of biological evidence than ever
before, by his having sufficient income to devote himself entirely to
science, and by his membershipof a distinct sub-culture which had
alreadyproducedseveralevolutionarytheories.But what of the content
of Darwin's work? If we are to understandthis in greaterdetail we must
do more than link Darwinian theory with broad featuresof nineteenth-
centurycapitalism.We must seek the social origins of the more specific
elementsof Darwin's analysis.
Darwin was uniqueamongbiologistsof his day in devising a long-term
programmeof recording the details of variation in plants and animals
under domestication.This was his solution to the problem of showing
clearly that changesin biological structuredo occur and that they can be
producedby selectiveinheritance.In pursuit of this objective, Darwin
'collectedfacts on a wholesalescale... by printedenquiries,by conversa-
tions with skillful breedersand gardenersand by extensive reading'
(Darwin in Sandow,1938, p. 321; seealso Vorzimmer, 1969). Most of his
observationson domesticvariation were, in fact, takendirectly from the
work of breeders(Young, 1971a);and in order toobtain this information
Darwin spenta great deal of time mixing with businessmen,commercial
breedersand fanciers.Therecan be no doubt that Darwin's treatmentof
domesticvariation was 'rooted in the practical activities of plant and
animal breeders';activities whose successwas measured,not by the
validity of members'knowledge, but by the amount of their financial
profit (Sandow, 1938, p. 332). The assumptionsand actions of these
practicalmen were absorbedby Darwin. They enteredinto his scientific
assertionsand they provided him with criteria for warranting his own
knowledge-claims.For instance,in clinching his argumentthat selected
inheritanceis the sourceof variation in domesticanimals,Darwin refers
to the fact that 'breedersof animals would smile' at any contrary
opinion. He goes on to cite several cases where large profits were
obtainedby selectivebreedingand he finishes with the statementthat
'Hard cashpaid down over and over againis an excellenttest of inherited
superiority' (Darwin in Sandow, 1938, p. 322).
It appears,then, not merely that Darwin's work was madepossibleby
the high level of attainment reachedin commercial breeding in nine-
teenth-centuryEngland as well as by the other aspects of capitalist
developmentalreadymentioned,but also that his observations,conclu-
sions and criteria of adequacyin relation to domesticvariation were in
somedegreetaken over from commercialbreeders.The perspectiveof
the breedersguided Darwin's detailed reasoningabout domesticvaria-
tion and, thereby, his inferences about the importance of selective
inheritanceas the sourceof evolutionaryadaptation.But the significance
of Darwin's closerelationshipwith the breedersdoes not end there, for
their proceduresalso provided the central metaphor or interpretative
themewhich informs the rest of his evolutionarytheory. I will return to
this metaphorshortly. However, in order to understandthe sourcesof
Darwin's analysismore fully, we must digressbriefly to considerseveral
other notions which he took over from theological and philosophical
debateabout the future of society and man's place in nature (Young,
1969).
Darwin'suseof suchnotions is evidentin his treatmentof adaptation.
The variation found in domestic plants and animals is adaptivein the
sensethat particular forms are selectedby breedersin accordancewith
their own purposesand the demandsof the market for which they cater.
It doesnot follow necessarilyfrom this that variation in natural settings
is also adaptive. Darwin assumed, nevertheless,that domestic and
natural selection are equivalent in this respect. He adopted the 'tacit
assumptionthat every detail of structure,exceptingrudiments, was of
some special, though unrecognizedservice' in natural as well as in
artificial environments(Darwin in Young, 1971a,p. 468). Thereseemto
havebeentwo sourcesfor this presupposition.One was William Paley's
writings on natural theology, which were intended to show that all
featuresof the naturalworld weredesignedby a beneficentGod. Darwin
admittedthat he was not able to annul the influenceof this belief, which
was then very widely held. In Darwin's work the idea that God actively
constructseach.speciesto a preconceiveddesign is abandoned;but the
idea that every item of biological structure plays a useful function is
retained.
The secondsourcewas Lyell's philosophicalprinciple of the uniform-
ity of nature. Darwin used this notion to argue that the processes
underlying variation in natural and domesticsettingsmust be identical;
although, of course,the meanswhereby specific variations are selected
certainly differs. It is importantto realisethat thereis nothinginherentin
the principle of uniformity which requiresus to acceptthat domesticand
naturalvariation arein the samedomain of phenomena.The principle in
no way specifiesthe rangeof its own application(seeChapter2). This is
illustrated by the fact that it took Darwin and Lyell many years to agree
that human beings came within the range of phenomenacovered by
evolutionarytheory. Lyell objectedto the idea that uniformities applic-
able to animals also applied to human beings. It is clear then that,
although Darwin sometimesclaimed to be following a 'true Baconian
method'and simply to be collecting facts, he actually employedpresup-
positions taken from philosophical and theological debate, and used
them selectivelyto extendand define the scopeof his scientific analysis.
Indeed this point can be taken much further. For virtually every
interpretative notion used by the biological evolutionists had been
employed earlier in the course of debate about society and human
progress(see Young,1969 and 1971a, for documentationof this point).
Once Darwin had concludedthat domestic and natural variation are
equivalent, he completedhis theory by extending his interpretationof
domesticbreedingto speciesin naturalsettings.In doing this, he retained
to a surprising degree the terminology of intentions and purposes
appropriateto the actionsof commercialbreeders(Young, 1971a).Even
in the formal presentationof his theory, he askedhis readerto conceive
of naturalselectionas beingcarriedout figuratively by a 'being infinitely
moresagaciousthan man' who operatedon biological populationsin the
wild so as to select 'exclusively for the good of each organic being'
(Darwin in Young, 1971a,p. 461). This metaphorcontinuesthroughout
Darwin's exposition.
It may be said that natural selectionis daily and hourly scrutinizing

throughoutthe world, everyvariation, eventhe slightest; rejectingthat
which is bad, preservingand adding up all that is good; silently and
insensiblyworking, wheneverand whereveropportunityoffers, at the
improvement of each organic being in relation to its organic and
inorganicconditionsof life. (Darwin, 1859, p. 84)
The metaphorof natural selection, which implied that there was an
agentwho did the selecting,causedDarwin manyproblems.For the rules
of scientific interpretationdevelopedin the seventeenthand eighteenth
centurieshad sought to banish anthropomorphismfrom the scieritific
repertoire. Indeed, it was partly on thesegroundsthat previous evolu-
tionary schemes,such as that of Lamarck, had been found wanting.
Accordingly Wallace, Lyell and other fairly sympatheticcommentators
criticised Darwin for using this kind of terminology and for thinking
'unscientifically'.But Darwin, althoughhe effectively rewrotethe whole
original text of The Origin of Speciesin subsequenteditions, did little to
reducethis elementin his analysis. Darwin's refusal to changehis form
of presentationis particularly surprising becauseit created confusion
qver what he recognisedto be 'the only novelty' in his analysis,that is,
his account of the mechanismsof natural selection. Whereas Darwin
professedthat 'natural selection'was a metaphorfor objective, imper-
sonallawsof nature,many of his readerstook the phraseliterally. One
reasongiven by Darwin for none the less retaining the term 'natural
selection' was 'that it is constantly used in all works on breeding'
(Darwin in Young, 1971a,p. 464). This justification illustratesDarwin's
strong commitment to the perspectiveof the breeders.But there must
havebeenotherconsiderationsinvolved, to accountfor his generaluseof
a voluntaristic vocabularyand his unwillingness to drop his metaphor
onceit had beenshownto confuserather than clarify his argument.One
such considerationwas that Darwin's metaphorenabledhim to avoid
having to demonstratethat domesticand natural variation were equiva-
lent. For this demonstrationwas impossible.The evidenceavailable on
natural variation was insufficient to establish the connection. Thus
Darwin usedthe metaphorof naturalselectionanda terminologyderived
from the work of breedersto bridge over a major gap in his argument.
He began with artificial selection and familiar examplesin order to
convincehis readersthat biological forms could be selectedin accordance
with externalrequirements.He assumedthat in natural settingsbiologi-
cal forms were functionally adapted. And he used the language of
domesticselectionin his analysisof naturalcontextsin order to bring his
readersto makethe 'leap of faith' requiredto seenatural and domestic
adaptationas equivalentprocesses(Young, 1971a).
I havesuggested,then, that the form of argumentusedby Darwin in
his treatmentof naturalselectionwas a consequence of his commitment
to the anthropomorphicperspectiveof commercialbreedersand of his
inability to provide detailed empirical evidence to support a thesis of
which he was personallyconvincedbut which he was unableto demon-
strateformally. It follows from this latter point that Darwin's accountof
the mechanismof naturalselection,the linchpin of his theory, cannotbe
treatedsimply as a summarystatementof regularitiesobservedin nature.
Let us look briefly at how Darwin, and Wallace who independently
reachedthe sameconclusion,arrived at the 'laws of naturalselection'.In
the first place, it is clear that their formulation was not gradually
achievedby the systematicaccumulationof evidence; although it was
precededby mu,ch data-collection.Both Darwin and Wallace had long
had enoughevidenceto be convincedthat naturalevolution occurredand
werelooking for an explanation,when they eachexperienceda dramatic
flash of insight to the effect that Malthus's 'laws' governing human
populationscould be applied with even more force to animal and plant
life. As a result, they suddenlysaw the corpus of availableobservations
from a new perspective.
There is clear evidencehere of informal, indeed subconscious,think-
ing; of the kind of gestalt switch emphasisedby Hanson, Kuhn and
Holton. In this case, a solution to a major interpretative problem in
biology was suddenlyseento be solvableby the displacementof concepts
from another area of cultural activity (Schon, 1963). In order to
understandthis transferof ideas, it is important to realisethat Malthus
was the most widely discussedauthor in Britain in the early years of the
nineteenthcentury.Malthus'sanalysisservedas a basiccultural resource
for numerousareasof intellectual life at that time. Moreover, not only
was his argumentemployedby all the major evolutionarywriters, but it
was usedin a variety of different ways. Malthus himself arguedthat the
growth of human populations would prevent social progress from
continuingbeyonda certainlevel. Paleyusedthe sameargumentto show
that biological specieswere stable. Lyell drew upon Malthus to demon-
stratehow specieswere eliminated. Whilst Darwin and Wallaceused the
same basic idea to account for the origin of species. Malthus had
undoubtedlyhit upon a powerful, multi-purposeinterpretativeformula-
tion. Therewas somethingabout the Malthusiandoctrinewhich fascina-
ted the educatednineteenth-centuryBritish mind and which expressed
what Holton would call 'the style of thought of the age'.
Malthus'sthesisgrew out of the late eighteenth-centurydebateabout
the natureof man andthe future of society. Utopian writers like Godwin
and Condorcethad arguedthat man is infinitely perfectible,that human
reasonis supremeand that completesocial harmonyis attainable.This
conceptionof social progresswas a dominanteighteenth-centurytheme.
Malthus argued,in contrast,that human and social perfectability were
limited by the operationof a basic law of society; namely that whereas
population tends to increasegeometrically, the means of subsistence
increasesonly arithmetically. It follows that population growth will
always outrun subsistenceand will be kept in check only through the
elimination of the 'poor and inept' by the ruthlessagenciesof hungerand
poverty, vice andcrime, pestilenceand famine, revolution and war. This
central part of Malthus's analysis was taken over by Darwin and by
Wallace, and applied to the selectionof biological organismsin natural
settings. The best-adaptedbiological forms were seen as surviving the
struggle for life in the wild, in exactly the same way that the fittest
individuals were thought to survive the rigours of industrialisationin
laissez-jaireBritain or equivalentpressuresin other societies.
Malthus'soriginal argumentwas so influential becauseit madesenseof
someof the disturbingsocial changesaccompanyingindustrialisationand
becauseit explained these events as the inevitable consequencesof
natural law (Sandow, 1938). Many members of nineteenth-century
societywereengagedin a bitter strugglefor existence.Malthus soughtto
show that this had always beenso and always would be so. Thus what
Darwin and Wallacedid was to transposeinto the domain of biological
theory an interpretationwhich had receivedits impetus and its justifica-
tion from considerationof social phenomena.However, they carriedout
this transposition without having the detailed evidence necessaryto
documentthe biological laws which had to be assumedto operateif the
Malthusian interpretation were to be accepted. Darwin was already
convincedof evolutionarychangeand of the equivalenceof domesticand
natural adaptation. What he neededwas a formulation which could
explain adaptationin natural settings and provide a parallel to the
actions of breederson domestic populations. As he had no detailed
evidenceon which to build up a pictureof how 'naturalselection'worked
(it could be presumedthat this lack of evidencewas due to the fact that
the processwas so slow and becausethe fossil record was fragmentary),
Darwin completed his analysis by adopting the ready-made thesis
available in Malthus. But although the Malthusian doctrine could be
redesignedto meet Darwin's requirements,it provided no more than a
generalinterpretativeformula. It is hardly surprising,given its origins in
social debate,that Malthus'sargumentdid not bring Darwin any closer
to specifying, for example, how inheritance is transmitted or how
particular structurescope with environmentalpressures.
In proposingthe theory of evolution by meansof the mechanismof

natural selection he was not really supplying a mechanismat all.
Rather,he was providing an abstractaccountat a generallevel of how
favourablevariationsmight be preserved.He had to keep his account
at a certainlevel of abstraction-since,as he confessed,he could neither
specifythe laws of variation nor the precisemeansby which variations
were preserved.The acceptability of his account dependedon its
plausibility and its ability to explain in very generalterms the sort of
processwhich was involved. (Young, 1971a,p. 469)
Darwin, then, becameconvincedinformally of the parallel betweenthe

selectiveproceduresof breedersand the effects of Malthusian pressures
on speciesin the wild; but he could not show in detail how artificial or
naturalselectionworked. In the formulation of his theory, therefore,he
coveredover the gapsin his knowledgeof Malthusiannatural laws, as he
covered over the assumptionthat domestic and natural vanatlOn are
equivalent,by use of the anthropomorphicmetaphorand the device of
the figurative 'wiser being'. Yet, despite the lack of evidence, the
metaphoricalargumentand the continual weakeningof his major claims
in responseto criticism, Darwin's theory was immenselyinfluential and
highly successfulin gaining widespreadacceptanceof the generalidea of
organicevolution. Two factors seemto have contributedsignificantly to
this success.On the one hand, as Young (1971a) has shown, Darwin's
voluntaristic terminology made it easierto forge a theological position
which could encompasshis conclusions.God came to be seen as the
'wiser being' who had designedthe laws of nature so as to ensurethe
beneficentadaptationof biological forms. But equally important was
Darwin's relianceon the Malthusian argumentwhich Victorians found
so persuasive.As we have seen,the only novel part of Darwin's theory
was his applicationof Malthus'sargumentto accountfor the emergence
of new species(although Darwin's researchtechniqueof acquiring data
from breederswas also exceptional).It appears,therefore,that it was the
combinationof the evolutionaryhypothesiswith the powerful Malthusian
thesis purporting to show how evolution actually worked, which made
Darwin's argumentso convincing to many and a sourceof such dismay
to others.
The heavy relianceon Malthus by Darwin, and to a lesserextent by
other evolutionists, is particularly interesting becauseit seemsto have
beenbasedon certainassumptionsaboutthe natureof their society. This
can be seenmost clearly in the caseof Wallace.In the late 1870s,Wallace
read and was convinced by Henry George'ssocialist interpretationof
society.Wallace cameto believethat voluntary co-operationand reform
were important social forces and could replacestruggleand competition
as the central agenciesof social change.As Wallacechangedhis view of
the basis of social life, so he rejected Malthus's analysis. Malthus's
theory, he cameto see,had no bearing 'whateveron the vast social and
political questions which have been supported by reference to it'
(Wallacein Young, 1969, p. 133). This seemsto show that Malthus was
convincing only so long as it was assumedthat the marked inequality,
the ruthlesscompetition, the absenceof welfare legislation, and so on,
which were characteristicof nineteenth-centuryBritain, were necessary
features of society. Only on this assumptionwas it possible to treat
Malthus'sformulation as a natural law of society whichwas sufficiently
well establishedto be transposedto the biological realm. Thereare some
grounds,therefore,for suggestingtentativelythat the use of Malthusian
doctrineby Darwin and his colleagueswas madepossibleby their sharing
a seriesof backgroundassumptionsabout the natureof social life which
were derived from dominant features of their own society; and that
similar presuppositionscontributedto the influence of Malthus's writ-
ings and to the successof evolutionarytheory.
The caseof Darwinian theory seemsto illustrate most of the main
points madeby Holton. Fundamentalscientific conclusionswere reached
by informal, metaphoricalproceduresof reasoningwhich, despite stren-
uous attempts at formalisation, left distinct traces on the form and
contentof the ensuingknowledge-claims.Scientistsdrew intuitively on
externalinterpretativeresourceswhich were usedto define the natureof
their intellectual problemsand to fill in the gaps in their analysis. The
major interpretationsproposedwerenot baseddirectly on observationof
biological phenomena.Rather they were taken over from practical
activities and from the wider realm of philosophical, theological and
social debateto provide the framework within which observationswere
given their scientific meaning. During the period that Darwinian theory
emergedthereseemsto havebeena body of relatedideas,associatedwith
the themeof struggleand adaptation,which was the main interpretative
resourcefor a wide variety of intellectual endeavours.Wheneverthese
wereadoptedby membersof the emergentbiological researchspecialism,
efforts weremadeto separatethem from their social origins and to devise
formulationsin which, as far as possible,they could be presentedsimply
as descriptionsof observedregularities (Young, 1971a). Consequently
backgroundassumptionsoriginating in specific social relationships(as
betweenDarwin and the breeders)or in broaderfeaturesof society (like
those implied in the Malthusian doctrine), have been gradually hidden
from sight. As a result, most sociologists and many historians have
acceptedat face value Darwin's own Baconianaccountof the epistemo-
logical statusof his conclusions.They havethereforefailed to realisejust
how fragile the observationalbasis of Darwin's theory was and how
much acceptanceof his claims dependedon commitment to shared
presuppositions.Given that Darwin's theory influencedso many aspects
of modernbiology, it may well be that someversionof thesepresupposi-
tions hasbecomebuilt into the technicalculture of biological science(see
Roseand Rose, 1976,ch. 6).
In the pagesimmediatelyaboveI havelooked at one casein sufficient
detail, I hope, to be convincing, insteadof skimming more superficially
through a numberof instances.The critical reader,however, may think
that too much reliance has been placed in this section on the one
example.For the caseof Darwinian theory might be misleadingin two
respects.First, as I pointed out above, the scientific community has
probably becomeincreasingly able to satisfy its own cultural require-
ments during the last hundred years or so; and biological research,in
particular, has undoubtedlybecome more specialised,more technical
and less closely linked to broad social debatethan it was in Darwin's
time. Thusthe Darwinian illustration may seemout of date. Secondly,it
may be that nineteenth-centurybiology was more open to outside
influence (or less separatedfrom what we now see as distinct areasof
discourse)than has ever been true of physics and chemistry, which are
further removedintellectually from the social realm. Thereis a modicum
of truth in both these propositions.The academicresearchcommunity
has been able to achieve a. very considerableautonomy and social
segregation;and the connectionsbetweenscientific thought and social
life havetendedto be moreobviousin the caseof biology than in the case
of chemistry and physics. Nevertheless,in recent years a number of
historical studieshave been completedwhich show that external influ-
enceson the contentof sciencehavebeenconfinedneitherto the distant
past nor to the life sciences(seealso Young, 1971b,and Roseand Rose,
1976,on modernbiology). Let me simply list someof the fields for which
a reasonablecase has been made and refer the reader to the sources
given: thermodynamicsand the conservationof energy in nineteenth-
century physics (Brush, 1967; Elkana, 1974); the theory of relativity
(Feuer, 1971; Holton, 1973); quantumtheory (Forman, 1971; Holton,
1973); and present-dayorganic chemistry(Slack, 1972).
We can be fairly certain, then, that the analysisof Darwinian theory
doesexemplify processeswhich occur in the 'harder' sciencesand in the
presentcentury. It seems,therefore, to be possible to supplementthe
accountsketchedin Chapter3 of the processesof knowledgeproduction
inside the researchcommunity with a more macro-sociologicalanalysis
(this kind of generalapproachis advocatedin Stark, 1958); although,of
course,muchmorecomparativeanalysisis neededthancan be attempted
here.Oneway of linking internal and externalprocessesmay be through
the notion of 'interpretativefailure'. In other words, it seemslikely that
scientiststurn to other cultural areaswhen basic interpretativeproblems
prove particularly difficult to resolve with their existing resources.In
suchsituations,scientistsarelikely to look beyondtheir own community
to other relatively systematicand co-ordinatedbodies of analysisor to
coherentpractical traditions. Although interpretativefailure seemsto
havebeeninvolved in the caseof quantumtheory as well as Darwinian
theory, it does not always seemto be present.The crucial factor in the
caseof organicchemistryappearsto havebeenthe longstandingrelation-
ship betweenthe researchcommunityandthe chemicalindustry.This is a
useful reminder that external demandsexpressinggroup interestsmay
influencehow scientistsconceptualisetheir field, irrespectiveof external
cultural resources(seealso Nelkin, 1977, for a discussionof ecology,and
Johnston,1978, for a more generalview). Moreover, bodies of culture
are carried by social groups and the movementof cultural resourcesis
often mediatedthrough the relationshipsbetweencollectivities. (This is
probably least true of that commonsenseknowledgewhich plays an as
yet mysterious role in scientific thinking.) However, I have by now
establishedthe generalpoint that the content of scienceis affected by
socialandcultural factorsoriginatingoutsidescience.In the next section,
I shall consider the movement of cultural resourcesin the opposite
direction, that is, from the scientific community into the wider society.
We will find that structuredsocial relationshipsand group interestsare
crucially important in understandingthe processesinvolved.
THE USE OF SCIENTIFIC CULTURE IN EXTERNAL POLITICAL SETTINGS

Onemajor concernin the sociologyof knowledgehas beento show how
society at large influences the production of specialisedknowledge. In
the last section I tried to show that scienceneed not be exemptedfrom
this kind of analysis.A secondimportant objective for the sociology of
knowledgehas been to explore how knowledgeis used in the courseof
political activity. I intend now to consider briefly this second theme.
Sociologists' accounts of science and the political process have been
formulated in accordancewith the customary assumptionsabout the
characterof scientific knowledgeand the nature of the scientific ethos
(see Blume, 1974). As a result, scienceand scientistshave been seenas
occupyinga rather specialposition in political life.
In the first placeit hasbeenassumed,with little recognitionof the need
for close empirical documentation,that scientists are the bearersof a
type of knowledgewhich is unaffectedby the social context in which it is
used. 'The theories,models, proceduresand formulae of science... are
generallybelievedto allow one trainedin their usesimply to calculatean
unambiguouslycorrect answer' (Mazur, 1973, p. 251). Secondly, it has
beenacceptedthat the supposednorms of science,such as disinterested-
ness and universalism, require scientiststo act in a politically neutral
mannerand that scientistscontinueto abide by thesenorms outsidethe
researchcommunity(Barber, 1952; Brooks, 1964; seealso Ezrahi, 1971).
Clearly, if thesetwo assumptionsare correct,the increasingparticipa-
tion of scientistsin the political sphereis likely to changedramaticallythe
characterof the political process(Lakoff, 1977); that is, the realm of
'politics and ideology' will be reduced,and perhapseventuallyelimina-
ted, as the range of application of certified scientific knowledge is
extended. Thus'if one thinks of a domain of "pure politics" where
decisionsaredeterminedby calculationsof influence, power, or electoral
advantage,and a domain of "pure knowledge" where decisions are
determinedby calculationsof how to implementagreed-uponvalueswith
rationality and efficiency, it appears... that the political domain is
shrinking and the knowledgedomain is growing ... ' (Lane, 1961, pp.
61-2; seealso Bell, 1960). Given the traditional sociologicalconception
of science,this line of argument,sometimescalled the 'end of ideology'
thesis, is quite reasonable.The assumptionson which it is basedare,
however, clearly inconsistent with the approach which I have been
developing. Let me try to show, therefore, where the argument is
inadequateand how we can depict more accuratelythe way in which
scientistsemploy their cultural resourcesin the political context.
In the traditional analysisof scienceand politics which has led to the
'end of ideology' thesis, scientific culture enters in the form of the
standardset of social norms and in the form of context-freeknowledge.
Let me discussthe social norms first. Thesenorms have typically been
conceived,as we have seen,as a set of rules unambiguouslyspecifying
certainkinds of socialaction. In the realm of political analysis,they have
been interpretedas requiring scientiststo adopt a disinterested,politi-
cally neutral, concernfor objective data. Active engagementin politics
hasbeenconsideredas alien to scientistsand as 'essentiallydestructiveof
scientific endeavour'(Haberer,1969, p. 1). However, we have seenthat
the normative resourcesof the scientific community are by no means
limited to those which have been customarily acceptedas defining the
scientific ethos.Furthermore,we haveseenthat the norms of scienceare
best conceived, not as clear prescriptionsspecifying certain kinds of
action, but as standardisedverbal formulations from which scientists
select in order to establishinterpretationsof their actionsappropriateto
particularsocial contexts.There are now a number of historical studies
availableof the connectionsbetweenscienceand government,particu-
larly in the USA, which are consistentwith this revisedview of scientific
culture. They showthat when scientistshaveenteredthe political context
they have drawn selectivelyon their cultural repertoirein a way which
has furtheredtheir collective interests.They haveusedtheir resourcesto
constructa political image or ideology particularly suited to the Ameri-
can 'democratic'setting.
In the first half of the nineteenthcentury, American scientistsdid not
presenta uniform image of sciencein their dealingswith laymen. This
seemsto have been becausethey were involved with a variety of lay
audiencesandas a result producedvarying accountsof what sciencewas,
what scienceought to be and what part scienceplayedin society. As the
century progressed,however, and as the scientific community increas-
ingly emergedas a separatesocial entity, a more coherentview beganto
crystallise.
Previously,sciencehad been'sold' to the public in termsof its contri-

bution to important American values-utilitarian, equalitarian,
religious-or even as a meansof social control, dependingupon the
speaker'sbest estimateof his audience.But in the 1870'sfor the first
time, great numbersof scientific spokesmenbeganto vocally resent
this dependenceupon valuesextraneousto science.The decade,in a
word, witnessedthe development,as a generally sharedideology, of
the notion of sciencefor science'ssake.(Daniels,1967,p. 1699)
This ideology was further elaboratedand formalised, over the years,

particularly in the courseof political negotiationover the provision of
support for science.
From the late nineteenthcentury until the presentday, one of the
crucial factors influencing scientists'relationswith governmenthas been
the cumulativeincreasein the size of the scientific community and in the
costofresearch.Increasingly,scientistshavecome to recognisethat only
central governmentcan provide funds on a scale sufficient to maintain
the pursuit of scientific knowledge (Price, 1963; Sklair, 1973). At the
same time, scientists have striven vigorously, and with considerable
success,to maintain what Gouldner calls the 'functional autonomy'
achievedby academicor 'pure' scienceduring the last century. In the
courseof their negotiationswith government,scientistshave arguedfor,
and have been granted, extensiveand increasing support in terms of
researchfunds, educationalfacilities and personnel,coupled with mini-
mal regulation from outside-particularlywith respectto control over
the elite activity of pure research.The argumentsemployedby scientists
in these negotiations have come to take a fairly standardisedform
(Greenberg,1969; Tobey, 1971; Haberer, 1969; Weingart, 1970). It has
beenargued,not only that scientific knowledgeis intrinsically valuable,
but also that, becauseit is the only truly valid type of knowledge, it
necessarilyleadsto practical benefit. Scienceis depictedas being unique
in its cumulativeacquisition of unquestionablefacts; which are obtain-
able only so long as scientistsareallowed to approachthe study of nature
with values which curb human tendenciestowards bias, prejudice and
irrationality. Thesevalues are describedby scientistsin terms, such as
independence,emotional discipline, impartiality, objectivity, a critical
attitude, and so on, which are virtually identical to those used in the
customarysociological analysis. Moreover, the parallel does not end
there. For scientists themselveshad argued explicitly, several decades
before sociologistsof sciencedid so, that scienceand democracywere
especiallycompatibleowing to their pursuit of commonvaluesand their
common recognition of the need for scientific autonomy (Mulkay,
1976a). For example, during the 1920s and 1930s in the USA, an
influential scientific pressuregroup actively sought to gain acceptance
within governmentand among the general public of the notion that
'American democracyis the political version of the scientific method'
(Tobey, 1971, p. 13).
When sociologists first began the empirical study of the scientific
community, they seem to have taken such public pronouncementsby
scientists more or less at face value and to have modelled their own
analysisupon them. Sociologistsseemto have beennoticeablylacking in
scepticism towards the public statementsof scientists, comparedwith
those of other groupings-probably because sociologists accepted
without questionthe covert epistemologyfrom which scientists'accounts
drew much of their strength. In the light of the analysis in previous
chapters,however, we can see that scientists' typical account of the
nature of scienceand scientific values is quite inadequate.We cannot
accept, therefore, that these standardisedformulations were used by
scientistsin the political context simply becausethey were accurateor
were the only ones available. Given that scientistscould have derived
quite different, yet still entirely plausible, accountsfrom their cultural
repertoire, it seemsreasonableto concludethat this particular version
was chosenbecauseit helpedscientiststo achievetheir central objectives
in the political context of North America. For if scientists select
descriptionsand justifications from the availablevocabularyin accord-
ancewith their interestsand the natureof the social context within their
own community, that is, when dealingwith personswho have first-hand
experienceof the social world of science,there is every reasonto expect
that they will do likewise in the courseof their interactionwith laymen,
who will find it even more difficult to challengetheir accounts.It is
certainlyclearthat scientistshaveusedtheir standardportrayalof science
to justify political claims (Greenberg,1969). Thus the epistemological
elementhas beenusedto justify increasedsupport for science:'Science
must be kept healthy becauseit is the only sourceof valid knowledge.'
And the descriptionof the supposedvalues of sciencehas beenused to
justify its continuedautonomy: 'Sciencehasa clear codeof ethicswhich
will only be disruptedby outsideintervention.Suchinterventioncan only
lead to the distortionof scientific resultsand eventuallyto economicand
military decline.' This selectivecharacterisationof science,by scientists,
in the political context amounts to the creation of a professional
ideology.
So far in this section Ihaveoffered no more than the briefest sketchof
the implications for the political analysis of science of my previous
discussionof its normative culture (for further analysis, see Mulkay,
1976a).If this interpretationis broadly correct, it meansthat the special
compatabilityof scienceand democracyis little more than a conception
createdby scientiststhemselvesfor their own practical purposes.It also
meansthat the 'endof ideology' thesisdoesnot apply when scientistsare
involved in political negotiation about science itself. In this latter
context, scientists seem to resemble other groups in pursuing their
sectionalinterestsandin developingan ideologyin orderto improvetheir
chancesof success.But what of situationswherescientists'own interests
as a community are not directly involved and where scientiststake part
merely as providers of certified knowledge?It appearsto follow from
earlier argumentsabout the context-dependence of knowledge-claims
that scientists'useof their technicalrepertoirewill not differ in principle
from their useof social formulations; that is, the way in which scientists
linterpretand draw on their expertiseoutsidethe researchcommunitywill
vary with the social setting in which they are operatingand with their
position in that setting.
In recentyearsseveraldetailedstudieshavebeencarriedout of the use
madeof scientific knowledgein the courseof political qebate.The main
conclusionof thesestudiesis that scientific knowledgedoes not reduce
the scopeof political action, but rather it becomesa resourcewhich can
be interpretedin accordancewith political objectives. This is clear, for
example,in Nelkin's studyof the CayugaLake controversy(1971; 1975).
The central scientific issuein this casewas the environmentaleffect of
building a nuclear power plant at Cayuga Lake. The State Electric
Company,after it had been challengedon this issue by local scientists,
spentover 1Y2 million dollars on research.The recipientsof this money
producedresultswhich were interpretedas providing clear refutation of
previous criticisms and as entirely vindicating the company'soriginal
plan. Various local groups, however, still refused to accept the com-
pany'sproposal,raisednew problemsand continuedto producealterna-
tive scientific analyses.Participantswereno moreableto agreeabout the
'scientific realities' after five years of debatethan they had been at the
outset and the dispute ended with the political defeat of the Electric
Company,despiteits massiveexpenditureon research.
The usemadein this caseof scientific knowledge-claimsseemsnot to
be unusual; nor is the selective use of such claims confined to issues
which fall within the domain of 'immature' scientific disciplines (King
andMelanson,1972; Nelkin, 1975).Opposingpartiesin political disputes
involving technical issues can usually obtain the services of reputable
scientistswho will provide data to buttresstheir policy and to undermine
that of their opponents(Benveniste,1972). By now it is clear why this is
possible.We haveseenthat the formulation of scientific facts dependson
prior commitmentsof various kinds, that thesecommitmentsare often
madein accordancewith participants'positionin a specificsocialsetting,
and that they influence the informal acts of interpretation which are
essentialto give meaningto observations.We would expect, therefore,
that scientistsoccupyingdiffering positions in a political context would
often bring different presuppositionsto bear and that their informal
reasoningwould be subtly influencedby the assumptionsof the group to
which they were affiliated. This view receives support from the few
studiesat presentavailable. Thesestudies indicate that scientists'entry
into the political arena affects their interpretation of their technical
culturein threeways: it influencestheir definition of technicalproblems;
it influences the choice of assumptionsintroduced in the course of
informal reasoning;and it subjectsscientiststo the requirementthat their
conclusionsbe politically useful.
The selective definition of problems is particularly obvious in the
CayugaLake controversy. In this case scientists defined the technical
issues in various different ways which, although they may have been
reconcilablein principle, were not in fact reconciled during the debate
but were instead used to generatedifferent technical programmes,to
'justify' different policies and to support different interests. For
example, in one scientific report, sponsoredby the Electric Company,
the scientific problem was defined as involving only the lake drainage
areaand the conclusionwas reachedthat the environmentaleffect of the
power plant would be negligible. Somescientistsin the local community,
however, maintained that the total parameterof the lake had to be
consideredand that when this was done the conclusionsof the company-
sponsoredresearchabout thermal pollution had to be rejected. Other
scientiststhoughtthat this was too restricteda framework and urged that
the input of the power station be consideredas part of the total lake
system. This perspective,once again, gave rise to a different scientific
analysisand to different practical implications. Mazur's (1973) examina-
tion of the debatesover the effects of radiation and fluoridation upon
public health illustrates the sametendencyfor scientiststo define issues
differently and consequentlyto reach divergent conclusions. Mazur
shows that in both these debatessome scientistsconcernedthemselves
primarily with acute poisoning and accordingly judged the risks from
radiation or fluoridation to be low; whilst others thought in terms of
chronic poisoning,the risk of which appearedto be significantly higher.
As in the CayugaLake example,the lengthy and critical nature of these
disputesas well as the failure to reconciledifferencesshowsthat scientists
became remarkably committed to these narrow definitions of the
problem.
Scientists' choice of a particular definition of a technical problem
cannotitself be decidedby observationand systematicinferencealone.
Rather, it precedesand is presupposedin observation and analysis.
Moreover,the choiceof such a definition in thesepolitical debatesoften
seemsto havedependedon a prior social commitment(seeNelkin, 1971,
and Mazur, 1973). Thus in the studies already mentionedit seemsthat
thosescientistswho spokeon behalf of the Water Authorities, the AEC
or the Electric Company,choseperspectiveswhich defined the issuesin a
manner favourable to their patrons' policies. Similarly, those who
representedopposinginterestsworked from quite different definitions
which enabledthem to reinterpret and to challenge their adversaries'
conclusions,and to defendtheir own collectivity from what they saw as a
threat.
Mazur notes that it is possibleto treat thesedisputesas arising from
failures of communicationwhich could perhapsbe remediedby showing
both partiesthat they weredefining the problemdifferently and that they
were,therefore,arguingat cross-purposes. He stresses,however,that this
seldom, if ever, happensin practice. It appearsto be very difficult for
scientistsengagedin public debateto adopt a Mannheimianstrategyand
to redefinetheir problemsat a higher level where apparentlydivergent
views can be reconciled. It is much more usual for differences in
underlyingdefinitions and in basicpremisesto becomeobscuredor to be
ignoredas both sidespresenttheir findings as 'the definitive facts' about
radiation or whatever. Mazur and Nelkin have shown that the various
parties in any particular dispute tend to use an identical pattern of
rhetoric. As in informal negotiation within the researchcommunity,
observationalclaims are embeddedin a series of highly standardised
formulations of a social kind, which are usedto discredit one'soppon-
ents and to strengthenone'sown assertions.Clearly neither side in this
kind of debateis engagedin formal scientific demonstrationwithin a
commonscientific framework. Both are engagedin informal negotiation
and are seekingto substantiateconclusionswhich areappropriateto their
social commitments.
In addition to influencing scientists'definition of the problem, social
commitmentsinfluence the detailed processesof scientific reasoning.
Mazur observesthat:
... complex technical problems of the state-of-the-artrequire subtle

perceptionsof the sort which cannot be easily articulatedin explicit
form. When it is necessaryto make a simplifying assumption,and
many are reasonable,which simplifying assumptionshould be made?
When data are lacking on a question, how far may one reasonably
extrapolatefrom data of other sources?How trustworthy is a set of
empirical observations?These questionsall require judgementsfor
which there are no formalised guides and it is here that experts fre-
quentlydisagree.(1973, p. 251)
Not only do scientistsdisagree,but they sometimesintroduce non-

technicalresourcesfrom the political setting into their analysesin order
to resolvetheir interpretativedifficulties. For instance,Mazur discusses
scientists'treatmentof the relationshipbetweenlow-level radiation and
the incidenceof leukaemia.He shows that severalmodelsor interpreta-
tions have been formulated which appear to be consistent with the
availabledata. The elementof judgementwhich is essentialin choosing
betweenthesemodels seemsfrequently to be dependenton, or at least
associatedwith, scientists'views about public health policy.
This theoretical ambiguityhas major implications for the technical

debateover permissibleradiation standards.It should be noted that
the 'threshold'model implies that doselevels below the thresholdwill
not harm the population (through leukaemia). The 'linear' model
implies that therewill be someincidenceof leukaemiano matter how
low the dose to the population. The two models differ, then, on
whether or not there is a 'safe' level of radiation exposurefor the
population. The ambiguousnature of the dose-effect curve is well
recognisedby radiation biologists, and many assume the 'linear'
model, not necessarilybecausethey considerit true, but becauseit is
the most conservativemodel for purposesof public safety. (Mazur,
1973, p. 254)
In sucha situation,one possibleoption is that of admitting openly that
no clear scientific conclusionis possible.But scientistshave not normally
acted in this way. One reason for this may be that scientistsenter the
political context as purveyorsof certified knowledge.They have nothing
to offer otherthan the supposedcertaintiesof science;and if they were to
presenttheir conclusionsas no more than 'plausible guesses'basedon
uncertain foundations, they would carry little political weight. More-
over, scientistsare expectedto and are usually inclined to presenttheir
conclusionsin termsof the formal calculusof science.But in order to do
this, scientistshave to commit themselvesto particular informal judge-
ments which subsequentlybecomehidden from view behind an imper-
sonal terminology and a rhetoric which presentsscientific findings as
objective representationsof the external world.
For instance, it is generally agreed among scientists that there is a
connectionbetweenexposureto ionising radiation and the occurrenceof
leukaemiaand thyroid cancer. The evidenceof a connectionbetween
radiation and other forms of cancer,however,is believedby many to be
less compelling. There is, therefore,a considerablerangeof opinion on
this issue, each opinion being supportedby differing estimatesof the
reliability of particular experimentsand observations,by varying inter-
pretationsof the causal processesinvolved, and so on. In the political
context, the important question is whether given levels of radiation
constitutea high or low risk to public health. Scientistshave responded
to this questionby calculatingthe numberof cancercasesper year which
will be produced by low-level radiation. The figures produced have
varied enormously,at least partly becausethose responsiblefor these
calculations have begun from different basic judgements about the
relationshipbetweenradiation and the various types of cancer. At one
extreme there have been a few vocal critics of the Atomic Energy
Commission,who haveassertedas a 'scientific law' that all major forms
of cancerare producedby radiation and who have consequentlycalcula-
ted the risks of radiation to be very high. At the other end of the
spectrumthere havebeenscientists,such as thoseworking for the AEC,
claiming that this estimate 'does not square with the facts' and that
calculations should begin from the established premise that only
leukaemiaand thyroid cancerare causedby radiation.
What has happenedhere and in other instancesis that scientistshave
quantified and applied formal techniquesof inferenceto their informal
judgements;they havepresentedthe resultsas incontrovertiblefacts; and
they have sought to persuadethe uncommittedby using the social and
technicalrhetoric of their professionalcommunity. All areasof scientific
researcharecharacterisedby situationsin which the establishedtechnical
culture permits the formulation of severalreasonablealternatives,none
of which canbe shownconclusivelyto be morecorrect than another.It is
in making choices between such alternatives, whether at the level of
broad definitions of the problem or at the level of detailedanalysis,that
scientists' political commitments and the pressures of the political
context come into play most clearly.
In this section I have tried to show that when scientistsenter social
contexts outside the researchcommunity, such as the wider realm of
political activity, they selectfrom and reinterprettheir cultural resources,
both technical and social, in responseto the social context and in
accordancewith their position in it. This brief and incompletediscussion
of the movementof cultural resourcesout of the researchcommunity
supplementsthe prior examinationof externalinfluenceson the content
of science.Thesetwo sectionscompletethe preliminary analysisoffered
in this book by showing that the processesof negotiationand cultural
reinterpretationwithin the researchcommunity are linked to similar
processesoccurring in society at large. They show that scientific know-
ledge must be seen as one part of a complex movement of cultural
resourcesthroughout society-amovementwhich is mediatedthrough
and mouldedby the changingpatternof social relationshipsand by the
clash of group interests.
BRIEF SUMMARY AND CONCLUDING REMARKS

I beganthis book by showingthat, becausesociologistshavecustomarily
regardedscientific knowledgeas having a specialepistemologicalstatus,
they havetreatedthe productionand legitimation of scientific formula-
tions as a specialcasewithin the sociologyof knowledge.The contentof
scientific knowledge has been excluded from sociological analysis
becauseit has beenassumedthat scientistshave found ways of ensuring
that their conclusionsare determinedby the natureof the physicalworld
itself. Oncesociologistshad acceptedthat scienceprovided an objective
representationof the external world, it appearedto follow that the
scientific communityhadto havecertain distinctive characteristics;for it
seemeddifficult to conceivehow objective knowledgecould be contin-
ually generatedby a communitywhich did not havethesecharacteristics.
In particular, the scientific researchcommunity was seenas necessarily
having an intellectually open and universalisticnormative structure, as
being politically neutral and as operating most effectively in societies
which allowed scienceconsiderableautonomy.
The standardphilosophical view of science, which sociologists of
knowledge took for granted in treating science as a special case,
furnisheda fairly coherentaccountof scientific observation,the relation
betweenfact andtheory, the uniformity of natureandthe criteria usedto
validate scientific knowledge-claims.In Chapter2 I tried to show that
this traditional philosophicalanalysiswas besetby gravedifficulties and
I tried to sketchthe broadoutline of an alternativeview which grows out
of recent philosophicaldebate.I arguedfor example,that the principle
of the uniformity of nature is best seen, not as an assumptionthat
sociologiststhemselveshaveto makeaboutthe physicalworld, but rather
as a part of scientists'resourcesfor constructingtheir accountsof that
world. I also argued that the factual content of scienceshould not be
treatedas a culturally unmediatedreflection of a stableexternalworld.
Fact and theory, observationand presupposition,are inter-relatedin a
complexmanner;and the empirical conclusionsof sciencemust be seen
as interpretativeconstructions,dependentfor their meaning upon and
limited by the cultural resourcesavailableto a particularsocial groupat a
particularpoint in time. Similarly, generalcriteria for assessingscientific
knowledge-claimscannotbe applied universally, independentlyof social
context,as most sociologistshavepreviouslyassumed.Thesecriteria are
always open to varied interpretationsand are given meaningin terms of
particular scientists' specific intellectual commitments,presuppositions
and objectives.In short, I arguedthat the cognitive/technicalresources
of scientistsare open to continual change of meaning; that there is,
therefore, nothing in the physical world which uniquely determines
scientists' conclusions; and that consequently it is appropriate for
sociologiststo look closely at the ways in which scientistsconstructtheir
accountsof the world and at the ways in which variations in social
context influence the formation and acceptanceof scientific assertions.
In the third chapter I showed that the longstanding sociological
analysisof the normativestructureof sciencewas inadequate,quite apart
from its inconsistencywith the revisedphilosophicalpositionproposedin
Chapter2. I suggestedthat what had beentaken by sociologiststo be a
completeset of basic principles specifying proper conduct for scientists
engagedin researchshould be seenas no more than part of a complex
social repertoirewhich scientistsuse flexibly in the courseof negotiating
the meaningof their own and their colleagues'actions. In addition, I
examineda numberof recent casestudieswhich appearedto show that
there is no clear separationin sciencebetweenthe negotiationof social
meaning and the assessmentof knowledge-claims. Both social and
cognitive/technicalformulations have to be selectedand interpretedby
participantsin particular instances;and both kinds of resourceacquire
their specific meaningas they are combinedin the sequenceof informal
interactionplus formal demonstrationwherebyknowledge-claimscome
to be ratified.
Oncewe haveabandonedthe orthodox philosophicalview of science,
it becomes possible to acceptthat the social negotiationof knowledgein
various kinds of scientific researchcontext is a legitimate, indeed a
fundamentaland as yet largely unexplored,topic for the sociology of
knowledge. But it must never be assumedthat the production of
knowledgeby scientific specialistsis entirely divorced from the wider
social and cultural context. It is in recognisingthis that the argumentI
have advancedhere links up with the more traditional concerns of
sociologistsof knowledge. In the final chapter I have used the caseof
Darwiniantheoryto show in detail how cultural resourcesfrom societyat
largecanenterinto the very form and contentof scientific claims as well
as, possibly,playing a significant part in their acceptance.Sometimesthe
connectionbetweenscienceand thewider societyis establishedby means
of direct social contact betweenscientistsand outsiders;but it can also
arise in a more diffuse way through scientists'ability to select from and
reinterpretcultural resourcesgenerallyavailableduring a specific period
to the membersof a particular society or a particular social class.
In this last chapter I also consideredbriefly the implications of my
overall thesisfor anothercentral concernof the sociologyof knowledge,
namely, the political actions of knowledge-producers.I suggestedthat,
as one would expect in the light of my previous argument, scientists'
knowledge-claimscan be affectedby their position in a political context
and that elementsof a political contextmay become builtinto scientists'
assertionsabout the natural world. I also tried to show that the growing
engagementof scientistsin the political arenain no way signalsthe end of
political ideology, as had been widely assumed.I argued instead that
scientists'own claim to be politically neutral was itself ideological, in the
sensethat it constitutesa selectiveemploymentand interpretationof the
cultural resourcesavailable to scientists, in a way which favours the
vestedinterestsof their specialisedcommunity.
The argumentpresentedin this book and summarisedimmediately
aboveopensup numerouslines of empirical researchand analysiswhich
havepreviously beenlargely ignored by sociologists.For instance,once
we ceaseto take it as self-evident that scientific knowledge-claimsare
assessedby clear-cut, pre-establishedcriteria, it becomes possible to
acceptthat non-technicalconsiderationsmay systematicallyinfluencethe
allocation of scientific credit. Consequently,it becomes possible to
approachthe study of social ranking in sciencein a radically new way
and to explore for the first time how far the phenomenaof power and
dominationare presentwithin the researchcommunity. This possibility
was discussedat the end of the first chapter. In the secondchapter,one
of the most interestingideas to emergewas that scientific propositions
are not stablein meaning,but are reinterpretedas they move from one
social context to another.This processof reinterpretationis at present
little understoodand is in needof detailedsociologicalinvestigation.For
example,it would be particularlyhelpful to havestudiesof the variations
introducedinto specificgeneral-purpose formulationsin variouscontexts
as well as careful analysisof how thesevariations were devisedto meet
the requirementsof thesecontexts.
In Chapter3, the central theme wasthat of the social negotiationof
scientific knowledge.As I stressedin that chapter,my treatmentof this
topic left manygapsand, therefore,manyavenuesfor further study. One
important point which did become clear was that the evaluative
repertoire of scientists is much more complex and extensive than
sociologistshavepreviouslyassumed.Accordingly, if we are to achievea
better understandingof the nature of social negotiation in science,we
must havemore studieslike that of Mitroff which provide wide-ranging
and detaileddocumentationof the moral languageof science.But this
alone, althoughessential,is not enough.For we must also find various
ways of ascertaining just how the repertoire of social evaluations
contributesto the interpretationand acceptanceof specific knowledge-
claims.
In this final chapter the main theme has been the movement of
interpretativeresourcesinto and out of the researchcommunity. A few
political scientistssuchas Nelkin and Mazur have alreadybegunto study
how scientists adapt their specialised knowledge to the pressuresof
political debate.Such studies are, however, still few in number and I
hope that this book will convince some sociologists that this is a
potentially fruitful realm of inquiry. But an equally important and
relatedtopic, which hasbeencompletelyignored by social scientistsuntil
very recently,is that of the influenceson scientific knowledgeoriginating
outside the social networks within which scientific researchis carried
out. Thus it has becomeclear in the courseof the presentchapterthat
oneof the major tasksnow facing sociologistsof knowledgeis to portray
the dynamic social processeswhereby scienceabsorbs,reinterpretsand
refurbishesthe cultural resourcesof modern industrial societies.From
the new perspectiveadvancedhere, scienceshould not be treated as a
privileged sociological case and kept separate from other areas of
cultural production. Rather, every effort should be made to investigate
scientists' debt to the wider society and to delineate the complex
connectionsbetweencultural productionin scienceand in other areasof
social life.
Theseare just a few of the more obvious topics for further research
which grow out of the analysispresentedin this book. In addition there
are numerousissueswhich are less easily perceived, yet which require
careful examinationin the light of this new conceptionof the sociology
of science.Perhapsthe most important of theseis the question of the
relationshipbetweenscienceand technology.From the standardview of
sciencethis relationshipis relatively unproblematic.For effective techno-
logy is seenas a simple by-product of objective knowledge. But if we
stress the socially and culturally contingent character of scientific
knowledge,we must be preparedto questionthe widespreadassumption
that moderntechnologyis on the whole a derivative of basic scientific
researchand/or to move towards an analysis of the social meaning of
technology. I do not intend to pursue this kind of complex issue any
further here, however. I hope that many readers will do this for
themselvesand that they discover in the text various interestingtopics
and unresolved issues worthy of systematic study that I have not
identified explicitly. If this happens,this book will haveachievedone of
its central aims, namely, that of helping to make the study of sciencea
lively region within the sociology of knowledge.
References
Agassi, J. (1963), 'Towards an historiographyof science', History and Theory, vol. 2,
pp. 1-23.
Barber,B. (1952), Scienceand the SocialOrder (New York: The FreePress).
Barnes,B. (1974), Scientific Knowledgeand Sociological Theory (London: Routledge&
KeganPaul).
Barnes,B., and Dolby, R. G. (1970), 'The scientific ethos:a deviant viewpoint', European
Journal ofSociology,vol. 2, pp. 3-25.
Barnes,B., and Law, J. (1976), 'Whatevershould be done with indexical expressions?',
TheoryandSociety,vol. 3, pp. 223 -37.
Bell, D. (1960), TheEndofIdeology(Glencoe,III.: The FreePress).
Ben-David,J. (1968), FundamentalResearchand the Universities(Paris:OECD).
Ben-David,J. (1971), TheScientist'sRolein Society(EnglewoodCliffs, NJ: Prentice-Hall).
Ben-David. J. (1977), 'Organization,social control, and cognitive changein science',in
J. Ben-David and T. N. Clarke (eds), Culture and Its Creators (Chicago and London:
University of ChicagoPress).
Benveniste,G. (1972), ThePolitics ofExpertise(Berkeley,Calif.: GlendessaryPress).
Berger, P., and Luckmann,T. (1967), The Social Constructionof Reality (London: Allen
Lane).
Blissett, M. (1972), Politics in Science(Boston, Mass.: Little, Brown).
Bloor, D. (1976),KnowledgeandSocialImagery(London:Routledge& KeganPaul).
Blum, A. (1970), 'The corpusof knowledgeas a normative order',in J.C. McKinney and
E. A. Tiryakian (eds), TheoreticalSociology(New York: Appleton-Century-Crofts).
Blume, S. S. (1974), Toward a Political Sociology of Science(New York and London:
Collier Macmillan).
Bohm, D. (1965). The Special Theory of Relativity (New York: Benjamin).
BOhme,G. (1975), 'Thesocialfunction of cognitivestructures'.in K. Knorr, H. Strasserand
H. G. Zilian (eds), Determinantsand Controls of Scientific Development(Dordrecht,
Holland: Reidel).
Bohme, G. (1977), 'Cognitive norms, knowledge-interestsand the constitution of the
scientific object', in E. Mendelsohn,P. Weingart and R. Whitley (eds), The Social
Production of ScientificKnowledge(Dordrecht, Holland: Reidel).
Bohme,G., Daele,W. van den, and Krohn, W. (1972), 'Alternativenin der Wissenschaft',
Zeitschriftfar Soziologie,vol. 1, pp. 302-16.
Bohme, G., DaeIe, W. van den, and Weingart, P. (1976), 'Finalizationin science',Social
ScienceInformation, vol. 15, pp. 307-30.
Borger, R., and Seaborn,A. E. (1966), The Psychologyof Learning (Harmondsworth:
Penguin).
Bourdieu, P. (1975), 'The specificity of the scientific field and the social conditionsof the
progressof reason',Social ScienceInformation, vol. 14, pp. 19-47.
Braverman,H. (1974), Labor and Monopoly Capital (New York and London: Monthly
Review Press).
Brooks, H. (1964), 'The scientific advisor',in R. Gilpin and C. Wright (eds), Scientistsand
National Policy Making (New York: Columbia University Press).
Bruner, J. S. (1974), Beyondthe Information Given (London: Allen & Unwin).
Brush, S. G. (1967), 'Thermodynamicsand history', The Graduate Journal, vol. 7,
pp.477-564.
Bunge, M. (1967), 'Technologyas applied science', Technologyand Culture, vol. 8, pp.
329-47.
Burtt, E. A. (1924), The MetaphysicalFoundationsof Modern Science(London: Rout-
ledge& Kegan Paul).
Carnap, R. (1939), Foundations of Logic and Mathematics (Chicago: University of
Chicago Press).
Carnap, R. (1966), PhilosophicalFoundationsof Physics(New York: Basic Books).
Chubin, D. E. (1976), 'The conceptualizationof scientific specialties', The Sociological
Quarterly, vol. 17, pp. 448-76.
Cole, J. R., and Cole, S. (1973), Social Stratification in Science(Chicago and London:
University of Chicago Press).
Collins, H. M. (1974), 'The TEA set: tacit knowledgeand scientific networks', Science
Studies,vol. 4, pp. 165-85.
Collins, H. M. (1975), 'The sevensexes:a study in the sociologyof a phenomenon,or the
replication of experimentin physics,' Sociology,vol. 9, pp. 205-24.
Collins, H. M., and Harrison, R. G. (1975), 'Building a TEA laser: the capricesof com-
munication', Social Studiesof Science,vol. 5, pp. 441-50.
Collins, H. M., and Pinch, T. J. (1978), 'The constructionof the paranormal: nothing
unscientificis happening',in R. Wallis (ed.), RejectedKnowledge:SociologicalReview
Monograph (Keele: University of Keele).
Curtis, J. E., and Petras,J. W. (1970), The SociologyofKnowledge(New York: Praeger).
Daele, W. van den, Krohn, W., and Weingart, P. (1977), 'The political direction of scien-
tific development', in The Social Production of Scientific Knowledge (Dordrecht,
Holland: Reidel).
Daniels, G. H. (1967), 'The pure-scienceideal and democraticculture', Science,vol. 156,
pp. 1699-705.
Darwin, C. (1859), On the Origin of Species(London: Murray).
De Gre, G. (1955), Scienceas a Social Institution (New York: RandomHouse).
Deutsch, M. (1959), 'Evidence and inference in Nuclear research',in D. Lerner (ed.),
Evidenceand Inference(Glencoe,Ill.: The Free Press).
Duhem, P. (1962), The Aim and Structureof Physical Theory (New York: Atheneum).
Durkheim, E. (1915), The ElementaryForms of the Religious Life (London: Allen &
Unwin).
Durkheim, E. (1938), The Rulesof SociologicalMethod (Glencoe,Ill.: The Free Press).
Easlea,B. (1973), Liberation and the Aims of Science(London: Chatto & Windus).
Edge, D.O., and Mulkay, M. J. (1976), Astronomy Transformed(New York: Wiley
Interscience).
Elias, N. (1971), 'Sociology of knowledge: new perspectives',Sociology,vol. 5, pp. 149-
68, and 355-70.
Elkana, Y. (1974), The Discovery of the Conservationof Energy (London: Hutchinson
Educational).
Elliot, H. C. (1974), 'Similarities and differencesbetweenscienceand common sense',in
R. Turner (ed.), Ethnomethodology(Harmondsworth:Penguin)pp. 21-6.
Ezrahi, Y. (1971), 'The political resourcesof American science',ScienceStudies,vol. 1,
pp. 117-33.
Fay, B. (1975), Social Theory and Political Practice (London: Allen & Unwin).
Feuer,L. S. (1971), 'The social roots of Einstein'stheory of relativity', Annalsof Science,
vol. 27, pp. 277-98, 313-43.
Feyerabend,P. (1975),AgainstMethod(London: NLB).
Fisher, R. A. (1936), 'Has Mendel'swork beenrediscovered?',Annalsof Science,vol. I,
pp.115-37.
Forman, P. (1971), 'Weimar culture, causalityand quantumtheory, 1918-1927:adapta-
tion by Germanphysicistsand mathematiciansto a hostile intellectual environment',in
R. McCormach(ed.), Historical Studiesin the Physical Sciences,No.3 (Philadelphia
Penn.: University of PennsylvaniaPress).
Fox, R. (1974), 'The rise and fall of LaplacianPhysics',Historical Studiesin the Physical
Sciences,vol. 4, pp. 89-136.
Frake, C. O. (1961), 'The diagnosis of diseaseamong the Subanun of Mindanao',
AmericanAnthropologist,vol. 63, pp. 113-31
References 125
Frank, G. P. (ed), (1961), The Validation oj Scientific Theories(New York: Collier).
Frankel, E. (1976), 'Corpuscularoptics and the wave theory of light: the science and
politics of a revolution in physcis', Social Studiesoj Science,vol. 6, pp. 141-84.
Galileo, G. (1953), A Dialogue Concerningthe Two Chiej World Systems,ed. Santillana,
G. de (Chicago: University of Chicago Press).
Gaston,J. (1973), Originality and Competitionin Science(Chicagoand London: Univer-
sity of ChicagoPress).
Giddens,A. (1978), 'Positivism and its critics', in T. B. Bottomoreand R. Nisbet (eds),
A History oj SociologicalAnalysis(New York: Basic Books).
Gilbert, G. N. (l976a), 'The transformation of researchfindings into scientific knowledge',
Social Studiesoj Science,vol. 6, pp. 281-306.
Gilbert, G. N. (1976b), 'The developmentof scienceand scientific knowledge:the caseof
radarmeteorresearch',in G. Lemaine(ed.), Perspectiveson the Emergenceoj Scientific
Disciplines(The Hagueand Paris: Mouton).
Gilbert, G. N. (1977a),'Competition,differentiationand careersin science',SocialScience
Injormation, vol. 16, pp. 103-23.
Gilbert, G. N. (1977b), 'Referencingas persuasion',Social Studiesoj Science, vol. 7,
pp. 113-22.
Glass,B. (1953), 'The long neglectof a scientific discovery: Mendel'slaws of inheritance'
in G. Boas et al. (eds), Studiesin Intellectual History (Baltimore, Maryland: Johns
Hopkins Press).
Gouldner, A. W., The Coming Crisis oj WesternSociology(London: Heinemann).
Grandy, R. E. (1973), Theories and Observation in Science(Englewood Cliffs, NJ:
Prentice-Hall).
Greenberg,D. S. (1969), The Politics oj American Science(Harmondsworth:Penguin).
Haberer, J. (1969), Politics and the Communityoj Science(New York: Van Nostrand
Rheinhold).
Habermas,J. (1972), Knowledgeand Human Interests(London: Heinemann).
Hagstrom,W. O. (1965), The Scientific Community(New York: Basic Books).
Hanson, N. R. (1965), Patternsoj Discovery(Cambridge:CambridgeUniversity Press).
Hanson,N. R. (1969), Perceptionand Discovery(San Francisco:Freeman,Cooper).
Harris, E. E. (1970), Hypothesisand Perception(London: Allen & Unwin).
Harwood, J. (1976), 'The race-intelligencecontroversy',Social Studiesoj Science,vol. 6,
pp. 369-94, and vol. 7 (1977), pp. 1-30.
Hempel,C. G. (1965), Aspectsoj ScientificExplanation(New York and London: The Free
Press).
Hesse,M. (1974), The Structureoj Scientific Injerence(London: Macmillan).
Hessen, B. (1931), 'The social and economic roots of Newton's Principia', in N. I.
Bukharin et al. (eds), Scienceat the Crossroads(2nd ed 1971) (London: Frank Cass).
Hirsch, W. (1961), 'The autonomyof sciencein totalitarian societies',Social Forces, vol.
40, pp. 15-22.
Hoijer, H. (1964), 'Cultural implications of some Navaho linguistic categories',in D.
Hymer (ed.), LanguageandCulture in Society(New York and London: Harper& Row),
pp. 142-53.
Holton, G. (1973), ThematicOrigins oj Scientific Thought (Cambridge,Mass.: Harvard
University Press).
Johnston,IR.(1976),'Contextualknowledge',AustraliaandNew ZealandJournal oj Socio-
logy, vol. 12, pp. 193-203.
Johnston,R. (1978), 'Goal direction of scientific research',in W. Krohn, E. Layton and
P. Weingart(eds), The Dynamicsoj Scienceand Technology,Sociology of the Sciences
Yearbook, Vol. II (Dordrecht, Holland: Reidel).
Johnston,R., andShepherd,J. (1976), Scienceand Rationality (Manchester:University of
Manchester,Departmentof Liberal Studiesin Science,SisconPamphlet).
Joravsky,D. (1970), The LysenkoAjjair (Cambridge,Mass.: Harvard University Press).
Kemp, R. V. (1977), 'Controversyin scientific researchand tactics of communication',
SociologicalReview, vol. 25, pp. 515-34.
King, L. R., and Melanson,P. H. (1972), 'Knowledgeand politics', Public Policy, vol. 20,
pp. 82-lOl.
King, M. D. (1971), 'Reason,tradition and the progressiveness of science',History and
Theory, vol. 10, pp. 3-32.
Knorr, K., Strasser,H., and Zilian, H. G. (1975), Determinantsand Controls of Scientific
Development(Dordrecht, Holland: Reidel).
Kroeber, A. L. (1944), Configurations of Culture Growth (Berkeley, Calif., and Los
Angeles: University of California Press).
Kuhn, T. S. (1962), The Structureof ScientificRevolutions(Chicagoand London: Univer-
sity of Chicago Press)(enlargeded 1970).
Kuhn, T. S. (1963), 'The function of measurementin modern physical science', in H.
Woolf (ed.), Quantification (New York: Bobbs-Merrill), pp. 31-6l.
Lakatos, I. (1968), 'Criticism and the methodologyof scientific researchprogrammes',
Proceedingsof the Aristotelian Society,vol. 69, pp. 149-86.
Lakatos,1. (1970), 'Falsificationand the methodologyof scientific researchprogrammes',
in I. Lakatos and A. Musgrave(eds), Criticism and the Growth of Knowledge(Cam-
bridge: CambridgeUniversity Press).
Lakatos, I. (1976), Proofs and Refutations(Cambridge:CambridgeUniversity Press).
Lakatos,I., andMusgrave,A. (1970), Criticism andthe Growth ofKnowledge(Cambridge:
CambridgeUniversity Press).
Lakoff, S. (1977), 'Scientists,technologistsand political power', in I. Spiegel-Rosingand
D. J. de Solla Price (eds), Science,Technologyand Society(London: Sage).
Lane, R. (1966), 'The decline of politics and ideology in a knowledgeablesociety',
American SociologicalReview,vol. 31, pp. 649-62.
Law, J. (1975), 'Is epistemologyredundant?A sociological view', Philosophyof Social
Science,vol. 5, pp. 317-37.
Law, J. (1976), 'Theories and methods in the sociology of science: an interpretative
approach',in G. Lemaine et al. (eds), Perspectiveson the Emergenceof Scientific
Disciplines (Paris: Mouton, and Chicago: Aldine).
Law J., and French, D. (1974), 'Normative and interpretative sociologies of science',
SociologicalReview,vol. 22, pp. 581-95.
Lemaine, G. et al. (eds) (1976), Perspectiveson the Emergenceof Scientific Disciplines
(The Hague/Paris:Mouton, and Chicago: Aldine).
Lemaine,G., and Matalon, B. (1969), 'La Lutte pour la vie dansla cite scientifique',Revue
Franfaise de Sociologie,vol. 10, pp. 139-65.
Lewis, C. I. (1956), Mind and the World Order (New York: Dover Publications).
MacKinnon, E. A. (1972), The Problem of Scientific Realism (New York: Appleton-
Century-Crofts).
MacLeod, R. (1977), 'Changingperspectivesin the social history of science',in I. Spiegel-
Rosingand D. J. de Solla Price (eds), Science,Technologyand Society(London: Sage),
pp. 149-96.
Mannheim, K. (1936), Ideology and Utopia (New York: Harcourt, Brace & World).
Mannheim,K. (1952), Essayson the SociologyofKnowledge(London: Routledge& Kegan
Paul).
Marcuse,H. (1962), One-DimensionalMan (London: Routledge& Kegan Paul).
Marx, K. (1904), A Contribution to the Critique of Political Economy(Chicago:C. H.
Kerr.)
Marx, K., and Engels, F. (1965), Manifesto of the CommunistParty (Moscow: Progress
Publishers).
Marx, K. (1973), Grundrisse(Harmondsworth:Penguin).
Marx, K. (1974), Economicand Philosophical Manuscripts of 1844 (Moscow: Progress
Publishers).
Mason, S. F. (1962), A History of the Sciences(New York: Collier).
Mazur, A. (1973), 'Disputesbetweenexperts',Minerva, vol. II, pp. 243-62.
McHugh, P. (1971), 'On the failure of positivism', in J. D. Douglas(ed.), Understanding
References 127
EverydayLife (London: Routledge& Kegan Paul).
Medawar, P. (1969), The Art of the Soluble(Harmondsworth:Penguin).
Mendelsohn,E. (1964), 'The emergenceof scienceas a professionin nineteenth-century
Europe',in K. Hill (ed.), The Managementof Scientists(Boston, Mass.: BeaconPress).
Merton, R. K. (1957), Social Theory and Social Structure(revisedand enlargeded) (New
York: The Free Press).
Merton, R. K. (1970), Science, Technologyand Societyin Seventeenth-Century England
(New York: Harper & Row) (originally publishedin 1938).
Merton, R. K. (1973), The Sociology of Science(Chicago and London: University of
ChicagoPress)(ed. with an introduction by N. W. Storer).
Merton, R. K. (1975), 'Structuralanalysisin sociology',in P. Blau (ed.), Approachesto the
Studyof Social Structure(New York: The Free Press),pp. 21-52.
Mitroff, I. I. (1974), The SubjectiveSide of Science(Amsterdam:Elsevier).
Mulkay, M. J. (1969), 'Some aspectsof cultural growth in the natural sciences',Social
Research,vol. 36, pp. 22-52.
Mulkay, M. J. (1972), The Social Processof Innovation (London: Macmillan).
Mulkay, M. J. (1974), 'Methodologyin the sociologyof science',Social ScienceInforma-
tion, vol. 13, pp. 107-19.
Mulkay, M. J. (1976a), 'Norms and ideology in science',Social ScienceInformation, vol.
15, pp. 637-56.
Mulkay, M. J. (I 976b), 'The mediating role of the scientific elite', Social Studies of
Science,vol. 6, pp. 445 -70.
Mulkay, M. J. (1977a), 'Connectionsbetween the quantitative history of science, the
social history of scienceand the sociology of science',Proceedingsof the International
Seminaron ScienceStudies(Helsinki: Academyof Finland), pp. 54-76.
Mulkay, M. J. (1977b), 'Sociology of the scientific researchcommunity', in I. Spiegel-
Rosingand D. J. de Solia Price (eds), Science,Technologyand Society(London: Sage),
pp.93-148.
Mulkay, M. J., Gilbert, G. N., and Woolgar, W. (1975), 'Problem areasand research
networksin science',Sociology,vol. 9, pp. 187-203.
Mulkay, M. J. and Williams, A. T. (1971), 'A sociologicalstudy of a physicsdepartment',
British Journal of Sociology,vol. 22, pp. 68-82.
Musson, A. E. and Robinson, E. (1969), Scienceand Technology in the Industrial
Revolution(Manchester:ManchesterUniversity Press).
Nagel, E. (1961), The Structureof Science(London: Routledge& Kegan Paul).
Nagel, E., Bromberger,S., and Griinbaum, A. (1971), Observationand Theoryin Science
(Baltimore and London: JohnsHopkins Press).
Nelkin, D. (1971), 'Scientistsin an environmentalcontroversy',ScienceStudies, vol. 1,
pp.245-61.
Nelkin, D. (1975), 'The political impact of technicalexpertise',Social Studiesof Science,
vol. 5, pp. 35-54.
Nelkin, D. (1977), 'Scientistsand professionalresponsibility: the experienceof American
ecologists',Social Studiesof Science,vol. 7, pp. 75-95.
O'Neil, W. N. (1969), Fact and Theory (Sydney: Sydney University Press, and London:
Methuen).
Outhwaite, W. (1975), UnderstandingSocial Life (London: Allen & Unwin).
Pinch. T. J. (1977), 'What does a proof do if it does not prove?', in E. Mendelsohn,
P. Weingart and R. Whitley (eds), The Social Production of Scientific Knowledge
(Dordrecht, Holland: Reidel).
Polanyi, M. (1951), The Logic of Liberty (London: Routledge& Kegan Paul).
Polyani, M. (1958), PersonalKnowledge(London: Routledge& Kegan Paul).
Polanyi, M. (1969), Knowing and Being (London: Routledge& Kegan Paul).
Popper,K. R. (1959), The Logic of ScientificDiscovery(New York: Harper & Row).
Popper,K. R. (1963), Conjecturesand Refutations(London: Routledge& Kegan Paul).
Price, D. J. de Solla (1963), Big Science,Little Science(New York: ColumbiaUniversity
Press).
128 ScienceandSociologyoj Knowledge
Radnitzky, G. (1974), 'From logic of scienceto theory of research',Communicationand
Cognition, vol. 7, pp. 61-124.
Ravetz, J. R. (1971), Scientific Knowledgeand Its Social Problems(Oxford: Clarendon
Press).
Robbins, D., and Johnston,R. (1976), 'The role of cognitive and occupationaldifferen-
tiation in scientific controversies',Social Studiesoj Science,vol. 6, pp. 349-68.
Rose, H., and Rose, S. (1976) (eds), The Political Economy oj Science (London:
Macmillan).
Rudwick, M. J. (1976), 'The emergenceof a visual language for geological science',
History oj Science,vol. 14, pp. 149-95.
Ryle, G. (1949), The Conceptoj Mind (London: Hutchinson).
Sandow,A. (1938), 'Social factors in the origin of Darwinism', The Quarterly Reviewoj
Biology, vol. 13, pp. 315-26.
Scheffler, 1. (1963), The Anatomyoj Inquiry (New York: Knopf).
Scheffler, 1. (1967), Scienceand Subjectivity(New York: Bobbs-Merrill).
Schon, D. A. (1963), Displacementoj Concepts(London: Tavistock)..
Sklair, L. (1973), OrganizedKnowledge(London: Hart-Davis, MacGibbon).
Slack, J. (1972), 'Classstruggleamongthe molecules',in T. Pateman(ed.), Countercourse
(Harmondsworth:Penguin).
Smart, J. C. (1973), 'The reality of theoreticalentities', in R. E. Grandy (ed.), Theories
and Observationin Science(EnglewoodCliffs, NJ: Prentice-Hall).
Stark, W. (1958), The Sociologyoj Knowledge(London: Routledge& Kegan Paul).
Storer,N. W. (1966), The SocialSystemojScience(New York: Holt, Rinehart& Winston).
Storer, N. W. (1973), Introduction to Merton, R. K., The Sociologyoj Science(Chicago
and London: University of ChicagoPress).
Strawson,P. F. (1959), Individuals (London: Methuen).
Sullivan, D., White, D. H., and Barboni, E. J. (1977), 'Co-citation analysesof science',
Social Studiesoj Science,vol. 7, pp. 223-40.
Thackray, A. (1970), 'Scienceand technology in the industrial revolution', History oj
Science,vol. 9, pp. 76-89.
Tobey, R. C. (1971), The American Ideology oj National Science (Pittsburgh, Pa.:
University of PittsburghPress).
Toulmin, N. J. (1968), Foresight and Understanding(New York: Harper & Row).
Vig, N. J. (1968), Scienceand Technologyin British Politics (London: Pergamon).
Vorzimmer, P. J. (1969), 'Darwin's questionsabout the breeding of animals (1839)"
Journal oj the History oj Biology, vol. 2, pp. 269-81.
Webster,A. J. (1978), 'Scientific controversyand socio-cognitivemetonymy: the caseof
acupuncture',in R. Wallis (ed.), RejectedKnowledge:SociologicalReviewMonograph
(Keele: University of Keele).
Weingart, P. (1970), Die AmerikanischeWissenschajtslobby(Dusseldorf: Bertelsmann
UniversiUitsverlag).
Weingart, P. (1974), Wissenschajtssociologie (Frankfurt am Main: Atheneum Fischer
TaschenbuchVerlag).
Weingart, P. (1974), 'On a sociologicaltheory of scientific change',in R. Whitley (ed.),
Social Processesoj ScientificDevelopment(London: Routledge& Kegan Paul).
Whitley, R. (1972), 'Black-boxism and the sociology of science', in P. Halmos (ed.),
SociologicalReviewMonograph (Keele: University of Keele), pp. 61-92.
Whitley, R. (ed.) (1974), Social Processesoj ScientificDevelopment(London: Routledge
& Kegan Paul).
Whitley, R. (1975), 'Componentsof scientific activities, their characteristicsand institu-
tionalisationin specialtiesand researchareas',in K. Knorr, H. Strasserand H. G. Zilian
(eds), Determinantsand Controls oj Scientific Development(Dordrecht, Holland:
Reidel).
Wittgenstein,L. (1953), PhilosophicalInvestigations(Oxford: Blackwell).
Woolgar, S. (1976a), 'Writing an intellectual history of scientific development:the use
References 129
of discoveryaccounts',Social Studiesof Science,vol. 6, pp. 395-422.
Woolgar, S. (1976b), 'The Identification and definition of scientific collectivities', in O.
Lemaine et 01. (eds), Perspectiveson the Emergenceof Scientific Disciplines (Paris:
Mouton, and Chicago: Aldine).
Wynne, B. (1976), 'c. O. Barkla and the J phenomenon:a casestudy of the treatmentof
deviancein physics', Social Studiesof Science,vol. 6, pp. 307-47.
Young, R. M. (1969), 'Malthus and the evolutionists: the common context of biological
and social theory', Past and Presentvol. 43, pp. 109-45.
Young, R. M. (1971a), 'Darwin's metaphor: does Nature select?', The Monist, vol. 55,
pp.442-503.
Young, R. M. (1971b), 'Evolutionary biology and ideology-then and now', Science
Studies,vol. I, pp. 177-206.
Young, R. M. (1973), 'The historiographicand ideological contexts of the nineteenth-
centurydebateon Man's placein Nature',in M. Teich and R. M. Young (eds), Changing
Perspectivesin the History of Science(London: Heinemann).
Index
Agassi, J. 19 Daniels, G. H. III

Arago, F. 88-90 Darwin, c., Darwinian theory 18-19,36-7,
50-I, 100-9
Bacon, F., Baconianmethod I, 103 De Gre, G. 21-2
Barber, B. 23, 97, 110 Democracyand science96-7, 111-2
Barkla, C. G. 79-82 Deutsch,M. 98
Barnes,B. 23, 62, 98 Dilthey, W. II
Ben-David, J. 86, 96 Dolby, R. G. 23
Benveniste,G. 114 Duhem, P. 46-8
Biot, J. B. 88 Dunbar, R. T. 80-1
Blissett, M. 68 Durkheim, E. 1,3-7, 10, 12,60,97
Bloor, D. 12, 78, 91-2 Differentiation in science73, 97-8
Blume, S. S. I 10 Discovery 18, 21, 50, 98
Bohm, D. 36, 43-4
Bohme, G. 93-4
Easlea,B. 27, 50
Borger, R. 44
Edge, D. O. 41, 56, 71
Bourdieu, P. 77
Einstein, A. 15, 75, 77
Braverman,H. 6
Elkana, Y. 109
Brooks, H. 110
Elliott, H. C. 37,98
Bruner, J. S. 44-7
Brush, S. G. 109 Empirical acuity of scientific theories37-9,
41
Burtt, E. A. 43
End of ideology thesis 1I0-1, 1I3, 120; see
Capitalist society 5-7, 10, 101
also ideology
Engels, F. 6, 8
Carnap,R. 32
Everyday knowledge44
CayugaLake controversy114-6
External social influences on scientific
Celestial radio emission,first reports 56-7
thought 96-109, 120-1
Chambers,R. 100
External world, constraint on scientific
Chubin, D. E. 88
knowledge61
Cole, J. R. and Cole, S. 25, 26, 69
Ezrahi, Y. 110
Cognitive and technical norms 94-5
Collins, H. M. 26, 74-8, 83-6, 90-1
Commonsense knowledgeI, 37, 47, 98; see Falsification, principle of 54, 76-8
also everydayknowledge Feuer, L. S. 109
Compton, A. H. 79-80 Feyerabend,P. 35, 91
Condorcet,A. N. 105 Fisher, R. A. 51-2
Consonancewith observarion50, 52-3 Fluoridation and public health 115
Constitutive and contingent forums 83-8, Focke, W. O. 36
92 Formal and empirical sciences92
Copernicus,N. 31 Forman, P. 109
Corpuscularand wave theoriesof light 53, Fox, R. 90
87-90 Frake, C. O. 35
Counter-norms66-8, 70 Frank, G. P. 50, 53
Criteria of adequacy20, 55-7, 81-2, 89-90; Frankel, E. 87-90, 92
of truth 11-12,seealso knowledgeclaims, French, D. 93
assessment of Fresnel,A. 90
Curtis, J. E. I, 10 Functionalanalysis9, 25, 67
Index 131
Galileo, G. 31 Lakatos, I. 38, 41, 48, 53-4
Giddens, A. 27 Lakoff, S. 110
Gilbert, G. N. 38, 59, 69, 73, 78 Lamarck, J. B. 100, 104
Glass, B. 36 Laplace, P. S. 88-90
Godwin, W. 105 Law, J. 72-3, 92-3
Gouldner, A. W. 70-1, 112 Lemaine, G. 23, 96
Grandy, R. E. 31,41 Lewis, C. I. 49
Gravitational waves, research network Linguistic frameworks 35-6, 45-6
75-8, 90-1 Lukacs, G. 19
Greenberg,D. S. 112-3 Lyell, C. 103-5
Lysenko, T. D. 96
Haberer, J. 111-2
Habermas,J. 5 Mackinnon, E. A. 21
Hagstrom,W. O. 32 MacLeod, R. M. 97
Hanson, N. R. 27-8, 35, 37,42, 45-6, 49, Malthus, Malthus's law, Malthusian
91, 105 doctrine 18, 105-8
Harris, E. E. 46 Mannheim,K.1,10-17, 19,27,37-8,60,115
Harrison, R. G. 75 Marcuse, H. 5
Hempel, C. G. 33 Marx, K., Marxist 1,3,5-11,13,19
Hesse,M. 27, 33,46, 91 Mason, S. F. 73
Hessen,B. 7-8, 10 Matalon, B. 23
Hirsch, W. 97 Mazur, A. 110, 115-6, 121
History of science3, 27, 97 McHugh, P. 78
Hoijer, H. 35 Medawar, P. 20, 69
Holton, G. 73, 90, 98-100, 105, 108-9 Melanson, P. D. 114
Hooker, J. D. 101 Mendel, G., Mendelian 36-7, 50-2, 58
Hooke's law 53 Mendelsohn,E. 96
Huxley, T. H. 101 Merton, R. K. 1, 8-10, 12, 19, 22-5, 48, 60,
66, 68, 70, 94, 97
Mitroff, J: l. 23, 64, 66-70, 121
Ideology, ideology of science5,8-11, 111-3,
Moral languageof science63-73
120; seealso end of ideology thesis
Mulkay, M. J. 22, 25, 41,56-7,66,71,73,
Incommensurability38-9, 41
113
Induction 28, 53-4
Musgrave,A. 41
Informal reasoning98-9, 108
Interpretative analysis of norms 93-5; see
Nagel, E. 20, 30, 37-8, 40
also norms of science
Nageli, C. von 36
Interpretativefailure 109
Nelkin, D. 109, 114-5, 121
Inverse squarelaw 28-9
Neo-Kantianthought, 11, 15
Network theory of scientific meaning33-4,
Johnston,R. 53, 109 39-42,44-5
Joravsky, D. 96 Newton, I. 7-8, 10,28,40-1,99
Journals57, 68-9, 84, 86 Norms of science21-6,63-73,96-7,110-3,
J phenomenon79-82 118-9
Kemp, R. V. 78 Observableand unobservableobjects 30-2,

King, L. R. 114 37
King, M. D. 22 Observation11-12, 14, 16, 19-20,30-1,33,
Knowledge claims, assessmentof 9-10, 40, 42-50, 52-3, 118-9
13-14,20,23-6,49-61,65,67, 77-92, 94-5, Observational and theoretical laws 20-1,
104, 118-20 29-34
Kroeber, A. L. 100 Ohm's law 53
Kuhn, T. S. 27, 38,41,47-8,56,73,87,91, O'Neil, W. N. 20
105 Outhwaite, W. 11
Paley, W. 102, 105 Sociology of knowledge I, 2, 5, 10-12, 14,
Parapsychology, research network 83-7, 17,19,28,37,60-3,95-6,110,118,120,
91-2 122
Pareto,V. I Spencer,H. 100
Petras,J. W. I, 10 Standardview of science19, 25, 27, 37-9,
Phenomenalseeing49 41-3, 49, 63-5, 84, 118-9
Philosophyof science3, 27, 43, 63-5 Standardizationof scientific formulations
Pinch, T. J. 78, 83-6 58-9
Planck'slaw 56 Stark, W. 17-19,27,96,109
Plato 99 Storer, N. W. 23, 66, 68
Polanyi, M. 90-1, 97 Stratification in science25-6, 68-9, 120
Politics and Science110-8, 120; of research Strawson,P. F. 35
community 88-90
Popper, K. R. 27, 50, 54, 60 Tacit knowledge74-6, 90
Price, D. J. de Solla 73,112 TEA laser, researchnetwork 74-6
Psychology,experimental43-4 Technologyand science17, 121
Pulsars39-41, 46, 70-1 Thematain science99-100, 105-8
Tobey, R. C. 112
Radiation and public health 115-7 Toulmin, N. J. 58
Ravetz, J. R. 27, 38, 45, 50, 53-9, 90 Two-tier view of science30
Relativity of ideas 3, 12, 14-5, 37
Religion and science4, 7, 22-3, 85 Uniformity of nature, principle of unifor-
Replicability, replication 50-2, 76-8, 81 mity II, 17-19,27-9,49,101,103, 118-9
Representationof physical world 44, 60
Revolutionsin science73, 87-90 Vocabulary of evaluations/justifications
Rose, H. and Rose, S. 9-10, 109 70-2, 77-8, 91-5, 111-3, 120-1
Rudwick, M. J. 45 Vorzimmer, P. J. 101
Ryle, G. 34
Wallace, A. R. 101, 104-7
Sandow, A. 101-2, 106 Weber, A. 19
Scheffler, I. 19, 35-6, 38, 47 Weingart, P. 23, 112
Schon, D. A. 105 Whitley, R. 26
Seaborn,A. E. 44 Williams, A. T. 57
Secondaryqualities 43 Wittgenstein, L. 50
Shepherd,J. 53 Woolgar, S. 39, 73, 92
Sklair, L. 112 Wynne, B. 79-82, 91
Slack, J. 109
Social thought and scientific thought 11-18, Young, R. M. 88-9,96, 100, 102-4, 106-9
62

(Routledge Library Editions - Social Theory 60) Michael Mulkay - Science and The Sociology of Knowledge-Routledge (2014)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(Routledge Library Editions - Social Theory 60) Michael Mulkay - Science and The Sociology of Knowledge-Routledge (2014)

Uploaded by

Copyright:

Available Formats

ROUTLEDGE LIBRARY EDITIONS:

SCIENCE AND THE SOCIOLOGY

Trademark notice: Product or corporate names may be trademarks or registered

ISBN: 978-0-415-72731-0 (Set)

This book is copyright under the BerneConvention.All rights

GEORGE ALLEN & UNWIN LTD

© GeorgeAllen & Unwin (Publishers)Ltd, 1979

British Library Cataloguingin PublicationData

Mulkay, Michael Joseph

Typesetin 10 on 11 point Times by Red Lion Setters,London

I wish to thank Tom Bottomoreand John Law for readingthe first

If we cannotbearthe paradoxof acceptingthat genuineknowledge

The Customary Sociological

The sociologyof knowledgehasa long history. Its origins are sometimes

THE CLASSIC VIEW OF SCIENCE: DURKHEIM AND MARX

Durkheim reachesthis conclusion by postulating a unity between the

. . .the essentialideas of scientific logic are of religious i.e. social

Durkheim describesat a rather general level the social conditions

The bourgeoisie... hascreatedmore massiveand more colossalforces

Thus capitalism needs and promotes 'the developmentof the natural

The unity of thought and action, conceptionand execution,hand and

The centraltheme,then, of the Marxist analysisof scienceis to seethe

... nature becomespurely an object for mankind, purely a matter of

With the changeof the economicfoundationthe entireimmensesuper-

... sciencebecomesan ideology and scientists the ideologists. How

In this passage,the authorsmake a distinction betweenthe technical

MORE RECENT VARIANTS: MANNHEIM AND STARK

Are the existential factors in the social processmerely of peripheral

In the light of several statementsof this kind and in view of the

Similarly, every kind of knowledgewhich only certainspecific histori-

As the bourgeoisieachieved a position of social and political pre-

The particularity of the theory of knowledge holding sway today is

In responseto this situation, Mannheimtries to formulate an alternative

We will have to reckon with situationaldeterminationas an inherent

This last sentencebrings us back to the 'marginal, special case' of

. . .whereasman has more than onceshifted his vantagepoint for the

The dependenceof scientific knowledgeon a particular perspectiveon,

Socialdevelopmentsdo not determinethe contentof scientific develop-

Whatever its origins, the Darwinian proposition is, in Stark's view,

THE STANDARD VIEW OF SCIENTIFIC KNOWLEDGE

THE SOCIOLOGY OF SCIENCE

The institutional goal of science is the extension of certified know-

Scientists' belief in the goodnessof these mores has its historical

Revisionsof the Standard View

In this chapter I intend to describe some recent contributions to

THE UNIFORMITY OF NATURE

Hence, to gain knowledgeof the truth of theseprinciples [Hanson is

Hanson goes on to show with various specific examples that the

FACT AND THEORY

. . .an experimental law, unlike a theoretical statement, invariably

This two-tier view of scientific knowledge, taken for granted by most

. . .if observabilityis merelya matterof degree,then thereseemsto be

Let us turn, therefore, to the questionof whether a difference can be

It is not correct to speak,as is often done, of 'the experientialmean-

If this is so, if observationalterms have no experientialmeaningapart

. . . no feature in the total landscapeof functioning of a descriptive

We appear,then, to have reacheda conclusionwhich rejectstwo basic

. . . our typesof notationin physicsmay occasionallyrenderus insensi-

Hansonstressesthat it is impossibleto demonstratethis thesis other

It follows that although... the samecategorizationallows the expres-

Although linguistic frameworks are necessaryfor the statementof

Having crossedsnapdragonsand producedhybrid varieties, Darwin

Mendel himselfwas ableto transformsimilar findings into a statementof