Sei sulla pagina 1di 17

Patterns of Name Ordering Among Authors of Scientific Papers: A Study of Social Symbolism

and Its Ambiguity


Author(s): Harriet A. Zuckerman
Source: American Journal of Sociology, Vol. 74, No. 3 (Nov., 1968), pp. 276-291
Published by: University of Chicago Press
Stable URL: http://www.jstor.org/stable/2775535
Accessed: 20-11-2015 15:03 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.

University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to American Journal of
Sociology.

http://www.jstor.org

This content downloaded from 132.203.227.62 on Fri, 20 Nov 2015 15:03:51 UTC
All use subject to JSTOR Terms and Conditions
of Name OrderingamongAuthors
Patterns of Scientific
Papers:
A Studyof Social Symbolism
and ItsAmbiguity'
Harriet
A. Zuckerman

ABSTRACT
Withincreasingscientific collaboration,visibilityofindividualrole-performance
has dimin-
ished.Orderingofauthors'namesis an adaptive devicewhichsymbolizestheirrelativecontri-
butions to research.Interviewswith Nobel laureates and comparisonsoftheirname-order
practicesto thoseof otherscientistssuggestthat thissymbolis ambiguousand makes evalua-
tion of individualrole-performance A probabilitymodel of expecteddistributions
difficult. of
name ordersis used in measuringpreferences forparticularsequences,and these preferences
varywiththe authors'eminence.On the assumptionthat authors'names are listedin orderof
the value of theircontributions, laureatesshouldbe first-authorsmoreoftenthan otherscien-
tists; in fact,theyare not. Instead, theyexercisetheirnoblesseobligeby givingcreditto less
eminentco-workers increasinglyas theireminencegrows.They do so moreoftenaftertheprize,
and eminentlaureates-to-be foregofirst-authorshipmoreoftenthan thoseas yetunrecognized.
This noblesseoblige,however,has its limits;laureates' contributionsto prize-winning
research
are morevisiblethan contributionsto theirotherresearch.

Everysocialinstitution providesforthe In science,evaluation is based on the


evaluation of role-performance withinit. extentand quality of scientificcontribu-
Assessmentof the quality,quantity,and tions as appraisedby significantothersin
style of role-performance is requiredfor the field. This is nothingnew. What is
social systemsto operatewithsomedegree new is the profoundchange in the social
Evaluationis both a com-
of effectiveness. organizationof scientificwork which,in
ponentof social controlwhichcurbsexces- many cases, reducesthe visibilityof role-
sive departuresfrominstitutionalnorms performanceof individual investigators.
and a componentof the reward system In thepast, therewere,as now,discrepant
whichrespondsto variationsin role-per- judgmentsabout thequalityofa scientist's
formanceby promoting, demoting,retain- work,but rarelywas thereany ambiguity
ing, and removingparticipantsfrompar- about whohad donewhat.A scientistmay
ticular social systems.Evaluation, how- have had technicalassistantsworkingwith
ever,requiresthatrole-performance should him,but he was unambiguously theinvesti-
be visiblein somedegree.All thisholds in gator,not merelythe "principalinvestiga-
the institutionof sciencejust as it does in tor." When he put his findingsinto print,
the institutionsof religion,politics, the he was, with few exceptions,the sole
arts,and the economy.2 author. There could be no doubt whose
1 Revisionof a paper read at the AnnualMeeting scientific
performance was being assessed.
of the American Sociological Association,August, As the social organizationof scientific
1967. This researchwas supportedby a grantfrom work has become more complex-more
the National Science Foundation (GS 960) to the
Programin the Sociologyof Science,Columbia Uni- often collaborativeand sometimesintri-
versity.I want to thank Robert K. Merton forhis cately organized-the visibilityof indi-
guidanceand criticism;and OrvilleG. Brim,Jr.,and vidual role-performance is reduced. The
NormanStorerfortheirsuggestions.An earlierdraft
2 For a systematicdiscussion of the relations
of thispaper was read at the LaboratoryforSocio-
EnvironmentalStudies,National Instituteof Men- between visibilityof role-performance, evaluation,
tal Health. This maybe identified as PublicationNo. and social control,see R. K. Merton,Social Theory
A-489 of the Bureau of Applied Social Research, and Social Structure(Glencoe,Ill.: Free Press,1957),
Columbia University. pp. 320 ff.
276

This content downloaded from 132.203.227.62 on Fri, 20 Nov 2015 15:03:51 UTC
All use subject to JSTOR Terms and Conditions
A STUDYOF SOCIALSYMBOLISMAND ITS AMBIGUITY 277
changed organizationis reflectedin the diflicultiespersist in a changing social
rapid increase of multi-authored papers. system, adaptive mechanisms,effectual
As Derek Price, BernardBerelson,and I and ineffectual, designedto cope withthem
have variouslyreportedelsewhere,3 about typicallydevelop.And so it is here.As the
80 per centofpapersin chemistry are now numberofmulti-authored papersincreased
multi-authored (a fourfoldincreasesince and as the numberof co-authors(what I
the turn of the century),60 per cent in call the author-set)increased,new prac-
physicsand 40 per cent in the biological tices were introducedto lessen the am-
sciences,in contrastto the 1 per cent of biguityabout the distinctiveroles of co-
papers in such humanisticsubjects as authorsin doingresearchand in preparing
philosophy,language,and literature.Such published papers. Various patterns of
papers, of course, introduce ambiguity name orderingof co-authorswereevolved
about the respectivecontributions of the forthispurpose.5
joint authors. This is just the sortof device that pro-
This ambiguityhas interfered with the duces ambivalence.Intrinsically,it seems
operationof the rewardsystemin science trivialor humiliating to be concernedwith
whichis now,as it alwayshas been,better the orderin whichthe names of authors
geared to allocate credit to individual appear; forscientiststo be as interestedin
scientiststhan to groupsof them.4Collab- the orderof theirbillingas actorsor busi-
orative researchis, in varyingdegrees,a ness partners. But symbolically,name
functionalrequirementof contemporary orderingis an adaptive device that facili-
scientific investigation. At the same time, tates the allocation of responsibility and
rewardsin the formof promotions, prizes, credit among co-workersin otherwise
and resourcesforfurther worktypicallygo ambiguoussituationsinducedby the new
to individuals.This malintegration of the structuresof scientificresearch.It is de-
rewardsystemand the social organization signedto have the rewardsystemoperate
ofscientific researchis reflected in thediffi- witha degreeofequityand adequacy.The
cultiesof findingways to allocate relative ambivalencetowardname orderingis only
responsibilityand credit for particular one more instance of scientists'ambiva-
investigationsamong collaborators.When lencetowardrecognition fortheirworkand
3 DerekJ. deS. Price,LittleScience,Big Science the means used to achieveit.6
(New York: ColumbiaUniversity Press, 1963), Beyondthis,twoaspectsoftheordering
pp. 88 ff.;BernardBerelson,GraduateEducationin of authors'namesneed onlybe mentioned
theUnitedStates(New York: McGraw-HillBook here.First,nameordersconveyambiguous
Co., 1960),p. 55; HarrietZuckerman, "NobelLau-
reatesin theUnitedStates:A Sociological Studyof information about the relative contribu-
ScientificCollaboration" (unpublished Ph.D. dis- tions of co-workers since,as we shall see,
sertation,ColumbiaUniversity, 1965), chap. iii.
Othershave reportedsimilarfindings forlimited 5It maybe ofinterestthatthisstudywasunder
samplesand timeperiods.See Zuckerman, op. cit., wayat theverytimethatNormanKaplanwas ob-
fora listofthese. serving withregretthat"therehas beenno known
4 Scientists
arenotunawareoftheincongruity of systematic study ofthenorms pertaining
to . .. the
the organization of contemporary scienceand the arrangement ofnamesinmultiple-authored papers"
allocationofcreditto individual Sinceit ("Sociologyof Science,"p. 857 in R. E. L. Faris,
scientists.
is difficult
indeedto assesstheparticularcontribu- [ed.],Handbook ofModern Sociology
[Chicago:Rand
tions of collaborators, one physicistfacetiously McNally & Co., pp.
19641, 852-81).
suggested thatresearchgroupsin his fieldbe or- 6 See R. K. Merton,"Prioritiesin Scientific
ganizedintoteamswiththenotimprobable titles: Discovery: A Chapterin theSociology ofScience,"
"The Harvard-M.I.T.Yankees"and "The 'Reds' AmericanSociologicalReview,XXII (December,
ofDubna." The standing ofeachin its ownleague 1957),635-59,forhisanalysisofthetension between
could then be computedand rewardsproperly institutional normsrequiring humilityand theem-
allocated(R. E. Westonetal., "ModestProposal," phasisplacedon makinga distinctive contribution
PhysicsToday,XV [June,1962],79-80). to science.

This content downloaded from 132.203.227.62 on Fri, 20 Nov 2015 15:03:51 UTC
All use subject to JSTOR Terms and Conditions
278 THE AMERICANJOURNALOF SOCIOLOGY
customary practices differamong dis- sent the average scientistspublishingin
ciplinesand withinthem,and since name thesefields.
orders,like othersymbols,are sometimes Using these data, I can systematically
adapted to servepurposesotherthantheir compare name-ordering practices among
manifestones. As a resultof thisambigui- scientistsof varyingeminencein different
ty, circlesof scientistsascribe theirown fields.Later on, I shall considerthe prac-
meaningsto sequencesof authors'names, ticesofmenat different stagesofthescien-
meaningsthat often diverge fromthose tificcareer.
intendedby the authorsthemselves.For
PATTERNS OP NAME ORDERING
example, textual referencesto scientist
"X and his collaborators" whenscientistX Three principalpatternsof name order-
is listed thirdor fourthin a long list of ingare used in multi-authored papers.The
namesare indicatorsthat scientistreaders firstmay be describedas the "equality
sometimesignorenameordersin seekingto pattern"sinceit is designedto put all co-
allocate credit and responsibilityamong authorson a plane and, so faras possible,
co-workers. In theprocess,it is notunusual notto distinguishtheirrespectivecontribu-
forscientiststo requestand to receivein- tions. This type consistsof alphabetized
formationabout "who did what" through and reversealphabeticalsequencesaccord-
informal channelsof communication.7 ing to the initial lettersof authors' last
Little is known about actual patterns names. A classicallyliteralcase of alpha-
ofnameorderingemployedby scientistsof betizingwas contrivednot longago by the
varyingdegreesofeminenceand in various trio of physicistswho signed theirpaper
fields.To get someleverageon theseques- Alpher,Bethe,and Gamow.9
tions,patternsof name orderingfoundon The second consistsof patternswhich
the scientificpapersof forty-one of the fif- enhance the visibilityof one or another
ty-five Nobel laureateslivingin theUnited memberof the set of authors.These also
States are comparedto thoseon a random take two principalforms:the first-author-
sampleof articleslistedin the abstractsof out-of-sequencefollowed by an alpha-
physics, chemistry,and the biological betizedgroup(thus: ZABC), and the last-
sciences from 1920 to 1964-the period author-out-of-sequence preceded by an
corresponding to thedates ofthelaureates' alphabetizedlist (thus: XYZA). In both
publications.8Althoughthe author of a forms,the out-of-sequence authorappears
particulararticleappearingin theabstracts as a figureagainsta ground.He standsin
may be as eminentas a laureate,on the contrastto the otherswho appear as mere
average, randomlyselected authors are, elements of an alphabetical pattern.10
of course,far less distinguished and repre- 8 Articleslistedin ScienceAbstracts(physics),
ChemicalAbstracts, and BiologicalAbstracts
were
7 Althoughlaureatesagree that discussionof randomly sampledfortwoyearsofeachdecadefrom
contribution
differential whengener-
is permissible 1920to 1959andfortwoyearsintheperiod1960-64.
ous recognition is givento other,thenormofpri- Since multi-authored papers were comparatively
vacyis considered legitimate.A physicist laureate rareinthefirst twodecades,thesewereoversampled.
put it thisway: "Littlewood said thatone of the Absolutenumbers ofpapersdo notrepresentactual
primary rulesgoverning collaborationis thatone ofmulti-authored
distributions papers.
mustnotrevealtotheexternal worldwhodidwhat.
I thinkhe is correct. .. (a) becauseone'srecollec- 9R. Alpher,H. Bethe,and G. Gamow,"The
tionsarenotalwaysaccurateand (b)thereis norea- Originof ChemicalElements,"PhysicalReview,
sonfordoingso. I thinkthatifeverybody abidedby LXXIII (1948),803.
thisrule,in general, collaborationswouldbe happi- 10Thereare,ofcourse,variations on thistheme.
er." He refersto thesomewhat lessexplicitremark A groupconsisting of twoseniorinvestigators
and
byLittlewood inthemathematician's reviewofjoint severaljuniorcolleaguesmaylistthenamesofthe
workbyRamanujan,Hardy,andLittlewood (J.E. seniormenfirst orlastand thenalphabetizenames
Littlewood,A Mathematician's Miscellany (London: ofthejuniors.To simplify coding,theseweretreat-
Methuen& Co., 1953),p. 89. ed as alphabeticallyrandomsequences.

This content downloaded from 132.203.227.62 on Fri, 20 Nov 2015 15:03:51 UTC
All use subject to JSTOR Terms and Conditions
A STUDYOF SOCIALSYMBOLISMAND ITS AMBIGUITY 279
The thirdmajortypeofordering usually reates than by rank-and-file physicists.In
gives primevisibilityto the first-authorchemistryand biology,by contrast,lau-
and smallerincrements less apt to signalequality
ofvisibilityto each reatesare slightly
succeedingauthor.This is the alphabeti- of contributions by usingalphabeticallist-
cally random pattern. As we shall see, ings. As I shall show, Nobelistsin these
names are arrangedaccordingto definite fieldsattemptto confercreditupon their
criteriaotherthan the arbitraryone of al- co-workers in otherways.'2
phabeticalorder. Obviously, alphabetical sequences do
Each pattern requires separate con- not alwayssymbolizeequalityof contribu-
sideration. tion.In somecases, the alphabeticalorder
TABLE 1
PERCENTAGE OF PAPERS WITH ALPHABETIZED NAME ORDERS BY LAUREATES AND A SAMPLE
OF AUTHORs LISTED IN ABSTRACTS ACCORDING TO
FIELD AND SIZE OF AUTHOR-SET

SIZE Or AUTHOR-SET

2 3 4 5 6+ All

Physics:
Laureates. 71 (398) 55 (147) 63 (58) 54 (46) 83 (46) 67 (695)
Abstracts........ 64(1,410) 38 (643) 35(142) 39 (38) 26 (34) 57 (2,267)
Chemistry:
Laureates . 52 (638) 17 (337) 12(119) 6 (49) 11 (35) 35 (1,178)
Abstracts.. . 55(3,519) 25(1,162) 14(305) 14 (81) 12 (58) 45 (5,125)
Biological sciences:
Laureates....... 47 (750) 19 (380) 4(186) 1(100) 1 (78) 29 (1,494)
Abstracts...... 55(2,802) 22(1,061) 10(347) 2(112) 6 (63) 41 (4,385)
All:
Laureates ... . 54(1,786) 24 (864) 16(363) 15(195) 27(159) 39 (3,367)
Abstracts...... 57(7,731) 27(2,866) 16(794) 12(231) 13(155) 46(11,777)

DIFFERENTIAL USE OF NAME-ORDER will coincideby chance with the orderof


PATTERNS differentialcontributions.
A chemistlau-
ALPHABETICAL SEQUENCES
like the physicists,
chemistry, are morelikelyto
At firstappearance,Table 1 seems to alphabetizeauthors'namesthan are biochemists
indicatethat laureatesare a littleless apt whochoosepatterns muchlikethebiologists.
Forty-
to use the alphabetical equality pattern eightpercentofphysicalchemists'papersare al-
than theirless distinguishedcounterparts: phabetized comparedto 28 per centof papersby
biochemists.
39 per centofpapersby prizewinnersand
12Althoughdifferences
betweenfieldsaresignifi-
46 per cent of the otherscarryalphabetic cant,we are notyetin a positionto explainthem
orders.What we actuallyfindare greater satisfactorily. Data on thetypesofcollaborationin
differencesin theuse of thepatternamong whichlaureateshavebeeninvolvedsuggestthatthe
differentdisciplines.Alphabetizingis more rankstructure of researchgroupsmayaccountfor
them.Physicist forexample,
laureates, aremoreapt
frequently used in physicsthan in chemis- to have workedwithpeersthanare chemistsand
try or the biological sciences,"1
and it is bioscientists
whosegroupstendto be comprised of
much more oftenused by physicistlau- one seniorscientist and a numberofjuniors.Data
on thedistribution
oftypesofcollaborative groups
11It turnsout thatthespecialties
ofchemistry in thefieldsbeingconsidered
areneededto testthis
followthe patternsof theirrespectiveneighbor hypothesis (see Zuckerman, op cit. [n. 3 above],
sciences:laureatesworkingprimarilyin physical p. 313).

This content downloaded from 132.203.227.62 on Fri, 20 Nov 2015 15:03:51 UTC
All use subject to JSTOR Terms and Conditions
280 THE AMERICANJOURNALOF SOCIOLOGY
reate reportedthis familiartype of case authors.In these papers,if no preference
fromhis own experience."This paper I for name orderingwere being exercised,
was co-authorof did not have thisauthor- nameson abouthalfofthepaperswouldbe
ship question in it . . . because . . . the alphabetizedand the otherhalf would be
sequence of authorsin termsof contribu- reverse alphabetical. With three-author
tion was identical with the alphabetical papers,in theabsenceofchoice,about one-
orderand thisput me firstwhichis unusual sixth would have alphabetized orders,
[considering that]mynamebeginswithT." anothersixth,reversealphabetical,and so
Scientistsare abundantlyaware of this on as summarizedin Table 2. A measureof
type of ambiguity.In many cases, the preference foralphabetizedorders-or any
manifest"equality" patternsuppliesmis- one of the othername-orderpatterns-is
information. One of the several laureates thereforeprovided by the ratios of ob-
who collaboratedwith Enrico Fermi em- servedfrequencies ofpapersto thenumber
phasizedthefrequency ofthissymbolically of papers that would be expected if no
inept sequence when workingwith the preferences werebeingexercised.
great physicist: "Anytime I did some- In Table 3, we see that preference for
thing... with Fermi [and the group], alphabetizedname ordersbecomesnotice-
although we always co-authored these able among laureates and the sample of
things. . . in alphabetical order. . . [it] authorslistedin the abstractswhenpapers
has to be understoodthat Fermi did the have fourormoreco-authors. The tendency
lion's shareof the work." to choosealphabeticalordersas the size of
A secondartifactthat makes forambi- the author-setincreasesis much the same
guityin themeaningofalphabeticalorders amongboth groups.With four-author pa-
is theirincreasinglypreferred use as the pers,bothchooseto alphabetizeabout four
numberof co-authorsincreases,since the timesas oftenas wouldbe expectedif this
task of rankingthe contributions of many were not a matterof deliberatedecision.
authors,say,thelaureates,is thenespecial- The ratiosriserapidlyforfive-author papers
ly difficult, unpleasant,and apt to induce to nineteentimesas manypapers by the
conflict. A physicistputit thisway: "When laureates and sixteen times among the
a lot of peopleare involved... the alpha- others.Finally,when papers have six or
betical order. . . certainlysimplifiesmat- more authors, alphabetizingoccurs 270
ters.And I just don't like thisweighingof timesas oftenas would be expected(if it
each person'scontribution. It is almostim- were randomratherthan preferential) for
possible to do." In these cases too, the read- the laureates and 133 times for other scien-
er of scientificpapers is confrontedwith tists. Ratios based on small expectedfre-
alphabeticalorderswhich do not convey quenciestendto be inflated, of course,but
the unequal contributions of co-authors.'3 it is clearthatalphabetizing is preferred as
Table 1 cannottellus whetherthereis a the size of the author-setincreases beyond
preferencefor alphabetical name orders. four.14
There is also an increase,thoughless
This is evidentin thecase ofpapersby two
marked,in preferenceratios for patterns
13 At the same time,the intent of authorsbe- whichheightenthe visibilityof first-and
comesmorevisibleas the numberof authorsin- last-authors. These will be examinedlater.
creases.A two-author papermayindeedbe alpha-
betizedwithout thisbeingunambiguously identifiedThe alphabetically random sequence of
as theprinciple governingthe two-item sequence. names-presumablyreflecting deliberatese-
Withthreeauthors, becomes
alphabetizing percepti-
ble,and thisincreaseswiththenumberofauthors 14Bothlaureates
andthesampleofauthorslisted
up to a pointafterwhichtheadditionofitemsdoes in abstractschoosethe alphabeticalordersignifi-
notadd extrainformation.This,ofcourse,is simply cantlymoreoftenthanwouldbe expected.x2 =
oftheWeber-Fechner
a specialinstance principleof 50.6 forlaureates,X2= 54.6 forsample,1 d.f.,
perception ofpatterns. p < .001forbothgroups.

This content downloaded from 132.203.227.62 on Fri, 20 Nov 2015 15:03:51 UTC
All use subject to JSTOR Terms and Conditions
A STUDYOF SOCIALSYMBOLISMAND ITS AMBIGUITY 281
quencesgaugedin termsof relativecontri- So much for the comparisonof actual
bution-occursomewhatless oftenin suc- authorshippracticeswith patternswhich
cessivelylargerauthor-setsthan would be wouldoccurifauthorswerenotdeliberately
expectedif no preferences
werebeingexer- choosingsomesequencesratherthanothers.
cised."6 We still must deal with the question of
15 These are nonetheless in whether laureates' selections of name-
the mostnumerous
absoluteterms. orderpatternsare differentfromthose of

TABLE 2
PROBABLITIES THAT PARTICULAR PATTERNS OF NAME ORDERING
WiLL APPEAR ON MULTI-AUTHORED PAPERS IN THE
ABSENCE OF EXERCISED PREFERENCES

NUmBEROF AUTHORS

ORDERINGPATTERNS

2 3 4 5 6

Alphabetical...................... .500 .167 .042 .008 .001


Reversealphabetical...... ......... .500 .167 .042 .008 .001
Alphabeticalexceptforfirstname............. . 333 .125 .033 .007
Alphabeticalexceptforlast name.... ......... .333 .125 .033 .007
Alphabeticallyrandom..... ........ ......... ......... .667 .917 .983

TABLE 3
RATIOS OF OBSERVED TO EXPECTED FREQUENCIESOF PAPERSWITH DIFFERENT
NAME-ORDER PATTERNS BY LAUREATES AND A SAMPLE OF
AUTHORS OF PAPERS LISTED IN ABSTRACTS

NUMBER oF AuTHoRs
NAME-ORDERPATTERNS
2 3 4 5 6+*

Alphabetical:
Laureates............. 1.1 1.5 3.9 19.3 270.4
Abstracts............. 1.1 1.6 3.9 15.6 133.3
Reversealphabetical:
Laureates............. 0.9 0.9 0.7 4.0
Abstracts............. 0.9 0.9 0.9 0.6
X+Alphabetical:t
Laureates.............. .......... 0.9 1.0 2.3 8.1
Abstracts........................ 0.9 1.1 1.7 8.2
Alphabetical+X:t
Laureates............. ........... 0.8 0.9 1.2 3.6
Abstracts............. ........... 0.8 1.2 0.8
Alphabeticallyrandom:
Laureates.............. .......... ........... 0.8 0.8 0. 7
Abstracts............. ........... ........... 0.8 0.8 0.8
Numberof papers:
Laureates..... (1,786) (864) (363) (195) (159)
Abstracts..... (7,731) (2,866) (794) (231) (155)

* The incremental
probabilities
governing of paperswithsevenor moreauthorsare
the expectedfrequencies
so smallthatwe used theprobabilities of a six-authorpaper as thebase forcomputingthe numberof expected
papers.
t Theserefer, to thepatternoffirst-author-out-of-sequence
respectively, followedby an alphabeticalsequence
and thelast-author-out-of-sequence
precededby an alphabeticalsequence.
$ No casesappearedin thesecells.

This content downloaded from 132.203.227.62 on Fri, 20 Nov 2015 15:03:51 UTC
All use subject to JSTOR Terms and Conditions
282 THE AMERICANJOURNALOF SOCIOLOGY
rank-and-file scientists.Taking all multi- nearlycost-free.And second,the chances
authoredpapersby bothgroupsand classi- are thatlaureates,evenwhentheyare very
fyingthemaccordingto patternsof name youngor very old, contributedenoughto
ordering,we findthat these distributionsparticularresearchesto be consideredby
differsignificantlyfrom one another.16 theirco-workers as equals. In any case, we
Table 3 showsthat the differences between findthatlaureatesare no moreapt to have
eminentscientistsand theirless celebrated theirnames in sequenceswhichemphasize
colleaguesare concentrated in paperswith thecontributions ofparticularauthorsthan
at least fiveauthors,17 and in particularin are rank-and-file scientists-withtheexcep-
thechoiceofalphabeticalorders."8 Although tionofthelaureates'greateruse ofthelast-
availabledata do not permitus to testthe author-out-of-sequence pattern on papers
validityofany explanationofdifferences, it withmanyauthors.In thissense,the elite
is, as we shallsee,consistent withotherevi- has muchthe same preferences withregard
dencewe have,to assumethatthesediffer- to name orderingas otherscientistson the
ences can be accountedfor in two ways. majorityof theirpublishedpapers. But as
First of all, the laureates' names are so we shall see, laureates differfromtheir
familiar to the scientificaudience that colleaguesin the positionstheyoccupy in
theysacrifice littlebyhavingtheir nameordersratherthanin theirpreferences
relatively
names alphabetizedalong with those of forparticularsequencesas such.
several other scientists.Wherevertheir Special situationsaffecting theambiguity
names come in these sequences,they are of alphabeticalname orders.-Whenalpha-
visible.'9In thisrespect,and as we shallsee betizednameordersare adoptedas a matter
in others as well, adopting conventions of custom,20 scientistswhosenames begin
whichdo not signal theirpre-eminence is withletterslate in the alphabetand those
who collaboratewithdistinguished menare
18X2 = 163.7,4 d.f.,p < .01. The X2 value is
inflated by theabsolutesize of the samplesbeing consistently penalized.21 One physicistlau-
considered. reatespokeout ofhis ownexperience about
17 Differencesbetweenthe twogroupson selec- the consequences ofpersistentalphabetizing
tionof name-order patternson paperswithfewer formenwhosenamescomeat theendofthe
thanfiveauthorsare not significant. x2 forfour- alphabet:"Since publicationis suchan im-
authorpapers= 2.54,4 d.f.,p > .10. For papers portantsourceofjudgmentoftheindividu-
withfiveauthors, X2= 8.65,4 d.f.,p < .10;and for al's work... ifa givenpersonis alwaysput
papers with six or more authors,X2 14.63,
4 d.f.,p < .01. [latein lists]. . . thereis a tendencyforpeo-
18Examination x2 ple to thinkthathe is just a subsidiaryall
of computations indicates
that
thebulkofthedifference between thetwogroupsof the time."
scientistson paperswithfiveor moreauthorsis The laureatesimplythat the allocation
concentrated in the differencesbetweenthemin ofcreditto well-known scientistsforpartic-
theirtendency toalphabetize;
morethan60percent ular contributions-atthe expenseof sub-
ofthex2 valueis dueto thisdifference. ordinates-is especiallyproblematicin the
19 We shallhave moreto say on the matter of
nameordering and visibility It is suffi- 20 In somecircles,thisconvention
ofscientists. has become
cienthereto notethatStephenCole and Jonathan normative; remarked:
oneinvestigator "Gentlemen
Cole findno consistent relationshipbetweenthe publishwithnamesin alphabetical order"(G. R.
visibilityofphysicists' workand theirpositions in Wendt,Science,CXLV [July10, 1964],110-12).
name-order sequencesand thatthelaureates'com- 21 Thishas becomemoreacuteas automated in-
mentsarecompatible withtheColes'findings ("Visi- formation retrievalsystemsare designedto keep
bilityand the Structural Bases of Awarenessin trackofthenamesofthefirsttwoor threeauthors
Science,"AmericanSociologicalReview,XXXIII and to anonymize theothersin thecategoryof et
[June,1968],387-413).A laureatein medicine and al.-"that clubof scientific
unmentionables whom
physiology describedit thisway: "You usuallyno- theworldis toobusyto notice"(Editorial, Journal
ticethenamethatyou'refamiliar with;evenifit's oftheAmerican MedicalAssociation
[May11,19631,
last,it willbe theonethatsticks."' p. 497).

This content downloaded from 132.203.227.62 on Fri, 20 Nov 2015 15:03:51 UTC
All use subject to JSTOR Terms and Conditions
A STUDYOF SOCIALSYMBOLISMAND ITS AMBIGUITY 283
case of alphabeticalsequences.This is one expressedhis dissatisfaction with all the
versionof theMatthewEffectdescribedby alternativesin these words: "It has been
Mertonas "the accruingof greaterincre- triedto rotatethealphabeticalorder;it has
mentsofrecognition forparticularscientific been triedto put [only]the institution[as
contributions to menof considerable repute author] and none of these has worked
and the withholdingof such recognition out ... When you say, 'We will go alpha-
frommen who have not yet made their betical,'and you get a studentwhosename
mark."22 Whenthenameofan eminentman beginswith'A'; aftera while,you get fed
happensto comefirstin an alphabeticalse- up ... No perfectly standardand uniform
quence, the scientificcommunity,which practicehas beenestablishednorcan it be."
tendsto credithim with the workin any We findthat alphabeticalorderingpat-
case, is all the more likelyto do so. One ternsdesignedto symbolizeequalityofcon-
physicistwho has been listedfirstforthis tribution havetheirownshareofambiguity.
alphabeticalreasonconfessedthat it "was How do thingsstandforothername-order
just embarrassing." patterns?
To preventinjusticesthat wouldbe per-
TABLE 4
petratedby consistently alphabetizingau-
thors' names on a series of joint papers, PERCENTAGE OF PAPERS BY LAUREATES AND A
MATCHED SAMPLE OF SCIENTISTS ON WHICH
scientistshave seizedupon the expedientof
THEY ARE FIRST-AUTHORS, BY AGE AT PUB-
rotatingnames throughlists. A geneticist
LICATION*
reported:"With F, it was not at all clear
[howwe should orderour names]because
Matched
we did everything together... So we tend- Age Laureates Sample
ed to alternate. . . I had a generalfeelingof
'Gosh,we're equal; let's just alternateand 20-29......... 49 (103) 38 (34)
30-39......... 40 (304) 38(130)
give the [correct]impressionthat it's an 40-49 ......... 26 (578) 56 (80)
equal partnership.' " 50-59......... 20 (374) 37(131)
When onlya pair of collaborators is in- 60+.......... 18 (222) 25 (64)
volved,rotationofnameson a seriesofpa- All . . 27(1,581) 39(439)
pers-to equalizecredit-is easy to arrange.
As the numberof co-workers increases,the * Papersherereferto thosehavingthreeor moreauthors.

numberofpaperstheymustpublishto give
everyman first-authorship must,of course, NON-ALPHABETIC ORDERS
equal the numberof authors-and if every Lists of authorsappearingout of alpha-
authoris to be givenan equal chanceto oc- beticalorderare presumably intendedto in-
cupyeverypositionin the list,the number dicate unequal contributionsto the re-
of papers requiredwould obviouslybe nI. search.In practice,such orderingis deter-
To put the mattersimply,six co-authors mined both by this criterionand by the
wouldhave to publisha mere720 papersto relativeranksofauthors.
rotateeveryauthorthroughthe sequence. First-authorship.-With these two cri-
Such preciselymathematized allocationsof teriafornameordering, wewouldexpectthe
creditwouldscarcelybe visibleto thescien- laureatesto be first-authors moreoftenthan
tificcommunity and wouldreveallittlemore thosewhose contributions to sciencehave
thanan obsessiveconcernwiththeproblem. been less substantial.When the laureates
Rotation of names is not the only ex- are comparednot to a sample of authors
pedienttriedand discardedin thesearchfor in theabstractsbut to a samplematchedto
an equitableordering pattern.One laureate themforage, fieldofspecialization, organi-
22 R. K. Merton, "The MatthewEffectin Sci- zationalaffiliation, and initialletterof last
ence,"Science,CLIX (January 5, 1968),56-63. name,as in Table 4, we findthatthisis not

This content downloaded from 132.203.227.62 on Fri, 20 Nov 2015 15:03:51 UTC
All use subject to JSTOR Terms and Conditions
284 THE AMERICANJOURNALOF SOCIOLOGY
the case.23The laureatesare listedfirston laureatesand the rankand fileholdsin all
27 per centoftheirmulti-authored24 papers threefieldsof science.But thoughthe lau-
and the matchedsampleon 39 per centof reatesare lessoftenfirst-authors on all their
theirpapers. This is consistentwith the collaborativepapers,it mightstillbe that
Coles' findingthat the correlation between theyearmarktheircontributions by electing
thevisibility ofphysicists and theirpropen- to be firstin thenameorderthatespecially
sitytowardfirst-authorship is insignificant.25 heightens thevisibilityoffirst-author. This
Laureatesare, of course,highlyvisiblebut is, as we notedearlier,the"first-author-out-
not given to placingtheirnames firston of-sequence" pattern(ZABC). Yet heretoo
scientificpapers. The difference between thelaureatesarefirst-authors lessoftenthan
23 Bibliographies werereceivedfromfifty-five of the matched sample; they occupy firstplace
the 123 scientists drawnfromAmerican Men of on 34 per cent of these papers as against
Sciencefrom whomtheywererequested, permittingtheirless eminentcolleagueswho are first
comparison of fortymatchedpairs of scientists. 41 per centofthetime.
Thesepairswerematched in thefollowing respects:
(1) Age: Withinfiveyearsof thelaureate'sbirth
Whydo thelaureatestake thefirstposi-
date. Whenmorethan one bibliography was re- tion less oftenthan those in the matched
turned fora particular match,theoneclosestto the sample?Laureates'contributions to collab-
laureatein age was chosen.(2) FieldofSpecializa- orativeinvestigations are scarcelyapt to be
tionas listedinAmerican MenofScience.Although less significant than those of the matched
thenumbers ofmenin eachspecialty are thesame
for the laureatesand the matchedsample,the sample, and we knowthat theyare highly
proportion of papersare not,sincemenwerenot ranked.Theirname-order practicesand the
matchedforproductivity. (3) Type of Organiza- interview data, takentogether, suggestthat
tionalAffiliation: Laureateswereclassified accord- theyoftenabideby thestandardofnoblesse
ingto whether theyhad workedin independent re- oblige: their rank imposes special obliga-
searchlaboratories, industrial laboratories, or uni-
versities.The scientists in thesamplewerematched tions.PreciselybecausetheNobelprize-or
accordingly. Pelz, amongothers,suggeststhator- the othersymbolsof recognition received
ganizational contextaffects patterns ofpublication beforeit-confers high rank upon them,
(see DonaldC. Pelz,"CreativeTensionsin theRe- theyare expectednot to exercisetheirau-
searchand Development Climate,"Science,CLVII
[July, 1967],160-65;andDonaldC. Pelz andFrank thority and powerto thelimit.And so they,
in
M. Andrews,Scientists Organizations [New more often thantheless distinguished sam-
York:JohnWiley& Sons,1966]).(4) InitialLet- ple, are moved to give co-workers an en-
ter of Last Name: In orderto avoid variations largedshareof the credit.A biochemist ex-
in positionsin authorlists attributable to place plainedit thisway: "If I publishmy name
in thealphabetical sequence, men in the matched
samplewerematched tothelaureates inthisrespect. first,then everyonethinksthe othersare
Takingvariouscriteria ofeminence, suchas receipt justtechnicians ... ifmynameis last,people
of scientific awards,prestigeof departments, or willcreditme anywayforthe wholething,
holdingeditorships or electedpositions in scientific so I wantthemto have a bit moreglory."
organizations, the laureatesare, as a group,sub-
stantially moreeminent thanmenin thematched This remarkand otherslikeit signalthat
sample.However,thesemen whosenameswere the laureatesare well aware that theyare
drawnfromAmerican MenofScienceareprobably not penalized by their distinctivename-
moredistinguished thanmostscientists, sincethat orderpractices.The MatthewEffect, which
directory listsonlyabouta fourthof all scientific
workers. Finally,thereis no reasonto supposethat enhances thestanding ofeminent men at the
menwhoreturned theirbibliographies from expenseof theirless distinguished
differed co-work-
thosewho did not, with respectto name-order ers, assures them of recognition wherever
practices. theyare placed on authorlists.To thisex-
24 Multi-authored papersrefer to thosehavingat tent,theirgenerosity is cost-free.
least threeauthors,unlessotherwise As
specified.
on
practices
thelaureates'publication
notedearlier, 25 ColeandCole,op. cit. (n. 19 above).The Coles'
two-author fromthoseof the computations
papersdo not differ alsoincludesingle-authored paperson
sampleofauthors nordo they theassumption
listedintheabstracts thatvisibility is enhancedbothby
differfromthoseofthematchedsample. first-authorship
and single-authorship.

This content downloaded from 132.203.227.62 on Fri, 20 Nov 2015 15:03:51 UTC
All use subject to JSTOR Terms and Conditions
A STUDYOF SOCIALSYMBOLISMAND ITS AMBIGUITY 285
This interpretation, in termsof noblesse laureates-to-be wererecipients ofthegener-
oblige,is at least consistentwitha morede- osity of theirmentorsand socializedinto
tailedreviewof the data in Table 4 where thepatternofnoblesse oblige.We can identi-
laureates'publicationpatternsat various fyanticipatory socializationin thecase ofa
ages are comparedto thoseof thematched chemistwho engagedin name-order prac-
sample.In theirtwenties,the laureates-to- ticesassociatedwitheminenceand seniority
be were first-authors on half of theircol- earlyin his career:
laborativepapers.Both motiveand oppor- INTERVIEWER: Fromthe earliesttimes,when
tunitystructure joined to producethis re- you published withstudents, I
sult. First of all, they were intentupon noticeyournameappearedlast.
establishingtheirdistinctivecontributionsLAUREATE: Oh, that was just an affecta-
to jointenterprises. Recallingthosedays,a tion... I suppose.But there
biochemistmused: "When you are young are twocustoms . .. eitheryou
and not reallyveryfar advanced,you are put the students'namesfirst
muchmoreexcitedabout[beingfirst-author] orelseit'salphabetical.
thanyou are lateron. It's not veryixmpor- This combinationof having made sig-
tantnow." nificantcontributions-noneof the lau-
And second, when the laureates were reateswas a "late bloomer"-andgenerosity
young,theyworkedwitheminentscientists. on thepartoftheirseniorsmakesfora high
In fact, twenty-sixof the forty-onelau- proportion offirst-author listingsamongthe
reateswho wereinterviewed workedunder laureates'paperspublishedwhentheywere
men who were,themselves,Nobel prize- young.28 Different values of the same two
winners.And ofthetotaloffifty-five Nobel variablesare tentativelytaken to account
laureatesat workin theUnitedStates,thir- for the lower percentageof first-authors
ty-four workedunderolderlaureates.Their amongthematchedsample.Theircontribu-
mentorsoftenmade decisionsto put young tions to collaborativeresearchwere prob-
menfirstjust as the presentlaureatesnow ably less significant and theirownmentors
do.26As one laureateremembered: "When probablyless generousthan those of the
I workedwithT [an olderprizewinner] it laureates.
usuallyturnedout that I was listedfirst." As both the laureatesand the matched
His experienceis not unique. When the sample take controlover name-orderde-
laureates-to-beco-authoredpapers with cisions-and the interviewsindicate that
theirlaureatemasters,it turnsout thatthe thesedecisionsare generally made by senior
juniormenwerefirst-authors 60 percentof investigators-thelaureates place them-
thetimeand theseniorsonly16 percentof selvesfirstless oftenand thematchedsam-
the time.27Simultaneously,the young ple moreoften.By theirforties, mostofthe
2BLaureates nottrainedunderpriorprizewinnerslaureatesare distinguished men,and halfof
norunderotherequallyeminent menwereapparent- them have already received the Nobel
ly nottreatedso well.One ofthemobserved:"Oh, prize. They are, by this time,aware that
I've been a secondauthorall my i.fe.... First
you'resecondauthorbecauseyou'reyoungandthen theirvisibilityand eminenceare not di-
you'resecondauthorbecauseyou'retheseniorman
in thegroup."His reportis trueenoughforhim-he and oneyoungscientist and laterprizewinner-the
on only12 per cent of his papers. juniormanis first-author
is first-author 68 percentof thetime,
27Neitherlaureates-to-be nor their laureate moreoftenthanwouldbe expected ifnopreferences
masterswerefirst-authors on theremaining 24 per werebeingexercised.
centofpapers.Thissituation is a markedcontrast 28 Laureates describingtheir own attitudes
to papersco-authored by scientistpeers,bothof toward talentedyoung scientistssuggestthat
whomlaterbecamelaureates.On thesepapers,one generosity is easierto bestowuponthesemen(who
futurelaureateis justas aptas theotherto be first- werenotunlikethelaureatesat thesameages)than
authorin papershavingat leastthreeauthors.In upon men who have neitherthe abilityto solve
two-authorpapers-thosewithonelaureatementor problems norto formulate questionsforresearch.

This content downloaded from 132.203.227.62 on Fri, 20 Nov 2015 15:03:51 UTC
All use subject to JSTOR Terms and Conditions
286 THE AMERICANJOURNALOF SOCIOLOGY
minishedby their name-orderpractices. erably followingthe Nobel prize which,
Finally,as the scientistsbecome older, afterall, confersmuch prestigeupon its
the issue of name ordering,in particular, recipients. Table 5 showsthatthelaureates'
and competition forrecognition, in general, ratesof first-authored papers,alreadylow,
declinesin importancefor both groups.29declinesome9 percentfollowing theaward;
Furthermore, thecontributions ofoldermen the publicationpracticesof the matched
to particularinvestigationsareprobablyless sample do not change at all in the same
significant.Again,in the interviews, older period.3'
laureatessay with regretthat researchis This change,followingthe Nobel prize,
tiringand that, increasingly,they now is not as greatas we wouldhave expected,
serrvein a supervisory capacityY3 given the very substantialprestige be-
Up to now,I have ascribedthe inverse TABLE 5
relationshipbetween first-authorship of
PERCENTAGE OF MULTI-AUTHORED PAPERS ON
papers and age at publication among
WHICH LAUREATES AND A MATCHED SAMPLE
laureates to their growingeminenceand OF SCIENTISTS ARE FIRST-AUTHORS ACCORD-
visibility.Someadditionaldata supportthis ING TO TIME OF PUBLICATION
interpretation. First of all, rates of first-
authorshipshould decrease quite consid- DATE OF PUBLICATION RELATIVE

29 These changesin name-order


TO TIE AWARD*
practiceswith
agingand the laureates'reportsof corresponding
changesin theirinvolvement withit suggestthat 6 or More 0-5 Years 1 or More
first-authorship mayhavedifferent consequences for Years be-
before Years after
thevisibility ofscientists at different stagesofthe forethe
Award the Award
scientific
career.Although the Coles findno asso- Award
ciationbetweenname-order practicesand visibility
fora sampleofphysicists whoseagesspanthework- Laureates ..... 39(688) 22(477) 13(416)
Matched
ingyears,it maybe thatthetwovariablesaremore sample..... 43(153) 35 (82) 36(204)
stronglyrelatedat earlyphasesofthelife-work cycle
thanat laterones,whenvisibility has alreadybeen
achieved. * Time of publicationof papersof themenin thematched
samplewas determined by usingthe date of the prizeof the
laureatein each matchedpair. A paper by a memberof the
30 Even thoughlaureates are first-authorspro- matchedsamplepublishedin 1952,forexample,was classified
portionately lessoftenas theyage,one mightsup- as beinginthecategory0-S yearsbeforetheawardifhislaureate
pose thattheywouldcontinueto emphasizetheir twinreceivedtheprizebetween1952and 1956.
contributions bybeingfirst on thetwonameorders
whichenhancethevisibility ofthefirst-author-thestowedbytheaward.My analysis,however,
first-author-out-of-sequence and the alphabeticallydisregards thefactthatnot all Nobel laure-
randompatterns. The laureatesdeclineto empha- ates were equally distinguished beforere-
size theircontributions in thedoublesenseoffore- ceivingthe prize. Some were alreadyun-
goingfirst-authorship and havingan increasing
shareoftheirpapersovertimecarrythissequence, 31Sincethesedata and thosereported inTable4
thereby further enhancing thevisibility ofworkof are drawnfromthesameinformation, it is scarcely
theircollaborators. surprising thatthesamepatternof declining first-
authorship is shownin each. Whenlaureatesare
PERCENTAGE OF PAPERS HAVING NAME to age at thetimeoftheprize,
ORDERS WHCH ENuANcE THE Visnsm-
classified according
ITY OFFIRST-AUTHORS ON WHICH LAU-
wefindthatmenwhoreceived theawardbefore they
REATES AND A MATCHED SAM&PLE OF werefifty werefirst-authors moreoftenin theperiod
SCIENTISTS AR-E FIRST-AUTHORS, AC- justbeforetheawardthanwereolderlaureates.In
CORDINGTO AGE AT PUBLICATION thefiveyearsbeforetheprize,28 percentof the
youngwinners' paperswerefirst-authored by them
and 16 per centof thoseby oldermen.Afterthe
Ageat PublicationLaureates Matched
award,ratesof first-authorship are just thesame,
20-29............ 50 33 14 percentforyounger winners and 13percentfor
40-49............ 26 63 olderones.Differences betweenlaureatesand men
60+.............. 18 11
in thematchedsampleholdin eachfieldforevery
timeperiod.

This content downloaded from 132.203.227.62 on Fri, 20 Nov 2015 15:03:51 UTC
All use subject to JSTOR Terms and Conditions
A STUDYOF SOCIALSYMBOLISMAND ITS AMBIGUITY 287
crownedlaureates;they were membersof siderableincrements of prestigeupon these
theirown national academiesand foreign men.I want,then,to comparethe author-
ones,recipientsof major scientific awards, ship practicesof men who achieved sub-
and holdersof multiplehonorarydegrees. stantialeminencebeforereceivingtheprize
For thesemen,the Nobel prize only caps to thoseofmenwhohad not,in orderto see
distinguished researchcareers.Others,how- whetherthe exerciseof noblesseobligeis in-
ever,wereneitheracademicians,recipients deed more characteristic of eminentmen.
of importantawards,norholdersof honor- As Figure1 indicates,menwhowereemi-
ary degreesbeforetheywerenamedlaure- nent beforereceivingthe prize began to
ates.32The Nobel prize confersverycon- to theirco-workers
transfer
first-authorship
32 Twenty-sevenwere classifiedas eminentprior beforetheirless eminentcolleaguesdid so."
to the award of the prize. All were membersof the
National Academy of Sciences or of foreignacad-
cludeall forty-one laureateswhoreceivedtheprize
emiesand had, on the average,twiceas manymajor
prizes and medals as the non-academicians.More- before1964(see p. 399,H. A. Zuckerman, "Nobel
over, theyreceivedeightyhonorarydegreesamong
Laureatesin Science:Patternsof Productivity,
themcomparedto the total of one awarded to the
Collaboration andAuthorship,"AmericanSociologi-
cal Review,XXXII [June, 1967],391-403).
others.My earlierexaminationofchangesin produc-
tivityfollowingthe Nobel award for scientistsal- betweenthetwogroupscannotbe
33 Differences
ready eminentand forthose who were not is based attributed in timeelapsedbetween
to differences
only on data forlaureatesnamed in 1960 or before. publicationof prizewinning researchand the re-
These computationson name-orderpractices in- ceiptoftheprize.Scientists,alreadyeminent at the
50

45

40

o
0 35
H
< 30

= 25
U-

< 20
LAUREATES NOT
z EMINENT BEFORE AWARD
ui 15
w
0- LAUREATES EMINENT
10
BEFORE AWARD
5

0
6 OR MORE 0-5 YEARS 1 OR MORE
YEARSPRIOR PRIOR TO YEARS AFTER
1O PRIZE PRIZE PRIZE

DATE OF PUBLICATION
papersby laureateson whichtheyare first-authors,
FIG. 1.-Percentageofmulti-authored according
to dateofpublication
andeminenceat timeoftheaward.

This content downloaded from 132.203.227.62 on Fri, 20 Nov 2015 15:03:51 UTC
All use subject to JSTOR Terms and Conditions
288 THE AMERICANJOURNALOF SOCIOLOGY
In the periodjust beforethe award,laure- more frequentlyas the eminenceof indi-
ates who werealreadydistinguished scien- vidualscientists
increases.It doesnotimply,
tiststookfirstplace in authorlistsonly 17 however,thatlaureatesare entirelyselfless
per centof the time,but menwhohad not or whollyunconcernedwith recognition;
yet made theirmarkdid so twiceas often. they "stake theirclaims" withouthesita-
Differencesbetween the two groups of tion to work to whichthey believe they
laureates,moreover, cannotbe attributed to made signal contributions. Two laureates
variations in name-orderpractices with emphasizedthisconsideration whendescrib-
aging.Takingthosewho receivedtheprize ing name-ordering patternson papers re-
beforetheywerefifty, forexample,we find portingresearchforwhicheach receivedthe
that the previouslyeminentwere first- prize: "I thinkboth Joe and Jim[histwo
authorson only15 per centof theirpapers co-workers] regardedme as theoriginator of
publishedjust beforetheaward,buttheless the idea and the main promoterof the ex-
eminentlaureates-to-be werelistedfirst48 periments. I have neverfeltwitheitherof
per centofthetime.Afterthe award,first- them,theslightest awkwardness as towhose
authorshipis farless frequent-10per cent idea it was or anythinglike that ... And
ofthealreadyeminentand 19percentofthe this was why I was first.. . But I would
newly eminentlaureates' papers carried also liketo say thatwithoutany one ofus,
theirnames first.34 In this respect,future it would not have been carriedthrough."
laureates-youngand old-who had notyet Justas directly, theotherlaureatereported:
becomemembersof the scientific elite are "Why was my name first?... I suppose
muchliketherank-and-file scientistsin the one should publish in alphabetical or-
matchedsample.These laureates-to-be pre- der.... [but]it was reallymy job that I
sumablyfeltobligedto earmarktheircon- had been workingon and thenhe came in.
tributionsto collaborativeeffortsto an ex- So I don't thinkit was unfairto him. He
tentthat theireminentcolleaguesdid not. did supplya keypieceofthething[but]his
Aftertheaward,thepracticesoftheprevi- time on it was small comparedto mine
ouslyeminentand the newlyeminent-re- anyway."
gardlessofage-converge:theyare first-au- These laureatessuggestthat theyadopt
thors12percentand 16percentofthetime, some symbolicdevice for earmarkingsig-
respectively.The Nobel prize apparently nificantcontributionsas their own. We
goesfartowardequalizingearlierdifferenceswould expectthesedistinguished scientists
in prestigeamonglaureates:"it also mini- to use one or anotherof the highvisibility
mizesthe tendencyamongprizewinners to positionsin name orders-eitherthe first-
highlighttheirown contributions to joint author-or the last-author-out-of-sequence
researchthroughsymbolsof authorship. or firstplace in the alphabeticallyrandom
These data are consistentwiththe sug- patternmoreoftenon paperswhicheven-
gestion that noblesseoblige is exercised tually broughtthem the prize than on
otherstheydeemedless significant.36 And
time of the award, received it nine years on the
average after publication, and the otherswaited 35 This is not to say that all laureatesare equally

more than seven years for the prize to be given. esteemed after the award. Scientistsare wont to
speak of "strong"prizes and "weak" ones-that is,
34Analysis of eminenceand authorshippractices
prizes which were awarded to men of unquestion-
of menwho receivedthe prizebeforethe age of fifty
able stature and accomplishmentand prizes given
and afterit suggeststhat eminenceis moreimpor-
to scientistswhose selectionby the Nobel Founda-
tant thanage in shapingauthorshipdecisions.There
tionwas not a foregoneconclusion.
are a fewpapersby the veryfewmenwho were more
than fiftyat the timeof the prize and not yet emi- 38 In retrospect,thelaureatesindicatedthat they
nent. The rewardsystemin science appears to be werewellaware oftheimportanceoftheworkwhich
sufficientlyeffectiveto minimize the number of eventuallybroughtthemthe prize. As one of them
men who have done significantwork but have not put it: "[I knew]it was great stuff.Absolutely...
achieved eminenceby the age of fifty. I mean a great piece of physics. . . It would have

This content downloaded from 132.203.227.62 on Fri, 20 Nov 2015 15:03:51 UTC
All use subject to JSTOR Terms and Conditions
A STUDYOF SOCIALSYMBOLISMAND ITS AMBIGUITY 289
it turnsoutthatthisis thecase. Takingonly authorshipof thelaureate-to-be or withhis
those papers reportingprizewinningre- name in a positionof highvisibility.The
search,we findthat laureatesoccupyhigh laureates'reluctanceto foregoall symbols
visibilitypositionsin name-order sequences of theirprimeachievementsuggeststhat
46 per centof the timecomparedto 30 per noblesse obligehas its limits.
centon all otherpaperstheyhavepublished Last-autlzorship.-Itmight be assumed
withthesesamename-order patterns.37Nor that thenormativecriteriaforlast place in
do theystop here.They publishpapersre- authorlistsare reciprocals to thoseforfirst
portingaward-winning researchundertheir place; thatauthorswhocontribute leastand
names alone-still anothermeans of en- ranklowestshouldcomelast.Yet thelaure-
hancingthevisibility oftheirowncontribu- ates are last-authorsmore oftenthanthe
tions.8 Taken together,approximately 60 membersof the matchedsample;in 53 per
per cent of all papers that reportedprize- cent of theirpapersas against33 per cent
winningresearchappearedunder the sole fortheothers.39
The practice of these laureates again
graced any universityin the universe,it was that expressesthe norm of noblesseoblige.A
good." A fair number of them were ambivalent laureatein medicinesaid: "My namealways
about theparticularworkcitedfortheNobel prize- comeslast . . . I did receivequite a bit of
that is, they did not think that it was their best
work-but none of them suggestedthat the work recognition forthe workI did on X com-
honoredwas insignificant. These findingsare only pound [notthe workforwhichhe won the
apparentlyinconsistentwith the Coles' reportthat prize].Justso longas mynamewas there,it
positionin name ordersand citationsare unrelated. didn't make any difference where it ap-
They limitedtheirinvestigationto first-authorship
and single-authorship and did not take into account
peared. Don't you think that's true? So
eithername-orderpatternor last-authorship. why not put it last?" And anothereven
described it as a matterofetiquette:"You'll
Laureates symbolize their distinctivecontri-
find in recent yearsthatmynameis always
37

butions to prizewinningwork more oftenthan on


otherpapers. First,theyare slightlyless apt to use last. This isn'tbecauseI thinkI have con-
alphabetical orders in these papers than on other tributedleast. But I thinkthat, as senior
publications-19 per cent as against 26 per cent. personin thelab, it's moreappropriate that
More important,they place themselvesin highly
visible positionson those papers with name orders
it be there."
whichenhance the visibilityof a particularauthor. As we noted earlier,one name order,
PERCENTAGE OF PRIZEWINNING PAPERS AND OF OTHER
the last-author-out-of-sequence, actually
PUBLICATIONS ON WHiCH LAUREATES OCCUPY heightensthe visibilityof the last-author.
POSITIONS OFHIGH VISIBILITY It appears equally as oftenon multi-au-
thoredpapersby laureatesand thematched
Name-Order Pattern Prizewinning
Papers Other Papers sample, about one-fifth ofthe timeforeach
First-author out-of-sequence. 54 (120) 23 (233)
group. Once this pattern has been chosen,
Last-author
Alphabetically
out-of-sequence....
random ........
69 (107)
23 (156)
55 (185)
18 (288)
however, the laureates are more apt than
All........................ 46 (383) 30 (706)
other,less eminentscientiststo use it as a
device for signalingtheirdistinctivecon-
tributions. As Table 6 shows,laureatesare
38 As I have noted above, single-authored papers last-authors almosttwice as oftenon this
are rarelytheproductof one investigatorat workin
isolation. Single-authoredpapers, especially those
high visibilitypatternthan are men in the
summarizingjoint research or reportingspecific matchedsample,61 per centof the timeas
parts of it, are generallynot submittedforpublica- against 34 per cent.40 Here again, noblesse
tionwithoutconsultingwithco-workers.It is fairto
assume that collaborators of the laureates-to-be 39The differencebetweenthe two groupsis con-
accepted these single-authoredpublicationsas one centratedin chemistryand the biological sciences
of their legitimateprerogatives(see Zuckerman, forreasonsthat remainobscure.
"Nobel Laureates in the United States..." [n. 3 40 Althoughwe knowthat thelaureatesare, more
above], pp. 288 ff.). oftenthan not, the most eminentmembersof their

This content downloaded from 132.203.227.62 on Fri, 20 Nov 2015 15:03:51 UTC
All use subject to JSTOR Terms and Conditions
290 THE AMERICANJOURNALOF SOCIOLOGY
obligeapparenitly is exercisedto a pointand workingin collaboration.The variouspat-
no further.Moreover,the laureatesadopt ternswhich,at first, seemto haveunequivo-
thispatternof visibilityincreasingly often cal meaningsforscientistsdo conveyvary-
as theyget older,therebyearmarking their ingdegreesofambiguity and,to thisextent,
contributions withoutdemandingfirst-au- failto servetheirintendedfunctions. How-
thorship.In the biologicalsciencesand in ever,it is also the case thatsuchambiguity
chemistry, for example,the laureatestell in the meaningof name ordersreducesthe
us, last-authorship has prestige,even when stressof collaboration.As in otherdepart-
name ordersdo not call attentionto it. A mentsof social life,makingthingsexplicit
medicalresearcher remarked:"In our field, oftenintroducesstrainintosocial relations.
thelastnameofa groupis the'father'ofthe Laureates, having been faced with de-
effort.... This has become a patternin cisionson name orderingand theirconse-
medicineso thatthelast name does have a quences, are well aware of the tensions
certainposition."Our data bear this out; whichaccompanyemphasisonstrictweight-
ing of contributions. They agreethat some
TABLE 6
measureofambiguity in thisdelicatematter
PERCENTAGE OF PAPERS ON WHICH LAUREATES of allocatingcreditand responsibility helps
AND A MATCHED SAMPLE OF SCIENTISTS ARE to maintainpeace in collaborativesitua-
LAST-AUTHOR, WHICH ARE SO ORDERED AS
tions.A physicist'sremarksare representa-
To HEIGHTEN THE VISIBILITY OF THE LAST-
tiveofgeneralfeelings:
AUTHOR, ACCORDING TO FIELD*
I think[a standardsystemfordistinguishing
thecontributions ofco-authors] wouldtendto
Field Laureates Matched
Sample producea greatdeal of interference.... If a
systemwereset up to tryto dispensecredit
Physics.............. 52 (21) 41(27) withcomplete justice,I thinkwe'dendup with
Chemistry ........... 63(125) 10(10) almosteveryone . . . disappointed,in general,
Biological sciences.... 59(143) 35(55)
that other
withwhatthey'vegotten-feeling
All ............. 61(289) 34(92) peoplehave perhapsgottenmorethan their
share.. . . So I wouldbe opposedto trying to
The last-author outofphase cut thisveryfine.
in thispatternis alphabetically It seemsto me thatthede-
and is precededby an alphabeticalsequence(XYZA). cisionas towhether somebody shouldbe senior
authoror shouldappearsimplyalphabetically
laureatesin the biologicalsciencesand in is perhaps as fineas itshouldbereasonably cut.
chemistryare more often"fathers"of in-
vestigations;theyare last on 62 and 59 per It appearsthat,in the long run,the func-
centofthepapers,respectively, withnames tionalrequirements oftheevaluationsystem
listedin the alphabetically randompattern, and of collaborativegroupsare incompati-
comparedto laureatesin physicswho were ble and so call fordifferent and sometimes
last authorsonly28 per centofthetimeon conflicting procedures for making role-
papersofthe sametype. performance visible.
I have shownthat patternsof name or- One matterremainsfordiscussion.Until
deringhave beendevisedas adaptivemech- now, we have assumed that position in
anisms to heightenthe visibilityof role- name-order sequencesenhancesthevisibili-
performanceof individual investigators ty of individual role-performance and there-
bymakesitpossibleto allocatecreditamong
research groups, we do notknowwhether thisis the co-workers. Whetherthis is so or not, the
case formenin thematched sample.It maybe that fact is that scientistsbelieve that it does.
the mostdistinguished co-workers of the sample As with otherdefinitions of the situation,
members arelast-authors on papersofthistypeas
oftenas thelaureatesare. The reported differences consequences are apt to reflectsuchbeliefs.
in last-authorship arethensuggestive, notdecisive. It seemsto be the case that eminenceand

This content downloaded from 132.203.227.62 on Fri, 20 Nov 2015 15:03:51 UTC
All use subject to JSTOR Terms and Conditions
A STUDYOF SOCIALSYMBOLISMAND ITS AMBIGUITY 291
name-order positionsjointlyaffectvisibility amongfieldsand amongmen at different
in such a way that occupyingvisible po- stagesofthescientific career.41
sitionsin name-ordersequencesis impor- Since thisdeviceservesits purposeonly
tant forthe youngand unknownbut less partially, we should expect continuing
so forscientistswhosenames are familiar attempts to modifyexistingname-order
to their colleagues. If so, then there is practices-as wellas continuing dissatisfac-
some isomorphismbetween the conse- tion with them-until otherarrangements
quences of name orderingand differentialaredevisedwhichintegrate moreeffectively
concernwiththe matter.Finally,if name- the functionalrequirementsof research
orderpositionsdo not affectvisibilityin groups and of the evaluation systemin
any consistentfashion,they have latent science.41
functionsforco-workers. Individualcollab-
oratorsoftenwant symbolicrecognition of COLUMBIA UNIVERSITY
their unequal contributions,if only to 41Cross-culturalcomparisons willshowwhether
"keep the recordstraight"and to maintain thesepatterns haveuniform meanings fortheworld
equitywithinthe group.This may explain community ofscientists
and,in particular,
whether
whydiscussions of name orderingare any- the distributionof patternsdiffers
between social-
istandothersocieties.
neutral,evenwhenco-
thingbut affectively 42 The verygreatconcern voicedby physicists
workersbelieve that the publishedpapers about the use of "institutional
authorship"-that
are routine. is, thelistingoflaboratoriesratherthanindividual
I have noted that certainpatternsof investigators-suggests thatmorevisibility,
rather
name orderingare variouslyadopted by thanless,is desiredbyscientists. Forthecomments
of one distinguished on thismatter,see
physicist
eminentmen more oftenthan by others, S. A. Goudsmit,PhysicalReviewLetters,VIII
that preferencesfor name orders differ (March15,1962),229-30.

This content downloaded from 132.203.227.62 on Fri, 20 Nov 2015 15:03:51 UTC
All use subject to JSTOR Terms and Conditions

Potrebbero piacerti anche