Professional Documents
Culture Documents
(Routledge Library Editions - Social Theory 60) Michael Mulkay - Science and The Sociology of Knowledge-Routledge (2014)
(Routledge Library Editions - Social Theory 60) Michael Mulkay - Science and The Sociology of Knowledge-Routledge (2014)
SOCIAL THEORY
Volume 60
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.
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
ISBN 0-04-301093-8
ISBN 0-04-301094-6 Pbk
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
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)
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
8 Scienceand the Sociologyoj Knowledge
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:
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:
OBSERVATION IN SCIENCE
From within the standardview of science,observationis thought to be
scientifically adequatein so far as such distorting influences as bias,
Revisionsof the StandardView 43
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
44 Scienceand the SociologyofKnowledge
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.
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)
Revisionsoj the StandardView 47
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.
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.
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)
... 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
56 Scienceand the Sociologyoj Knowledge
explicit list oj criteria oj adequacy applying to a wide class oj
problems.(Ravetz, 1971, pp. 154-5)
SOCIOLOGICAL IMPLICATIONS
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)
84 Scienceand the Sociologyoj Knowledge
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
Cultural Interpretation in Science 85
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.
86 Scienceand the SociologyofKnowledge
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)