1995 - The Thomas Theorem and The Matthew Effect

Social Forces, University of North Carolina Press
The Thomas Theorem and the Matthew Effect

Author(s): Robert K. Merton
Source: Social Forces, Vol. 74, No. 2 (Dec., 1995), pp. 379-422
Published by: Oxford University Press
Stable URL: http://www.jstor.org/stable/2580486 .
Accessed: 27/05/2013 17:41
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.
Oxford University Press and Social Forces, University of North Carolina Press are collaborating with JSTOR
to digitize, preserve and extend access to Social Forces.
http://www.jstor.org
This content downloaded from 79.175.121.210 on Mon, 27 May 2013 17:41:43 PM

All use subject to JSTOR Terms and Conditions
The ThomasTheorem
and The Matthew Effect*
ROBERTK MERION, ColumbiaUniversityand
RussellSageFoundation
Eponymyin scienceis the practiceof affixingthe namesof scientiststo what they have
discoveredor are believed to have discovered,'as with Boyle's Law, Halley's comet,
Fourier'stransform,Planck'sconstant,the Rorschachtest, the Gini coefficient,and the
Thomastheorem.
This article can be read from various sociological perspectives.2 Most specifically, it records an epistolary episode in the sociointellectual history of what has
The definition of eponymyincludes the cautionary phrase, 'or are believedto have
discovered,"in orderto take due note of "Stigler'sLaw of Eponymy"which in its strongestand
"simplestform is this: 'No scientific discovery is named after its originaldiscoverer"'(Stigler
1980).Stigler'sstudy of what is generallyknown as "thenormaldistributioneor "theGaussian
distribution"as a case in point of his ironicallyself-exemplifyingeponymous law is based in
part on its eponymous appearancein 80 textbooksof statistics,from 1816 to 1976.
2As will become evident, this discursive composite of archival documents,biographyof a
sociological idea, and analysis of social mechanisms involved in the diffusion of that idea
departs from the tidy formatthat has come to be prescribedfor the scientificpaper.This is by
design and with the indulgent consent of the editor of SocialForces.But then, that only speaks
for a continuinglargenessof spirit of its editorialpolicy which, backin 1934, allowed the ironic
phrase "enlightenedBoojumof Positivism"(with its allusion to Lewis Carroll'simmortal The
Huntingof theSnark)to appear in my very first article,published in this joural better than 60
years ago.
*
I am indebted,once again, to HarrietZuckernan,Robert C Merton,CynthiaFuchs

Epstein,David L. Sills, and StephenM. Stiglerfor vetting a manuscript,to JenniferLee
andMaritsaPorosfor researchassistance,and to EugeneGarfieldforaid of otherkinds.
Direct correspondenceto Robert K. Merton, East Gallery, Low Memorial Library,
ColumbiaUniversity,New York,NY 10027.
C)The University of North CarolinaPress
Social Forces, December1995, 74(2):379-424

380 / Social Forces 74:2,December1995

"if men definesituationsas real,
come to be known as "theThomastheorem":3
they are real in their consequences"(Thomas & Thomas 1928:572).More
generally,that episode provides a strategicresearchsite for examning certain
substantiveand methodologicalproblemsin the sociology of science.Fromthe
methodological perspective, it provides a prime example of the basic if
presumablyobvious preceptthatit is one thingto establisha phenomenon(i.e.,
show that somethingis empiricallythe case) and quite anotherto explain it.
Obviousthis may be but, as we shall see, the two are neverthelesseasily (and,
I believe, often) conflated. The episode also exhibits the risks involved in
reductionist,single-factorexplanationsof a concretesocialphenomenon(which
becomeseven more markedin ex postthanin ex anteexplanations).Finally,the
episode provides an apparentinstanceof how socioculturalcontextsof science
and scholarship- in this case, the belatedthrustof the civil rightsmovement
toward equity for women - can make for an exclusive and premature
interpretationthat a particularcognitivephenomenonis sexist.
The cognitive phenomenonin this case consists of sociologists'frequent
ascriptionof the Thomastheoremsolely to W.I.Thomasratherthanto bothW.I.
and DorothySwaineThomas.I should emphasizefromthe start,however,that
thereis only incidentalinteresthere in tryingto adjudicateproprietaryclaims
to this basic sociological idea, although the introductionof private archival
materialsmay contributeto thatresult.Rather,the widespreadaccreditationof
the Thomastheoremto W.I.Thomasalone holds interestfor us hereprincipally
as a specificinstanceof a genericphenomenonin the reward-systemof science
and scholarship- what can be convenientlydescribedas "thepartialcitation
phenomenon,"thus substitutingfour words for the approximately20 words of
its definition,i.e., the widespreadaccreditingby scientificand scholarlypeers
of an (actuallyor apparently)jointcontributionto only a subsetof the collaborators.4
Some forms of the partialcitationphenomenonalmost provide their own,
intuitively evident, explanations.The partial citation of coauthorsis hardly
problematical,of course,for the rapidlyincreasingnumberof scientificarticles,
3 As I have noted elsewhere (Merton1984:282),the designationThomas theorem"does not,

of course, adopt the term theoremin the strict mathematicalsense (as, say, with the binomial
theorem).It refers,rather,to an idea that is being proposed or accepted as sound, consequential, and empirically relevant." In proposing the Thomas eponym for both mnemonic and
commemorativepurposes (Merton[1942]1973:273),I had fastened on the term theoremrather
than such less formidabletermsas dictum,maxim,proposition,
or aphorismin orderto convey my
sense that this was "probablythe single most consequentialsentence ever put in print by an
American sociologist" (Merton 1976:174).In any case, the word theoremwas rhetorically
employed in the same broad sense that had the mathematicallyminded Hobbes refering in
The Leviathanto "general rules, called theorems or aphorisms." (Reading this note, the
mathematicalstatistician,Stephen M. Stigler, reminds me that the redoubtableseventeenthcenturymathematicianJohnWallis - he of the Wallis theorem- had ample cause to destroy
the pretensionsof the mathematicallyminded Hobbes to being an actual mathematician.)
4 The word citationin the termpartialdtationphenomenon
is to be construedbroadly;not only
as a fonnal reference in a note or bibliographybut as any mode, including eponyms, of
referringto previous scientific or scholarlywork.

The ThomasTheoremand the MatthewEffect/ 381

chiefly in the physical and biological sciences,with large numbers- at the
extremes,literally hundreds - of listed coauthors (ScienceWatch 1995:1-2;
Zuckerman1968).Nor is it greatlyproblematicalin cases of worksby, say, four
or five authorswhere the citationpatternof "et al." has long since evolved in
the world of scholarshipto serve the sharedconvenienceof publishers,editors,
senior authors, and readers (if not the neglected collaborator6).Nor is it
problematicalthat citations of large-scale research reports often have the
institutionalizedsenior investigatoralone eponymized, as with The Kinsey
Reportor The ColemanReport.But the partialcitationphenomenonis surely
problematicalin the limiting case, as with the book in which the Thomas
theoremfirst appeared,when only one of two authorsis regularlycited.
The equitablepeer ascriptionof contributionsis no minor matterin the
social institutionof science which has evolved a reward system that consists
basically in rewardingscientistsby having knowledgeablepeers grant them
public recognitionfor their distinctivecontributions.All other rewardsflow
from it.5And so it is thatpeers will experienceand sometimespubliclyexpress
strong moral sentimentswhen they have reasonto believe - as in the case of
the Thomastheoremsome have felt thereis reasonto believe - that the norm
of rightfulaccreditationhas been violatedby systematicallybiassedascriptions
in the pertinentcommunityof scientists.Forlike all othersocialinstitutions,the
institutionof science has its (partlymanifest,partly latent)normativeframework, one that includesthe norm of equityin peer recognitionof contributions.
Along with letters exchanged some time ago between Dorothy Swaine
Thomasand myself,the core archivalmaterialsin this study of the allocationof
creditfor the Thomastheoremconsistprincipallyof a more recentexchangeof
letters between anotherpair of collegial sociologists which has one of them
interpretingthis instanceof the partialcitationphenomenonas a "pieceof (dare
I call it?) institutionalizedsexism ... in the new era."These letters - one of
them, crisp and pointed by a social scientist of amply merited international
fame;the otherby myself, and repletewith documentaryexhibitsrunningto a
good-sized articlein its own right - will be quoted verbatim.6As the more
voluble memberof thatepistolarypair,I shall subjectthatexchangeto analysis
in terms of patternsin the growth and transmissionof knowledge that I have
been investigatingover the years:such pattems as "establishingthe phenomenon" and the use of "strategicresearchsites" (Merton1987);"the retroactive
effect"in perceivingor imaginingadumbrationsand anticipationsof ideas in
5Forthe paradigmwhich maintains that the institutionaldynamicsof science derives from

the interactionof its "rewardsystem" and its "normativestructure,"see Part IV of Merton
1973.
6An apt procedure,one would think, since the sociological analysis of verbatimletters was
introducedby W.I. Thomasand FlorianZnaniecki;though, to be sure, on a ratherlargerscale
runningfrom page 217 to page 1114 of their classic work in five volumes, 7hePolishPeasantin
Europeand America([1918-20]1927). For a critical examinationof methodologicalproblems
involved in the use of letters in ThePolishPeasant,see Blumer1939:29-39.


the history of thought;7patternsof primaryand derivativeor serial diffusion
of knowledge (Merton1995);"sociallyorganizedskepticism"in the domainof
science and scholarship (erton [1942] 1973:267-78,339-40);the place of
eponyms (such as the Thomastheorem)and of otherforms of peer citationin
the reward system of science (Merton[1942]1973:273-74;[1957]1973:297-302;
"oral publication"as distinct from publicationin print (Merton
1988:619-23);
1980);"multiples"or independentmultiplediscoveriesand inventions(Merton
1973;Ogburn& Thomas1922)and the emergenceof citationanalysis (Garfield
1955;Garfield,Sher& Torpie1964).Above all else, however,the partial-citation
phenomenonof the Thomastheoremwill be analyzedin termsof the Matthew
effect (Merton& Zuckerman[1968]1973;1988).
This epistolaryexchangewhich involves contestinginterpretationsof this
one case of the partial citation phenomenonalso raises normativequestions
aboutthe peerrecognitionof scientificand scholarlycontributionsthatprovides
the ultimatesocial and moral,if not the legal, basis of intellectualpropertyin
science (Zuckerman1988a:526-27).
(As we have just seen in preview,such analysisin termsof one's enduring
thematic interests in the sociology of science is bound, alas, to entail an
intemperateabundanceof self-citation.It is some consolationto note, however,
that disciplinedcitationanalysesdo not mistakeself-citationas evidenceof the
peer recognitionthat is the ultimatecoin of the domainof scienceand scholarship.)
Diffusion of the Thomas Theorem
Like other important sociological ideas, the Thomas theorem has had its
adumbrationsandpartialanticipations.Recognitionof the subjectivecomponent
in humanactionhas had a long historyin sociologicalthoughtand a far longer
histoxy before we sociologists arrived on the historical scene. Among the
[c. 110 A.D.] 1926-28:
Ancients,we need only recall Epictetus(TheEncheiridion
II, 487, ?5), stating that "Whatdisturbsand alarmsman are not actions, but
opinions and fancies about actions."And if one has not had occasionto read
Epictetus recently, his aphorism (theorem?)may be recalled from the still
enduring eighteenth-centurymasterwork,TristramShandy,where Laurence
Steme quotes it - in Greek of course - on the title page. And among the
Moderns,there is Schopenhauer([1851]1974:1,326) also echoing Epictetus8as
he observes that "it is not what things objectivelyand actually are, but what
they are for us and in our way of looking at them that makes us happy or
unhappy."Be it said, however, that, as is generallythe case with the ex post
7ZAnticipationsa referto earlierideas, formulationsor findingsthat overlaplaterones but do

not focus upon and draw the same implications from them; "adumbrations,"to earlier
formulations or findings that, quite literally, foreshadowed later ones but only dimly and
vaguely. On these patternsin the history of thought,see Merton[1949,1957] 1968:13-25.
8Typicallystill for his time and place, Schopenhauerof course also echoed Epictetusin the
original Greekand the derivative Latin ratherthan in his own native German.


spottingof adumbrationsor anticipations,my havingnoticedthoseanticipatory
passages is probablyjust anotherinstanceof the "retroactiveeffect"(in which
undeveloped ideas that have remainedin oblivion are later brought into the
limelightonly becausenew and clearlyformulatedideas sensitizeus to earlier,
typically less developed, and previously ignored,versions).9The chances are
that I would not have takennote of those aphorisms(theorems?)by Epictetus
and Schopenhauer- as quotedeven in the variouseditions of Tristram
Shandy
on my bookshelves- had it not been for a half-centuryof close familiaritywith
the Thomastheorem.
Along with these venerableanticipations,the theoremalso had a striking
contemporaryand neighboringversion. In his lecturecourse at the University
of Chicago,W.I.Thomas'scolleagueGeorgeH. Meadhad observedin distinctly
sociological terms that "If a thing is not recognizedas true, then it does not
function as true in the community."But the Thomastheoremand the Mead
theoremexperiencednotablydifferentcognitivefates.In virtuallyself-exemplifying style, the Mead theoremdropped into permanentoblivion even afterits
posthumoustransitionfrom"oralpublication"in lecturesto publicationin print
(Mead1936:29).Not so with the Thomastheorem.
True,no notice of the theoremwas takenin any of the reviews of 7heChild
in Americathatappearedin the threeprincipalAmericansociologicaljournalsof
the time - the American
Journalof Sociology(whichwas then not only produced
and edited at the Universityof Chicagobut was also the officialjoumal of the
American Sociological Society), Social Forces(at the University of North
Carolina),and Sociologyand SocialResearch(at the University of Southern
California).But soon afterward,as we shall have occasion to see in detail,
Kimball Young, the prolific author of textbooks in social psychology and
sociology then at the Universityof Wisconsin,gave the theoremspecial notice
by selectingit as an epigraphfor chaptersin two successivetextbooks;one, in
his widely adopted SocialPsychology(Young 1930:397)and the other, in his
editedvolume,SocialAttitudes(Young1931:100).TheYoungepigraphsevidently
become an early conduit for diffusionof the quoted sentence.At any rate,the
very next year,the omnivoroussociologistHoward [P.]Becker,was interpolating the sentence twice in his amplifiedand Americanizededition of Leopold
von Wiese'sAllgemeine
Soziologie
([1924]1932:34,79)and faithfullyreportingthat
it was being quoted from Young'sSocialPsychology.
As is often the way with the genealogy of ideas, various types of errors
began to intrudewith enlargeddiffusion.Omnivorousreaderthough he was,
Beckermanaged to commit a triple errorwhen a dozen or so years later he
ostensibly located the Thomas sentence on page 79 of the condensed onevolume edition of the Thomasand Znanieckimasterwork,ThePolishPeasantin
EuropeandAmerica([1918]1927),where it is not to be found. Not contentwith
this mishap,Beckeradvancedthe extraordinaryconclusionthatthis nonexistent
ghost appearancewas "in contentat least . . . probablyZnaniecki's,"this even
9"Undevelopedideas" inasmuch as these earlier adumbrationsor anticipationswere not

singled out, elucidated, or followed up by furthertheoreticalor empiricalinquiry, either by
their original authors or by others.

384 / Social Forces 74:2, December 1995

though the concept of "definition of the situation' had appeared before The
Polish Peasantand then endured continuously as basic in Thomas's conceptual
framework but did not turn up at all, Polish colleagues assure me, in Znaniecki's early work and surely not in his later work in English.10
The Barnard sociologist Willard Waller (1938:20) introduced an error of
another kind in the course of diffusion. He misquoted the Thomas dictum,
understandably without citing its specific source, thus: "As Thomas has put it,
'If peopledefine things as real, they are real in their consequences."' In our own
time of acute sensitivity to deliberately or unwittingly gendered terms, one is at
first tempted to interpret Waller's substitution of peoplefor men as a quiet but
deliberate effort to de-gender the theorem. However, as one notices the further
substitution of things for situations,it seems more likely that the de-gendering
was the inadvertent result of a faulty memory.1
In the same year as Waller's would-be quotation, an effort was made to
pinpoint the lucid and elegant character of the formulation by describing it as
"W.I. Thomas's sociological theorem" (Merton 1938:333). Like Young's quotations of the theorem, this almost casual allusion made no precise reference to
the original source of what soon became abbreviated as "the Thomas theorem."
The long-standing and largely continuing absence of such specific citations and
texts persuade me that the widely neglected paragraphs which culminated in
the Thomas theorem both require and merit repetition here. We note that the
theorem caps the methodological case being made for use of
The behavior document (case study, life-record,psychoanalyticconfession) [which]
representsa continuityof experiencein life situations...
[E]venthe highly subjectiverecordhas a value for behaviorstudy. A document
preparedby one compensatingfor a feeling of inferiorityor elaboratinga delusionof
persecutionis as far as possible from objectivereality,but the subject'sview of the
For
situation,how he regardsit, may be the most importantelementfor interpretation.
his immediatebehavioris closelyrelatedto his definition
of thesituation,which may be in
terms of objectiverealityor in terms of a subjectiveappreciation-'as if' it were so.
Veryoften it is the wide discrepancybetweenthe situationas it seems to othersand the
situationas it seems to the individualthatbringsaboutthe overtbehaviordifficulty.To
takean extremeexample,the wardenof Dannemoraprisonrecentlyrefusedto honorthe
orderof the courtto send an inmateoutsidethe prisonwalls for some specificpurpose.
He excusedhimselfon the groundthatthe manwas too dangerous.He had killedseveral
personswho had the unfortunatehabitof talkingto themselveson the street.Fromthe
movementof their lips he imagined that they were calling him vile names, and he
behavedas if thiswere true.If mendefinesituationsas real,theyarerealin theirconsequences
(Thomas& Thomas1928:572;italicsadded).
As we see, this essentially methodological observation draws upon the basic
substantive concept of "defining the situation" which Thomas (1923:4243;
10This statementwill surely not be taken to detractfrom Znaniecki'smajorcontributionsto

sociology which, however, were of quite otherkinds than that representedby the theorem.My
own appreciationof those contributionsis summarizedin Merton1983.
11Onthe phenomenonof unintended genderingand de-genderingof language, see Merton,

"De-Gendering'Man of Science,"'1996.

The Thomas Theorem and the Matthew Effect / 385

1929:1-13) plainly regarded as his most significant contribution to the sociopsychological understanding of the formation of social personality and
character. Seen in its immediate context, the memorable theorem turns out to be
a generalization of a specimen of paranoid behavior. (As we shall also see in
due course, Thomas largely confined himself to reiterating this example the only
two other times he himself quoted the theorem, a singular circumstance that
raises an obvious problem for future exploration).
The longtime absence of a correct reference to the source of the theorem in
sociological writings quoting it also led me to conclude some time ago that it
had become known to American sociologists and their students largely if not
entirely through secondary discussions in print rather than through their having
read the original text. This assumption was reinforced by inquiries over the
years from colleagues near and far which, we shall see, asked for the exact
source of what had been described and analyzed as "the Thomas theorem"'in
"The Self-Fulfilling Prophecy" (Merton 1948). How such secondary diffusion
came to be and, in particular, how social mechanisms of initial diffusion
operated through such pathways as sociocognitive networks will be examined
in the last part of this article as we explain sociologically the early absence of
precise citations to the theorem and how that in turn helped produce the partial
citation phenomenon.
That once conjectural assumption that the theorem had largely become
known through secondary sources has now been empirically confirmed in a
study by R.S. Smith (1993) of "well over 100 introductory textbooks" of
sociology which found only one of the 40 texts that quoted or paraphrased the
theorem citing its source, replete with page number, while an unspecified
number of authors actually cited "The Self-Fulfilling Prophecy" as a mediating
source. Smith also found that most of the textbooks attributed the theorem to
W.I. Thomas alone and the few that referred to the book in which the theorem
appeared generally failed to cite Dorothy Swaine Thomas as its second author.
These empirical findings led Smith to a type of judgment which we have noted
is deeply imbedded in the normative framework of science about equity in
scholarly attributions:
[S]inceit [thetheorem]appearsin a co-authoredwork,andno particularauthoris singled
out as having writtenChapterXIII[in which the theoremappears],it seems reasonable
to suppose either authorcould have written this phrase.Consequently,unlessthereis
evidenceto the contrary,it would seem properscholarlypracticeto attribute
compelling
these words to both W.I. Thomasand DorothySwaine Thomas (Smith1993;I have
inserted italics to underscorethe scholarly care with which Smith allows for the
possibilityof countervailingevidence).
Having reported his empirical findings and having arrived at his contingent
normative judgment, Smith goes on to propose possible explanations of this
widespreadpracticeof excludingDorothySwaineThomasin referencingthe theorem.
This mightbe explainedby poor scholarshipon an individuallevel, althoughseveralof
those involved are nationallyand even internationallyknown authors.It can also be
explainedin termsof a structuralissue - the genderizationof the disciplineas part of
the processof professionalization.
By not citing DorothySwaineThomasthese authors
help sustaina view of sociologyas historicallya male domain(Smith1993).


However, in mounting this crisp study of "Dorothy Swaine Thomas and
the 'Thomas Theorem"' Professor Smith could not possibly have known that
both his normative observation about the proper citation of coauthored work
and his structural hypothesis of what might have led to the almost exclusive
ascription of the eponymous theorem to the male Thomas at the expense of the
female Thomas had been independently and concisely stated, without benefit of
systematic empirical study, in a letter written some five years before by a social
scientist of amply merited international fame.
The Inputation of "Institutionalized Sexism"
That letter, addressed not to me but to David Sills, with whom I was then
collaborating on the volume SocialScienceQuotations(Sills & Merton 1991), holds
vaied historical, methodological, and sociological interest as it briefly describes
a search for "the original source of the Thomas theorem." It bears not only upon
the early history of this important sociological idea but also upon the methodological requirement of putting to empirical test ex post hypotheses evoked by
data that are, of course, initially congruent with them. Moreover, as I have
hinted, the plainly informal letter and my lengthy thematic responses to it
exemplify various patterns in the transmission of knciowledgethat I had been
exploring for some time. Not least in point, the letter culminates in the
composite normative and explanatory conclusion that ascribing the theorem to
W.I. Thomas alone amounts to sexist eponymy.
Here, then, is the short evocative letter in its verbatim entirety.
August 16, 1988
David Sills
SocialScienceQuotations
111 EighthAve., Suite1503
New York,NY 10011
DearDavid,
I shouldhave the X. .. and Y ... citationsto you shortly.In the meantime,I bring
to yourattentionan interestingfootnotein the sociologyof sociology.Abouteightmonths
ago, I was desperatelylookingfor the originalsourceof the Thomastheorem.I couldfind
it nowhere,exceptin Mertonwho used it withouta footnote.I finallywas aboutto resign
myself to using as my citation:Cited in Merton,etc. Then your little brochure,"Social
ScienceQuotations:Guidelines"arrived,and lo and behold, on p. 15, as one of your
examples,you use the Thomastheorem,with a propercitation.
Welltrainedas I am in scholarlyskepticismof sources,I took out the bookfromthe
library.And lo and behold,you arein error,an errorwhichnot only you but Mertonand
indeedthe entireU.S.(world?)scholarlycommunityhavemadeconsistently.Forthe book
was not writtenby W.I.Thomas,but by WilliamI. andDorothySwaineThomas.Nothing
in the bookindicatesthathe wrotesome chaptersand she others.Theyarejointauthors.
And, at least in SocialScienceQuotations,this piece of (dareI call it?) "institutionalized
sexism"should not be perpetuated.
Yoursin the new era,
I anonymizethe authorof this informalnote as "SkepticalSocialScientist"
(hereafterSSS)]


Patterns in the Transmissionof ScientificKnowledge
Although this letterwas addressedto David Sills, I undertookto respondto it
since I was the one who had long ago electedto attributethe theoremexclusively to W.I. Thomas and since, as coeditor of SocialScienceQuotations,I had
preparedthe entryon the theoremas a specimenquotationin the "Guidelines"
for contributorsto thatvolume. I soon foundmyself resonatingto the considerable arrayof cognitiveand normativeideas and problemspackedinto the brief
but complexletterthatexemplifieddiversepatternsin the normativeframework
of science and the transmissionof scientificknowledge.Herewith,then, a few
of those patterns.
DIFFUSION
PATIERNSOFKNOWLEDGE
To begin with, it will be noticed that SSS matter-of-factlyadopts the eponym

"Thomastheorem"which had been introduceda half-centurybeforeto signal
the assessed importanceof the reverberatingThomassentence.It is unlikely,
however, that SSS had come upon that eponym where it first appearedsince
that was in an article published long ago (Merton1938) in the Philosophy
of
Science,a journalpresumablynot often read by social scientistsof a much later
time. Nor is it likely that this same social scientist writing in the late 1980s
would have come upon the eponym in 1948 when the theorem was being
analyzed and distinguished from the related concept of "The Self-Fulfilling
Prophecy"in TheAntiochReview,another journal not notable for its social
sciencereadership.I am inclinedto believe,therefore,thatlike those otherswho
had let me know that this was the case for them, SSS had come upon the
eponym in one or another of the three editions of SocialTheoryand Social
Structure(Merton1949, 1957, 1968)which had included both those articles.In
accord with the R.S. Smith (1993, 1995) finding that this book had become
something of a conduit for the Thomas sentence, here is SSS "desperately
looking for the original source of the Thomastheorem"and about to adopt
what some mightdescribeas the honorablebut,for many,the also unappetizing
expedientof quoting from a mediatingsource;to wit: "Citedin Merton,etc."
(justas we have seen HowardP. Beckerdo some 60 yearsbeforein quotingthe
theoremvia KimballYoung). As indicated by SSS and others, this citational
expedienthad resultedfrom the continuedabsenceof any precisereferenceto
the original source in all three editions of SocialTheoryand SocialStructure.12
SSSinformsus, Guidelinesfor
Contributors
toSocialScienceQuotations
Fortumately,
came along in the nick of time to provide "a propercitation.UU
12As we shall see from archival evidence yet to be examined, both adventitious and
theoreticalreasons made for this studied failure to supply the originallywell-known reference
to the exact source of the Thomastheoremin TheChildin America.
13his, then, may have been the first anticipatorycase, even before publication, in which
SocialScienceQuotationsfulfilled a manifestfunction that was describedthis way: "Of obvious
use to readerscoming upon quotationsnew to them, exactreferencesmay also prove useful for
swiftly locating the more familiar quotations. By leading readers back to the sources, such
detailed referencescan help them place even extended quotationsin their larger contexts. In


Although this straightforwardscholarlyletterreportsfull familiaritywith
the substanceof the Thomastheoremand with the eponym,it makesclearthat
such knowledgedid not derivefromhavingreadthe originalformulation.Thus,
and
the letterprovidesa distinctcase of the basic differencebetweenthe primary
orserialdiffusionof ideas,a subjectof distinctinterestto those of us
thederivative
at work on patternsof diffusionin science and technology(Coleman,Katz &
Menzel 1966;Merton1989,1995;Rogers[1962]1995;Zuckerman1989).Whena
first-handquotationdoes not provide an exactcitationto its source,it cannot,
of course, make for independent recourse to the original source and later
derivative or serial diffusion can only reproducethe quotationas mediated.
This,in turn,raisesthe sociologicalas well as normativequestion,which is here
only introducedratherthanexplored:how does it happenthat,unlikeSSSwho
was ready to cite the mediatedsource,many of the authorswho were plainly
quoting the theorem derivativelyratherthan directly have not done so? In
short,is there a norm for citing mediatedreferencesand if not, why not?
Is There A Norm for Citing Mediated References?
SSS'sannouncedintentionof citinghis mediatingsourceof theThomastheorem
may be defined by some peers as supererogatory,for there are evidently no
well-establishednorms governingsuch citationbehavior.At least, not if one
may judge from the comparativerarityof citationsto encyclopediasand other
referenceworksin scholarlypublications.Ormorespecifically,if one mayjudge
fromthe frequencywith which even maturescientistsand scholarshavewritten
sociologists of science to ask whether or how to ascribe"propercredit to an
authorfor drawingone's attentionto a valuablereference.'
These normativequestionsabout mediatedreferenceshad not yet crystallized back in mid-centurywhen the sometime sociologist of science Norman
Kaplan (1965) introduced the generic problem of "the norms of citation
behavior."Nor have thesequestionsaboutmediated
references
yet been examined
by the founder of the ScienceCitationIndexand of citation analysis, Eugene
Garfield(1983,1995),in his longtimeand still ongoing examinationof how the
norms and practicesof citationare acquired.(Seealso Cronin1984.)However,
we now see how this detail in the collegialletterfrom SSStacitlydirectsus to
a generic, difficult,and importantsociologicalproblem:the dynamics of the
emergenceof a new social norm;(on the genericproblem,see Coleman1990).
Justas anotherdetail tacitlydirectsus to the institutionalizednorm of socially
organizedskepticism.
this way, a book of quotations can extend an open invitation to the further reading or
rereadingof the original texts, beyond the quotationsthemselves"(Sills & Merton1991:xvi).
14Forexample, the queiy quoted in the text which was addressed in the first instance to
EugeneGarfieldand derivativelyto me by the authorof "Thoughtson Eponyms,"Howard B.
Burchell,M.D. (1985).


SOCIALLYORGANZED SKEPTICIM
Having long ago identified this technicaland moral norm embedded in the
cultureand the social structureof science (Merton[1942]1973:267-78;33940),
I
can only applaudSSS'smanifestadherenceto it. The termand concept,socially
organizedskepticism,refers to institutionalizedarrangementsfor the critical
scrutiny of knowledge claims in science and learning that operate without
dependingon the happenstanceskepticalbent of this or that individual.The
processof socializationin the cultureof sciencejoins with such social arrangements as published and unpublished"peerreview' that serve as agencies of
social controlwhich see to it, among otherthings,that authorsgenerallyabide
by the norm of indicatingtheirpredecessorsand sources.Thatnormhas many
cognitiveand social functions;Garfield(1983,1995)lists 15 of them.Amongthe
manifest cognitive functions are those of enabling scientists to consult pnror
sourcesto see whetherthey have been correctlyutilized and whetherthey also
providepertinentinformationnot includedin the mediatingsource.A manifest
socialfunctionis to pay homageto pioneersand otherpredecessors,alongwith
its largelylatentand correlativefunctionof thushelpingto maintainthe reward
system of science which, like all institutionalreward systems, initiates or
reinforcesincentivesfor role performance.
Exemplifyingsocializationin this normativepracticeof organizedskepticism, this committedscholarreports"havingbeen well trained... in scholarly
skepticismof sources,"and goes on to use the secondarywork of SocialScience
as a means of getting to the originalsource of the Thomastheorem
Quotations
that can then be examinedat firsthand.This soon led SSSto the discoverythat
the book in which the theoremfirst appearedwas actuallywrittenby William
I. ThomasandDorothySwaineThomasand that,in turn,led (as we have seen
that, in effect, it later led R.S. Smith) to the conclusion that ascribing the
theoremto W.I.Thomasalone amountsto "institutionalizedsexism."
This suggests the hypothesisthat such a conclusionis especiallyapt to be
drawnin a time of socially and culturallyinducedsensitivityto all mannerof
discriminatoryand exploitative-isms.For,as the sociologyof sciencehas noted
from its earliestdays, "the questionof the relativeimportanceof intrinsicand
externalfactorsin the determinationof the foci of scientificinteresthas long
been debated"(Merton[1938]1970:199)but thereis no questionthat social and
culturalcontextsdo variouslyinfluenceproblem-choiceand hypothesis-choice
(Zuckerman1978, 1994). In that process, certain contextually influenced
hypotheses ratherthan others soon leap to mind as plausible.In this case, it
appears,the context-ladenhypothesis that the failure to ascribethe Thomas
theoremto both Thomasesmust be an expressionof sexism.But of coursethat
hypothesisraises the methodologicalquestionwhetherjoint authorshipof the
book in which the theorem first appeared is enough to conclude that the
theoremwas itselfa jointproduct.In self-exemplifyingstyle,the cognitivenorm
of socially organized skepticism thus requires us in turn to examine that
conclusionby probingotherscholarlysourcesandpersonalarchives,as we shall
be doing in due course.
Meanwhile,one can see from the SSSletterhow adherenceto the norm of
organized skepticism can lead to gratifying experiences that presumably


reinforcecontinuingadherenceto it. Not least, the special pleasurethat comes
from erasing errorsin receivedknowledge.For the detectionof long-accepted
erroralso has its social and personalfunctions.It not only contributesto the
commonwealth of sciencebut resultsin system-inducedheightenedesteemby
knowing colleagues.I too once found myself resonatingto the kind of Eureka
glow that evidently came to SSS upon discoveringthe unscholarlyerrorthat
"not only [Sills] and Merton [but] indeed the entire U.S. (world?)scholarly
communityhave made consistently."I still recall the scholarlythrill of first
tentativelyspotting and then actuallydemonstratinga centuries-longerrorin
received scholarship.As it happens, this episode also involved the correct
ascription of a memorable socioculturalaphorism (theorem?)to its actual
author(s). The aphorism "If I have seen further, it is by standing on the
shouldersof giants"- which had diffusedwidely in the literaryas well as the
scientificcommunity,especiallyafterit becameknown thatIsaacNewton once
made it his own - had long been ascribed to what I demonstratedwas a
phantomsource in a work by the ancientpoet Lucanratherthan to its actual,
author,the twelfth-centuryBernardof Chartres.Muchalong the samelines that
SSSfearedmight be the case with the ascriptionof the Thomastheoremto W.I.
Thomasalone in the then forthcomingSills and Mertonreferencebook, Social
ScienceQuotations,the ubiquitousreferencebook Bartlett'sFamiliarQuotations
had been perpetuatingthe errorof an ancientsourcefor the Newton aphorism
- this, in no fewer than seven editionsfor almost a century.In the lattercase,
it became a scholar's comfort to find that even in this postmodern age of
deconstruction,evidentiarytruthcan still prevail.ForI can happily reportthat
only thirty years or so after that ghastly error of a ghostly source had been
demonstratedin On theShoulders
of Giants(Merton[1965,1985]1993:246-60),
it
was finally correctedin the fine sixteentheditionof Bartlett's([1882]1992:281b)
by its new scholarly editor, Justin Kaplan. And, in accord with the norm
governing scholarly acknowledgment,the correctionwas made with due
referenceto that long drawn-outand digressivedocumentingwork of mine.
But if the editor of Bartlett'sFamiliarQuotations
saw the light and rectified
what had been laboriouslyshown to be a fossilized errorof faulty scholarship,
not so, it appears,with the editors of SocialScienceQuotations.
In unyielding
style, they decided to continueascribingthe Thomastheoremto W.I.Thomas
alone.This,despite the advancewarningby theircolleague-at-a-distance
thatto
do so would not only perpetuatean almostuniversalerrorof quiteunscholarly
attributionbut would amountto a "pieceof 'institutionalizedsexism."'As we
shall soon see in detail, that apparentlyintransigentdecision was based on
fortunateaccess to a personalarchive.15
15This,however, is manifestlythe place to anticipatethe detailed documentaryanalysis a bit
by replicatinga memorandumon this decision dated "26 December1989":
THE THOMASTHEOREM
CanonicalVersion
To:
David
From: Bob
It seems to me that we should give readers or browsers the fruits of our unique
documented knowledge that though the theorem appearedin the jointly written TheChildin


Those recentobservationsby SSS and R.S.Smith on seeming inequitiesin
the citationhistory of the Thomastheoremprovide yet anotherspecimenof a
recurrentpatternin the domainof scienceand scholarship,to wit the patternof
multiple independentobservations,discoveries,and inventions.
The Pattern of Multiple Independent Observations, Discoveries, and
Inventions
As is well known, a longstandingtheory of the developmentof science and
technology is rooted in the theoreticallystrategic fact of the multiple and
independentappearanceof essentiallythe samescientificobservation,discovery
or invention;what, for brevity'ssake,can be describedas a "multiple"(Merton
1961,1963).Thispattem of independentmultipleshas been foundto hold for all
mannerof cognitive contributions,rangingfrom the great through the intermediate and small to the trivial. An apparent example of the truly trivial
multipleis providedin the triplyindependentobservationsby SSS(1988),Smith
(1993),and myself (Merton1976, [1948]1982) that the Thomastheoremwas
being ascribedsolely to W.I.Thomas,typicallywithouteven collateralreference
to DorothyThomas.To situate the specific empiricalcase in theoreticalterms,
the widespread citation pattern of the Thomas theoremwas identified at the
outsetof this retrospectivearticleas an instanceof the partialcitationphenomenon - the delimitedaccreditingof a presumablyjoint contributionto a subset
of the collaborators.16
Whether this example of a multiple in the ongoing
trivialas it is surely empirically
historyof the Thomastheoremis as theoretically
trivialwill becomeevidentonly upon examiningthe apparentcounter-example
of the typically full rather than partial citation of the important Ogbum-[Dorothy]Thomas(1922)paperon independentinventions.(On"thetrivialand
the importantin sociology,"see Merton[1961]1973:59-62;1987:16-19].)
America,it was neverthelesswritten solely by W.I.I won't burdenyou with anothercopy of the
14-page gloss on that fact when we were accused of 'institutionalized sexism' in having
ascribedit wholly to W.I. in the Guidelines to Contributors.
I believe that the quotation should be ascribed to W.I. just as it is in the Guidelines.
Plainly, a more extended arrow [our idiomatic term for explanatory notes appended to
quotations]is needed to explain the lone ascription.It might read this way:
The Childin America1928:572- Although the 'Thomas theorem' appears in this book
written jointly by W.I. Thomas and Dorothy Swaine Thomas,it is ascribedto him alone since
DorothyThomasreported[insisted?]in a letterto one of the editorsthat she had done only the
statistical portions of the book and that "the concept of 'defining the situation' was strictly
W.I.'s."
Of course, this could be condensed - or extended!

[In the event, this explanatory note was neither condensed nor extended but substantially
reproducedin SocialScienceQuotations.]
16Aswe shall see when we turn to the problemof explainingratherthan merely identifying
instances of the partial citation phenomenon,they often derive from the Matthew effect.


The Evocative SSS Letterin Retrospectand the RKMLetterin Prospect
SSS'ssuccinctletterof 1988 along with the Smithoral publicationof 1993have
thus providedmuch gristfor the sociologicalmill as we focus on the normative
question of proper ascriptionof the Thomastheorem.WhatSSS describedas
"aninterestingfootnotein the sociology of sociology"reflectsvariousmanifest
and latent patterns of peer interactionthat obtain within the institutional
frameworkof science and scholarship.And as we shall now see, the extensive
letter in response examines the normative issuesraised by SSS, supplies previous-
ly unpublishedarchivalevidenceaboutthe respectiveroles of W.I.and Dorothy

S. Thomasin formulatingthe Thomastheorem,
and addressesthe theoreticaland
methodologicalproblems involved in trying to explain this specimen of the
partialcitationphenomenon.
It will come as no surprise,of course, to find that,
like the SSSdocument,this one also exhibitspatternsthathave been identified
in the sociology of science. And so, like the letter from SSS,herewithreproduced verbatim,it too will be subjectedto ongoing commentaryin discursive
footnotesand bracketedtext.
RKM TO SSS&A VERBATIMDOCUMENT ON THE THOMAS THEOREMAS AN INSTANCE OF
THE PARTIALCITATIONPHENOMENON
SOCIAL SCIENCEQUOTMTONS
Who Said what, When, and Where
miten
WIDL S=LES
&l ecRuenrn'chCowsc
Dwr K. )zEOK
bAuasbiaUERii.,itV
diltor3I oardoRowK.
WAN,LV
MzwrO,Ctir,a*
BIIZWUNK U"JdiMi
dSfigGiR
GgwitT AUocAvuancedShut1*
CuLOR
Prvceto,
VAMANG3OORMNeT
S0NM PWk
PUblC Lib'VY
td
nO
GADmnLt vrmm we4StWdwtAdv
h4Lced
f Chicago
..iftv"
SaU
M ISAX&oDrfOu RoDwr N. UBELL
AowJevay
iaorScA
cA
S(4
SJOIU
tnfovd.Caliornuia
Compomencs
BonaT UwAswcums Iwc
MhlRhLAwcue,SuitC5O3
wlg,NYekQU
Wephout (2Za) 64548=A
RO BIN&
10 September1988
Dear [SSS],
I want to add my thanksto David'sfor your willingnessto put togethera batchof
quotationsfrom X--- and Y-. Knowingthat I had been thinkingof a much fuller
annotationin SSQfor "The ThomasTheorem,"David has askedme to respondto your
"footnotein the sociologyof sociology"concerningthe properattributionof the theorem.
I'm glad to try my hand at that since it should help me move towarda properannotation.
To begin with, you are surelynot alone in having searchedin vain for the prime
sourceof the ThomasTheorem.As you say, I failed to give a specificreferencewhen I
first happenedto refer to it as a theoremback in the 1930sand 40s. As a result,I've
periodicallyreceived requestsfor the precise reference.I enclose such a fairly recent
inquiry [as Exhibit8], this one from CynthiaEpstein (who I know won't mind my
includinghers as a specimendocument).
Now this is just the sort of informationwhich David and I intend to have the
scholarly apparatusof SSQ make instantly accessible.As you note, even our little


brochure, GUIDELINESTO CONTRIBUTORS,does provide "a proper citation" to the
theorem. But then, having exercised a proper "scholarly skepticism of sources" by looking
up the cited book source for yourself - this being the kind of behavior conforming to the
norm of "organized skepticism" which I proposed as a major element in the social
institution of science and scholarship back in the early 1940s - you go on to report your
findings in these words:
Lo and behold, you [David Sills] are in error, an error which not only you but Merton
[too] and indeed the entire U.S. (world?) scholarly community have made consistently.
For the book was not written by W.I. Thomas, but by William I. and Dorothy Swaine
Thomas. Nothing in the book indicates that he wrote some chapters and she others. They
are joint authors. And, at least in the Social Science Quotations, this piece of (dare I call
it?) 'institutionalized sexism' should not be perpetuated.
To make it clear that your observation registers a fairly newfound [general] sensitivity to
the matter of sexism, you sign off as
"Yours in the new era,"
[It will be observed that in one respect, the otherwise clear-spoken Skeptical Social
Scientist is here rather puzzling. Having noted in the first paragraph of the letter that the
Guidelinesfor Social Science Quotations which Sills and I had sent along to potential
contributors has "a proper citation" to the source of the Thomas theorem, that plainly
assiduous scholar nevertheless goes on to declare that we, like "the entire U.S. (world?)
community" have consistently erred in ascribing the bookto W.I. Thomas alone. My
response to this puzzle took this form:]
Now, you are altogether right of course in observing that the book in which the
Thomas theorem first appears - THE CHILD IN AMERICA:BEHAVIORPROBLEMS
AND PROGRAMS(New York:Knopf, 1928) - was "not written by W.I. Thomas, but by
William I. and Dorothy Swaine Thomas." Indeed, you will find that our SSQ brochure
indicates as much in its list of sources of the specimen quotations on page 19. (It refers to
W.I. and Dorothy S. Thomas, following the format of the INTERNATIONALENCYCLOPEDIA OF THE SOCIALSCIENCES.But I allowed friendship to taint scholarly precision.
All his friends referred to him as "W.I.",never as "William I." and W.I. clearly preferred
that usage [as, we shall see, did his collaboratorDorothy]. But we should have cited him
as he appears on the title page, not in this misplaced friendly but unscholarly fashion. As
the editor of the IESSin which SOCIALSCIENCEQUOTATIONSwill appear as Volume
19, David may overrule me on this [as in the event, he did] since I note that the
biographical entry in Volume 16 lists the book as having been written by W.I. Thomas
and Dorothy S. Thomas. And I recall that the authoritative volume, edited by Ed Volkart
for the Social Science Research Council back in the early 1950s, referred on the title page
to the "Contributions of W.I. Thomas to Theory and Social Research," not to William I.
Perhaps too many old friends of W.I. have improperly subordinated scholarship to
friendship. But this is scarcely the matter central to our discussion. We are in thorough
agreement, then, that the book was written by the two Thomases.
And you are also entirely correct in reporting that "Nothing in the book indicates
that he wrote some chapters and she others. They are joint authors." There is nary a word
in the book stating who thought or wrote what.
All this leads you to conclude that our ascribing the theorem to W.I. alone, rather
than to both W.I. and Dorothy, is plainly a case of sexism which may become further
institutionalized by the medium of SOCIAL SCIENCEQUOTATIONS.And you remind
us that we have entered a new era where old-style sexism no longer goes.


Still, I am confidentthat as a scholargiven to the carefulsiftingof fact fromfable,
you are ready to reconsideryour initialinterpretationsin light of new evidence.In this
case, your doubleindictmentof David and me as guiltv7ofscholarlyerrorand of sexist
bias to boot. With that in mind, I shall submitseveral exhibitsin the hope that you
will want to withdrawboth accusations(not thatany of us arewholly freefromerroror
immunefrom the contagionof inadvertentor of institutionalizedbias).18For we both
surely agree with those wise words of the 12th-centuryWilliamof Malmesbury(much
admiredby my mentor,the dean of historiansof science,GeorgeSarton[1931,I:255]who
wrote of him [that]"He was the best chroniclerof his time;the firstone afterBede who
triednot only to chronicleevents but to explaintheircausalrelations"):"Throwout such
dubiousstuff and gird ourselvesfor a factualnarrative."(Alas,my ancientnotes fail to
note the work in which this quotation[fromMalmesbury]is to be found.)
Exhibit 1 ITheThomasTheoremand the MatthewEffect]
The first exhibitbearingon the case of the ThomasTheoremis drawn frommy piece,
5 January1968,vol. 159,55-63,as this was
"TheMatthewEffectin Science,"(SCIENCE,
OF
slightly amplifiedwhen reprintedin my collectionof papers,THE SOCIOLOGY
SCIENCE(Universityof ChicagoPress1973).By way of orientation,I shouldreportthat
consistsin the
what I describedas 'the Mattheweffect'(afterMatthew13:12and 25:29)19
accruingof greaterrecognitionby peersforparticularscientificor scholarlycontributions
to scholarsof greatreputeand the withholdingof suchrecognitionfrom[theircollaborating] scholarswho have not yet made their mark.Here it is being suggested that the
Mattheweffectmight have operatedin the very case which is of centralinterestto us at
the moment.Thus:
The problemof achievinga publicidentityin sciencemay be deepened
by the greatincreasein the numberof paperswithseveralauthorsin which
the role of novicecollaboratorsbecomesobscuredby the brilliancethat
surroundstheirillustriouscoauthors.Evenwhenthereareonly two collaborators, the same obscuranteffect may occur for the junior who exhibits
of status. The role ascribedto a doubly or trebly
several "inferiorities"
stigmatizedcoauthormay be diminishedalmostto the vanishingpoint so
that, even in cases of latersubstantialachievements,thereis little recognition of that role in the earlywork.Thus, to take a case close to home. W.
I. Thomashas often been describedas the sole authorof the scholarly
book ThteChild in America,althoughits title page unmistakablydeclares
thatit was writtenby bothWilliamI. ThomasandDorothySwaineThomas.
It may help interpretthisrecurrentmisperceptionto considerthe statusof
the collaboratorsat the time the book was publishedin 1928. W. 1.
17Asthe reader will soon notice, these documentaryexhibits turned out to be a good many
more than merely "several."
SISSShad no way of knowing that I would find this charge of potential "institutionalized
sexism"particularlydistressing.For it was backin the early days of the civil rights movement
-a decade or so before the Civil Rights Acts of 1957 and 1960 - that I had attempted to
discrimination"as distinct
identify and to analyze the social phenomenonof "institutionalized
from acts of discriminationby individuals(Merton1948a:120, 101). That had seemed to me a
fundamentalsociological distinctionthen as it seems to me still.
19By way of reminder. "Unto every one that hath shall be given, and he shall have
abundance:but from him that hath not shall be taken away even that which he hath."

The Tomas Theorem and the Matthew Effect / 395

Thomas, then 65, was presidentof the AmericanSociologicalSociety in
of his longstandingrankas dean of American
belatedacknowledgement
sociologists,while DorothySwaineThomas(not to becomehis wife until
seven yearslater) was subjectto the doublejeopardyof being a womanof
sociologicalscience and still in her twenties.Althoughshe went on to a
distinguishedscientificcareer(incidentaily,beingelectedto the presidency
of the AmericanSociologicalSocietyin 1952), the earlybook is still being
ascribedsolelyto her illustriouscollaboratoreven by ordinarilymeticulous
scholars."9
19. See the ascriptions of the book, for example, in Alfred Schutz. Collected
Pape"s, 2 vols., edited and with an introduction by Maurice Natanson (The Hague:
Martinus Nijhoff. 1962), 1:348, n71; Peter McHugh, Defining the Situation (Indianapolis: Bobbs-MerillCo, 1968), p. 7.
Exibit 2 [my I do not takethe AlfredSchutz ascriptionof the bookto W.L aloneas
ipsofacto evidenceof a sexist bias]
The precedingquotedfootnoteindicatesthateven so meticulousa scholaras Alfred
Schutz(1962:,348,n. 71)has managedto ascribethe salientbooksolelyto W.I.To obviate
any need for you to searchout the passagein which he does so and thanksto our home
photocopier,I Canonizeit here (as I should perhapshave done in the publishednote):
The "definition ot the
to
the
so-called
"Thomas
theorem" well known
situation" refers
as
situations
men
define
real, they are real in
"If
to sociologists:
their consequences."71
1 It was first developedby WVlliam
Isaac Thomas in his book, The Chil i"
Amcri= BehaviorProblemsand Programs,New York, 1928, p. 572. See also W. I.
Thomas, Social Behaviorand Pmonaoi*y, edited by E. K. Volkart, Social Science
Research Council, New York, xg5z, pp. 14 and 8off.; the term "ThomasTheorem"
was coined by RobertK. Merton,Soca Thry and Social Strucaure, Glencoe,1949,
s79.
Youwill noticeat once the emphaticthoughprobablyunintendedway in which,the

to-meadmirablescholar,AlfredSchutz,castsDorothyThomasintolimboby ascribingnot
merelythe ThomasTheorembut the book to W.I. alone. Indeed,the fatefulmasculine
pronounin the phrase "his book"would seem to lend grist for your mill of sexism at
workin the case of the Thomases.Nevertheless,as you see frommy referenceto Schutz's
errorin attributingthe book wholly to W.I.,I do not imputea sexist bias that made for
the error. The reason? Our now fairly extensive studies of the Matthew effect in
collaborationsamongscholarsof greatlydifferingstandingin the field have foundthis to
as well.
and amongfemalecollaborators
operatequiteregularlyamongmalecollaborators
(Whichis not to say, of course, that sex or gender does not affect the probabilityof
achievingrecognitionfor one's scientificor scholarlywork;the intensivestudiesof men
and women scientistsby HarrietZuckermanand JonathanColeoverthe yearsbeing the


basis for that statement.)20 In view of all this, I adopted the clinical hypothesis that
Schutz's ascription of the book to W.I. alone was probably an instance of the Matthew
effect rather than a sexist predilection.
ascribethe theoremto W.Laloneand absolvethe

I reiteratively
Exhibit3 [Wherein
of totalsutjectivisml
theorem
This exhibit has me [once again] emphatically ascribing the theorem, not the book, solely
to W.I., even though it appeared in a co-authored book. This I have done regularly since
its logical character and, I confess, its assessed importance first led me to describe it as
'The Thomas Theorem". (Tis was not, as Alfred Schutz understandably cites it in his
foregoing footnote, in 1949 but back in 1938. But then, there was not the least reason for
this colleague at-a-distance to know that the term had appeared earlier in my article,
"Science and the Social Order" [Merton 1938:331-332]).
You will note that the following passage, again photo-copied to save you confirming
search, connects the theorem to other formulations but singles it out as the most succinct
and memorable formulation of much the same sociological idea. But I must not digress
into questions of sociological theorizing. Here, then, is a pertinent fragment [drawn] from
a piece of mine titled "Social Knowledge and Public Policy,' first published in 1975 and
reprinted in Merton [1976]; the fragment appears on pages 174-175, 177:
The Hazardsof Subjectivism
The idea of the subjectivecomponentin humanaction has a long lhistory
In sociology and lhadan even longer history before wvesociologists arrived an the historicalscene. It is an idea, moreover,that has been formulated in various traditions of sociological thought: in the notion of
(roughly: intuitive understanding)advanced by Max Weber
VersieheI&
(and many others influencedby him), Robert MacIver's "dynamicassessment,"FlorianZnaniecki's"humanisticcoefficient,"Talcott Parsons's
"voluntaristictheory of action," and Schutz's 'phenomenological perspective."The idea was succinctlyformulatedby W. r. Thomasin wvhat
is probablythe single most consequentialsentenceever put in print by an
Americansociologist:
If mendefinesituationsas real,theyare real in theirconsequences.20
Now, it is one thing to maintain,with Weber,Thomas,and the other
giantsof sociology,that to understandhumanaction requiresus to attend
systematicallyto its subjectivecomponent:what people perceive, feel,
believe,and want. But it is quite anotherthing to exaggeratethis sound
idea by maintainingthat action i5 nothing but subiective. That extravaganceleads to sociologicalBerkeleyanism(the allusionbeing, of course,
to the Englishchampionof philosophicalidealism, not to an American
conceivesof social
geographicor academicplace). Suchtotalsubjectivismn
realityas consistingonly in social definitions,perceptions,labels, beliefs,
assumptions,or ideas, as expressed,for example,in full generalityby the
criminologicaltheorist,RichardQuinney,when he writes that "We have
no reason to believe in the objective existence of anything.""'A basic
idea is distortedinto error and a great injustice is visited upon W. I.
Thomasivheneverhis theoremis thus exaggerated.
Exaggerationof a seminal truth produces its own brand of error.
Total subjectivism,which maintainsthat only social definitionsof the
20Cole (1979),Cole and Zuckerman(1984), [Zuckerman,Cole and Bruer(1991),Zuckernan
and Cole (1994)].


situation (or other subjective equivalents) determine the characterof
humanaction and its consequences,in effect manages to transformthe
ThomasTheoreminto this fallaciousmaxim:
If men do not definesituationsas real, they are not real in their
consequences.
...To correct the imbalancethat comes with total subjectivismand to
restorethe objectivecomponentsof social situationsto their indispensable
place,we plainly need this counterpartto the ThomasTheorem:
And if men do not definereal situationsas real, they are nevertheless
realin theirconsequences.
You evidently consider this ascription of the theorem wholly to W.I. (with an allusion
to Dorothy Thomas only in a footnoted citation to the book)21 as primafade evidence of
"sexism." In this case, however, presumably an expression of personal rather than
"institutionalized sexism." But I urge you to withhold reiteration of this harsh impeachment until you have examined Exhibits yet to come.
Exhibits4 and 5 fIhich bearrenewedwitnessto an longtimeeffortto deterthe

solelyto W.I.]
ascribingof sexismto scholarswhoattributethetheorem
A few prefatory remarks before I introduce these almost identical exhibits.22 Nearly six
decades ago, I elected to focus on social structures, social patterns, and social processes
in the domain of science and scholarship - not a subject of immense scholarly interest
back then. And, as you surely have no cause to know, some thirty years ago I first tried
to identify the institutionally distinctive reward system of that domain since it seemed to
me to provide part of the dynamics and patterning of scientific work. That was set out in
some detail in a longish paper titled "Prioritiesin Scientific Discovery"(1957). Part IV of
THE SOCIOLOGYOF SCIENCE (1973), designated as "The Reward System of Science,"
collects a variety of my articles focussed on the workings of that system. Among other
things, that 1957 piece proposed the strongly stated hypothesis that contradictions
21
In retrospect,I note that SSSand, for that matter,R. S. Smithin his 1993 paper,along with
countless others before and after them might have obviated their continuing search for the
provenance of the theorem had they happened upon Volkart'sreprintingof the concluding
chapterback in 1951 (Thomas1951) or even this citationto TheChildin Americain 1975.But as
GeorgeJ. Stigler's (1961)seminal paper, "TheEconomicsof Information"led us to see both in
principle and in fact, even in our new age of advanced information-technology,achieving
retrieval of sought-for informationcan exact prohibitive costs in terms of time, energy, or
money.
22 This archival essay-letter,which was plainly not intended for publication,had adopted
ShandeanMethod of compositioneinaugurated
"the non-linear,advancing-by-doubling-back
by LaurenceSternein his immortaleighteenth-centuryTristramShandyand hesitantlyadapted,
of Giants(which,it
just two centurieslater, in my own "ShandeanPostscript,"On theShoulders
may be remembered,turned up in the introductorypages of this paper as I was empathically
resonatingto the manifest pleasure expressedby SSSupon discoveringwhat was taken to be
the universal errorof attributingthe Thomas theorem to W.I.).Steme's TristramShandyhad
drawn upon the technique of "stream of consciousness" long before William James had
formulatedthat apt metaphoricconceptin his monumentalPrinciplesofPsychology([1890]1950,
I: 239) and longer still, before James Joyce and, to a degree, Virginia Woolf had put that
techniqueto work in their novels.


between the rewardsystem and the normativesystem of sciencemade for such social
pathologiesas the occasionalfeloniesof plagiarismand the creationof fraudulentdata.
(I stray from the subjectin hand to mention this only becauseof the intense current
interest,both scholarlyand popular,in such pathologicalphenomenain science.Backin
the 1950s,that effortto identifythe varietiesof misbehaviorin scienceand to theorize
about their systemic sources apparentlyseemed like little more than sociological
muckraking.)
Backto our immediatesubject,I also reportthat PartV of the same volume,THE

SOCIOLOGY
OF SCIENCE,is given over to an examinationof 'The Processes of
Evaluationin Science.' As one would suppose, it includesthe articleon the Matthew
effect and other pieces devoted to theoreticallyderived and empiricallyinvestigated
pattems of [peer]recognitionin science.
I reportthese longstandinginterests,in supremelyegocentricstyle, for a reason.
These ?s are designed to providea contextfor the pair of exhibits,4 and 5. Fromthat
contextit can be inferredthat rve had morethan a friendlyand ad hoc interestin the
peer perceptionsof my old, much admired,friends,W.I.and DorothyThomas.And so,
as [the precedingexhibitand] those [following]exhibitstestify,I have gone into print,
fromtime to time, to get the stoxystraightaboutthe sourceof the ThomasTheorem
The essential step, it seemed to me in light of certainpattems that occasionallU
emergein the world of scholarship,was to providea kind of anticipatoryprophylaxds,
in print,directedagainsteasy misinterpretations
and harshindictmentsof scholarswho
attributedthe theoremto W.I.ratherthanto bothW.I.and Dorothy.Perhapsgettingthe
facts in print would precludesuch indictmentsbased on skimpyprimafacie evidence.
Thus:
Exhibit4 [Evidence
thatmy ownattfribution
of theThomasTheorem
to WI. alonewas
neither'sexist'nora caseof theMattheweffect]
20. Whatwe may call the ThomasTheoremappearsjust once in the corpusof
W.I. Thomas'swritings:on page572 of the bookhe wrotewith DorothySwaine
(Thomas)ThomasentitledThe Childin America(New York:Knopf,1928). I
ascribethe theoremto W. I. Thomasaloneratherthanto the Thomasesjointly
not becauseof his genderor great senioritybut only becauseDorothyThomas
has confirmedfor me what many have supposed:that the sentenceand the
paragraph
in whichit is encasedwerewrittenby him.Thereis thus nothingin
thisattribution
thatsmacksof "theMatthewEffect,"whichin casesof collaboration
betweenscholarsof decidedlyunequalreputationhas us ascribeall creditto the
prominent
scholarandlittleor noneto the othercollaborator(s).On the Matthew
effect,seeMerton,op. cit., 1973,Chapter20.
(This wouldbe prophylacticfootnote is quoted from the second 'edition' of the
previouslycited paper,'Social Knowledgeand SocialPolicy,' as it appearson page 175
of Merton,[1975]1976.)
As you see, this note anticipatesmisinterpretations
of my having attributedthe
theoremto W.I.alone. Having been immersedall those many years in the sociologyof
science,I had inevitablybecomeawareof the frequentpatternof swift mis-imputations
23 Earlyscholarlystudies of the reward system of science include Gaston (1971),Hagstrom

(1974), and Zuckerman (1977); for an analytical overview, see Zuckernan (1988:520-26).
Drawing extensively upon sociological sources,the popular volume by the science joumalists,
Broad and Wade (1983), soon catalyzed the attention paid by the mass media to deviant
behaviorin science.
24
On re-reading,I find the term 'anticipatory' ratherredundlant.

The ThomasTheoremand tiheMatthewEffect/ 399

of ideologicalor otherbias. Hence,my evidentlyunavailingeffortto counterin advance
the chargethat my attributionsimplyexpressed(wittingor unwitting)"sexism"or, in a
parallelconstruction,inverted 'ageism." Or that the attributionwas simply another
instanceof the Mattheweffect.25
I should pause to reportthat I erredin writingthat 'the ThomasTheoremappears
justonce in the corpusof W.I.Thomas'swritings."Theleadingauthorityon that corpus,
EdmundH. Volkart,wrote me recentlyto say thathe had independentlymade the same
erroneousclaim (A comparativelyrarespecimenof independentsimultaneouserrorin
contrastto the many known cases of independentsimultaneousdiargx=.) Ed Volkart
had learnedfrom our longtimemutualfriend,EleanorIsbell- for so many years,the
indispensablestalwartof the SocialScienceResearchCouncil- that W.I. did put the
theoreminto printonce again.This,in his essay, "TheRelationof Researchto the Social
Process,"which appearedin a symposiumby TheBrookingsInstitutionin 1931underthe
title, ESSAYSON RESEARCH
IN THESOCIALSCIENCES3.
Uponexaminingthatessayby W.I.,I see thathe largelyreiteratesthe original?lfrom
THECHILDIN AMERICAin which the theoremappears,clearlymakingno effortto
elucidateor develop it further.Still,the fact, and it is truly a fact whateverthe radical
subjectivistsmight say, thatthe crucialsentence[re-]appearsin an essay by W.I.[alone]
would seem to provide publiclyaccessibleevidencethat he -knewit to be his and his
alone.
Speakingof reiterationsin print,I would have you turnthe page to Exhibit5. This,
as you will observeat once,is anotherwouldbeprophylacticnote thatlargelyrepeatsthe
one publisheda half-dozenyears before.It appearsin a revised introductionto 'The
Self-FulfillingProphecy,"as it appearson page 248 of Merton[1948]1982.
Exhibit5 RWichreportsbutdoesnot documentthefact thatDorothyThomasherself

hadconfirmed
thatshehadnothingto do withtheconceptof "definingthesituation")
**Whatwe have been describingas the ThomasTheorem appearson
page 572 of the book he wrote with DorothySwaineThomas in 1928: The
Child in America(New York: Knopfl. I ascribethe theorem to W.I. Thomas
alone rather than to the Thomasesjointly because Dorothv, who became
Dorothy Thomas Thomas when they were marriedeight years after that
book appeared, confirmed that the consequentialsentence and the paragraph in which it was encased were written bv him. Thus, nothing in this
attributionsmacksof "the MatthewEffect,"which operatesin cases of collaborationbetween scholarsof decidedlv unequal reputationto ascribeall
credit to the eminent scholarand little or none to the collaborator4s)-supplementarvnote, 1982.
25 In retrospect,I note here the pronouncedbut still undeveloped effort to avoid conflating
the 'partial ascriptionphenomenon'and possible hypotheticalexplanations
of the phenomenon
in terms of sexism, ageism, or, more generally, the Matthew effect. I trust that this will be
clarifiedin the final section of this paper which examinesthe sociologicaland methodological
import of both the phenomenonand its explanation
26
As we shall see, Thomasemphaticallyquoted the theorema third and last time in Blumer
(1939:85)but did not elect to apply or develop it further.That introducesa puzzle for future
exploration.What are we to make of W.L'sown sparse attentionto the specific theoremin his
ample writings on the governing concept of "definitionof the situation"?The very question
evokes SherlockHolmes's observationon the methodologicalsignificanceof the absent:"the
curious incident of the dog in the rnght-time;the dog did not bark. (For the imperishable
original, see Arthur Conan Doyle, [1894] 1953: 347; in perhaps more immediately relevant
context, the methodologicalfragmentreproducedin SocialScience
Quotations
[Sils & Merton
1991:52]).


The reasonfor the largelyreiteratedfootnoteis by now surelyclearto you as it had
becometo me. It was plainthata then-and-thereexplanationwas needed wheneverone
attributedthe theoremto W.I. alone. Else,some were bound to concludethat Dorothy
was being robbedof this partof her ampleintellectualcontributions.Thisvariantof the
note no longeralludesto possibleimputationsof genderor age bias.Instead,havingset
forth the groundsfor the attribution,it goes on to inform,or remind,readersthat the
had married- on 7 February,1935when he was 71 and she, 34 -and that
collaborators
Dorothyhad been a Thomasbeforeshe marriedanotherThomas.
Thatallusionto the post-marital"DorothyThomasThomas"must seem altogether
irrelevantto the subjectin hand.27 I agree.It is thereonly as an echo of Dorothy'sletter
to me in which she describesher role in the writingof "TheChildren[sic]in America"
and reportsthat "Theconceptof 'definingthe situation'was strictlyW.I.'s."Butreaders
of thoseprophylacticfootnotescouldnot surmiseall that.WouldthatI had been able to
publish in full Dorothy'scharmingletter to me and my reply to her. Then the story
would have been writ largeand plain.
But if I could not intrudethat correspondencein my latter-dayarticlesreferringto
the ThomasTheorem,I can surelyincludeit here.28
You have only to turn to the next page of this lengtheningresponseto your own
letterto Davidto find Dorothy'sletteras Exhibit6 and thenturnto the page[s]following
for Exhibit7.
27R.S.Smith (1993:4-5) found that threeof the 40 textbooksreferringto the Thomastheorem

"imply she [DST]was marriedto W.I.Thomas"at the time it appearedin print.He notes that
this was mistaken and goes on to observe: ". . . even if they were marxied,this begs the
question of why include such information? According to The ChicagoManual of Style (1982), it
is not scholarly practice to identify the marital status of coauthors.It is assumed that each
author is a contributorin his or her own right and so making such a point is irrelevantor
detractsfrom the purpose of citation.As Reinharzimplies in the title of her articleon Dorothy
Swaine Thomas ("Wasn'tShe the WomanMarriedto WilliamI.?")citing the work in this way
reinforcesthe patriarchalpractice of subsuming a wife's work under her husband's authorship." Thus led to the Reinharzarticlewith its nicely ironic subtitle, one arnivesat it only to
find next to nothing there about eitherthe personallife or the considerableworklfe of DST.As
this archivalletter to the SkepticalSocial Scientistmakes plain in virtually anticipatorystyle,
my own echoic allusion to their marriageno more exemplifies "the patriarchalpractice"of
subsuming"underher husband'sauthorship"Dorothy'svirtual lifetime of pioneeringwork in
demography(D.S.Thomas1938,1941;Kuznets& Thomas1957-64)and her pioneeringstudies
of the Japanesedetention camps during World War II (D.S.Thomas & Nishimoto 1946;D.S.
Thomas, Kikuchi & Sakoda 1952) than it exemplifies a tacitly "matriarchal"practice of
subsuming under his wife's authorshipW.I.'svirtual lifetime of work on his concept of "the
definition of the situation."It is only that even meticulous scholars of Thomasiana,such as
Janowitz (in Thomas1966:xvii)and Tate (1974),unaware that DSThad been "an ardentLucy
Stoner" - an American colloquialism for a married woman who insists on retaining her
birthname- have assumed that Dorothy Swaine became DorothySwaine Thomasonly after
marriage.
28Though persuaded then as now that discursive footnotes providing correlatives and
contexts serve a useful function, I hadn't the temerityto include Dorothy'sletter as a rather
extended footnote in papers that, unlike this one, were not focussed on the matter of
attributionsof the theoremto W.I.being takenas a (wittingor unwitting)expressionof sexism.

Ihe ThomasTheoremand the MatthewEffect/ 401

Exhibit6 [In whichDorothySwaineThomassuppliesthearchivalsmokinggun)
GEORGETOWN
UNIVERSITY
D.C. 20007
WASHINGTON.
CENTSR FOR
POPULATION RZSeARC14
September 10, 1973
Professor Robert K. Merton

Department of Sociology
Columbia University
New York CIty, New York 10027
Dear Bob,
Many thanks for your nice letter
on the Matthew Effect in Science.
and the copy of your paper
I have always enjoyed reading
the things both you and Harriet Zuckermanhave written.

just one point in this article
underlinings
that puzzles me.
refer to some index.
There is
I assume the
If so, you apparently have me
under the Swaintg> I assure you I was born a Thomasand then

married a Thomaswho was no relation.
I was an ardent Lucy Stoner
and also swore I would never change my maiden name which I didn't.
In regard to The Children in America W. 1. Thomasemployed me
as an assistant
since he had been told by the Rockefeller
get himself a statistian.
The statistical
group to
portions were mine and
I am sending you under seperate cover Volkart's book which makes

this clear.
W. I.'s.
The concept of "defining the situation"
was strictly
With cordial regards,

Sincerely,
Dorothy SwainmeThomas
Professorial Lecturer
DST:rjm


Exhibit7 [Replyto thesmoking-gun
letterby DST]
CENTER
FOR
ADVANCED
STUD-Y IN THE BEHAVIORAL
202 JunipersSeao Soulevowd* Stloaord,Colfornlo94305
SCIENCES
Telephone(1S1 321.2052
28 September 1973
Dear Dorothy,
I treasure your sentence
possible occasion as a lovely
and, unless you say no, I shall

piece of sociological
history.
quote it on every
The sentence:
"I was an ardent Lucy Stoner and also swore I would never
change my maiden name which I didn't."
As for the underlining that puzzled you, that is pure and, in this case,
meaningless chance.
I happened to xerox the version of the paper which appears
in the page proofs of my forthcoming collection
of papers, THE SOCIOLOGY
OF SCIENCE.
All the proper names in the book had been underlined for purposes of indexing (and
I assure you that Dorothy Swaine Thomas appears in the Index as such and not as a
Swaine).
During the almost forty years (') since I first met W. I. and you, I have
retained a happy image of the two of you together, in your every joint aspect.
Evidently,
that was more than a casual imprint.
As for the possible Matthew effect involving W. I. and you, I had only this
in mind. True,,you were very much the junibr assistant
but W. I., in his generous
way, saw to it that you were identified
as co-author on the title
page, It is
therefore of some interest
that repeatedly THE CHILDIN AMRICA is cited as being
by W. I. alone.
Of course, that has nothing at all to do with W. I.'s concept of
"defining the situation."
That was a basic sociological
idea for many when I was
a freshman first becoming excited by this oddly-shaped field known as sociology.
Yours ever,
Robert K. Merton
RKM:ja
Dorothy Swaine Thomas
Professorial
Lecturer
Center for Population Research
Georgetown University
Washington, D.C.
20007


[Butbeforeconcludingthis long responseto SSS,still anotherinterventionis needed
by way of context:Forlengthythoughit is, this now archivaldocumentneverthelessdid
not include an essential part of Dorothy'sand my correspondence;namely, my note
signalingthe pertinentpassagein the "MatthewEffect"paperwhichled to heraltogether
unexpectedsingling out of W.I.'sand her own distinctiveroles in writing 7heChildin
America.Here, then, is that note as the only-nowinterpolated
Exhibit 7A [which,alongwith Exhibit1, had evokedDorothy'ssmoking-gunreplythat
was reproducedas Exhibit6 in the 1988 letter to SSSJ
31 August 1973
ProfessorDorothyS. Thomas
Centerfor PopulationResearch
GeorgetownUniversity
Washington,D.C 20007
DearDorothy:
I had hoped to see you at the meetingsand that is why I hadn'twrittenyou before
now. And how wise you were not to come:980 in the shade all throughthe sessions.
I am flatteredto have you ask for a photographof me but I don't have one at hand
of me back in the days when I was president.But I'll see if I can dig one up and will
happilysend it on to you.
In the meantime,I thoughtyou mighthavesomeinterestin the enclosedpaperif only
becauseof whatI say aboutyou, W.I.,andTheChildin Amenca.on pp. 446-7. [Italicshave
been added to underscorethe relevantsentencethat evidently evoked the unsolicited
smoking-gunreply].
Yours,
It is of primeevidentiaryimportancethatneitherthis note nor iheindicateapassage
in the accompanyingcopy of "TheMatthewEffect"(whichappearsas Exhibit1 in the
letter to SSS) refers to the partialcitationof the Thomastheorem.As can be seen, the
passagerefersonly to the partialcitationof thebook.It then proceedsto accountfor this
instanceof the partialcitationphenomenonin termsof the Mattheweffect,which is said
to be all the more probablesince the less-knowncollaboratoris "subjectto the double
jeopardyof being a woman of sociologicalscience and still in her twenties."9 Yet, it
will be noticedthatonherozwninitiative,DorothyThomastakesoccasionto partitiontheir
distinctivecontributionsto the book to emphasizethat as W.I.'sstatisticalassistant"The
statisticalportionswere mineand ... [t]heconceptof 'definingthe situation'was strictly
W.I.'s."The concluding portion of this retrospectivearticle will collate the many
statements,bothpublicand private,to the sameeffectby bothDorothyand W.I.Butfirst
the conclusionof the long letterof responseto SSSback in 1988.]
And thereyou have the essentialdocumentation.Theseseven Exhibitscould surely
be extended in my old and new Thomasianfiles.*But perhapsthis documentationis
will not be perpetuatingany
enough to indicatethat SOCIALSCIENCEQUOTATIONS
"institutionalized
sexism"by attributingthe ThomasTheoremto W.I.Thomas.
Yours,truly,froman era not quite so new,
rt K. Merton
29 Readersmay want to refreshtheir memory by glancing back at Exhibit1.


* ... I append a few other Exhibitswhich might have been incorporatedin the letter
properwere this being put togetheron a word-processorratherthan an old-fashioned
SelectricII typewriter.
Exhibit8:
Exhibit9:
Cynthia[FuchsEpstein's]note of 21 May, 1981

R.K.M.replyto C.F.E.,26 May1981.[C.F.E.managedto put the gist of
this exchangeinto herbookthen in press:Womenin Law(BasicBooks,
1981),362n., thus providinganotherstill-rareprecisecitationto the
originalsourceof the theorem.
Titlepage and antecedentpage of my copy of the firsteditionof THE
CHILDIN AMERICAbought at the grand Leary'sbookstorein
Philadelphiaon the 23d of June1939
Exhibit10:
Exhibit 8
Columbia Universityin the Cihyof New York

0 064
pi
0.
C1.CLNILW
E
9CI'~t"lIer
t. IAL.
~T.C)fI&L
C?66
';c?1
oft 0E
.~CaI
I X
.*:-*
.9.
'
...
.@
*.*.
21 May, 1981
Dear Bob:
Checking through hundreds of copy-editor's
quieries on my book I had occasion to check
ST &SSfor the ThomasTheorem. To my surprise
I find no footnote to the original "If men define
things as real..."
I am convinced that if you chose to present
it without footnote then you were applying
some rule of commonknowledgethat makes
the footnote unnecessary- indeed, perhaps
vul gar.
So - I ask,
even implore
for a spelling
out
of the rule (as I lay choked in dusty tomes

pursuing thankless searches for forgotten
footnotes).
You427


Exhibit 9
Columbia
Universiltyin the City of NewYork j
t-NIVERSI
ty
New York.N.Y. 10027
PROFESSO1;
Fayes%waltter
'Wil*
26 May 1981
Dear Cynthia:
the Thomas
Having a specific
page reference*for
Theorem will in truth make your book quite distinctive.
As you may have noticed,
few cases
there are precious
in fact,
to the famous
I don't recall
many -.referring
its source.
sentence
which do accurately
pinpoint
This gives
the impression
that ST&SS may indeed have
served as a conduit
over the past thirty
years for
the Thomas Theorem.
I enclose
At any rate,
the page
reference.
I should explain
it to W.I. Thomas,
that I attribute
rather
than to both Dorothy Thomas and him because
I
She was emphatic that
once asked Dorothy about it.
were entirely
both idea and formulation
W.I.'s.
Incidentally,
she once told me also that part of the
reason -- perhaps,
only a very small part -- for her
marrying W.I. is that 1t allowed
her (as a Lucy
Stoner)
to retain
both her maiden name and to take
on her husband's
name as well.
I write this just hours
off on our much-awaited
before
holiday
and I take
Harriet
and London.
in Italy
With much love,

Yours,
Robert
Professor
Cynthia
Center for Social
817 S.I.A.
* Encl.
xerox
K. Merton
Epstein
Sciences
of p. 175 in rkm, SOCIOLOGICALAMBIVALENCE
Enc.


Exhibit 10
v4':
* hL4e4iztcA.:
THE
IN AMERICA
CHILD
L
BEHI-A'IOR
AND
PROBLEMS
WILLIAM
PROGRAMS
I. THOMAS
AND
_______
gi
DOROTIHY
NqLW
Y'ORK
SWAINE
/ILFREDl)d
TIIOMAS
KNOPF).
MicMxxviii
Socia Mechanisms Generatingthe PartialCitation Phenomenon

As indicated at the outset, this article is not primarily concerned with adjudicating claims to the origin of the Thomas theorem, and surely not with doing so on
behalf of W.I.Thomas or Dorothy Swaine Thomas. Both simply took that origin
for granted. We shall be examining further crucial evidence, from both a private
archive and from publications, bearing on the recently disputed origin of the
theorem, but our primary objective remains to understand the partial citation
phenomenon and its place in the reward-system of science by analyzing this
particular instance of that phenomenon as a strategic research site. The case of
the Thomas theorem provides a strategic site if only because I happen to have
first-hand and fine-grained archival information about the origin of the theorem,
its early citations and subsequent citational history, the latter having been


usefully amplified by the Smith (1993)survey of the theoremin textbooks.3"
And thoughsome of thatarchivalmaterialappearsin the foregoingdocumentaxy letter to SSS, it leaves a variety of specific historicalpuzzles and generic
sociologicalproblemsstill to be specifiedand solved if we are to understandthe
specific case of partial citationand, throughit, the generalphenomenon.We
begin with specific historicalpuzzles - how did it happen that the Thomas
theorem was singled out for attentionand how did it happen that its book
source was not cited from the start?- and then attempt to account for the
continuationof this patternof citationin termsof a genericsocialmechanismin
the culturaltransmissionof knowledge,namely,obliterationby incorporation.
Obliterationby Incorporation
To begin with, how did this one sentencecome to be selectedfor incorporation
in canonicalknowledgefrom some 12,000? sentencescomprisingthat book of
583 pages? After all, the Thomastheoremwas neitherthe core subjectnor the
main theme of TheChildin America.But if it was not centralto the book, it was
central- indeed it was a climacticformulation- for those social scientistsin
W.I. Thomas'sinvisible collegemwho for some time had been drawingupon
his concept of definitionof the situation.It crystallized a new phase of that
evolving conceptby adoptinga pragmaticposition to say much in little about
the subjectivecomponentof actionbeing trulyconsequential.And so members
of thatinvisiblecollege promptlyfastenedonto this new focus on consequences
of definitionsof the situation.
As was briefly noted in reviewing the early diffusion of the theorem,an
integralmemberof thatinvisiblecollegewas the sociologistand socialpsychologist KimballYoung and, to the best of my knowledge, it was he who first
isolated the sentencefor conspicuousattention.This,it may be remembered,he
did soon afterpublicationof T7he
Childin Americaby adoptingit as an epigraph
twice: the first time for Part Four of his textbookSocialPsychology(1930:397)
and the next year for his chapterin an edited anthology,SocialAttitudes(Young
1931:100).The first epigraph ascribesthe quotationsimply to "Thomas"sans
given name or initials (ust as the two other epigraphson that same page are
ascribedto [the tentativelyinferredErnestW.] Hobsonand [thesurely inferred
so
It turns out that RS. Smith has expanded his 1993 presentationto the EasternSociological
Society and that this enlarged paper is scheduled for publication in T7eAmericanSociologist
under the title, "GivingCreditWhereCreditIs Due: DorothySwaine Thomasand the 'Thomas
Theorem.'
31In a felicitous strokeof terminologicalrecoinage,Derekde Solla Price ([1963]1986,passinv
see Index) adapted RobertBoyle's seventeenth-centurytenn 'invisible college" to designate
past and present informalcollectives of closely interactingscientistslmited to a size "thatcan
be handled by inter-personalrelationships."See also Diana Crane,InvisibleColleges(1972).

John]Wesley). The second book, which the dedicatedYoung3 had dedicated

to his master,W.I.,ascribesthe quotationmoredefinitelyto "W.I.Thomas."But
as is commonlythe case with epigraphs,neitherof these gave a bibliographic
referenceto its source.Thus began the unpremeditatedpracticeof quotingthe
Thomassentencesansthe full scholarlycitation.
In examiningits early diffusion,we also noted thatthe first descriptionsof
the sentence as "the Thomastheorem"likewise provided no bibliographical
citation.And this for reasonsas happenstantialas Young'shavingfirst quoted
the sentencein epigraphs.The first such descriptionof the quotedsentenceas
"W.I.Thomas'ssociologicaltheorem"(Merton1938:333)was merelya passing
allusionthatclearlyrequiredno citationwhile the firstdetailedreferenceto the
truncated"Thomastheorem"(Merton1948)appearedin TheAntiochReview,a
journal for the common reader which did not look kindly on footnotes in
general and on bibliographicalcitations in particular.3C(he one citational
footnotein this articlewas not to the theorembut referredto a publicationthat
appearedwhile proofs were being correctedand was barely negotiatedinto
being.) And when the article was reprinted, as it often was, the citation
remainedabsent.
Thus it was owing to peripheraland surely unplannedcircumstancesthat
these early appearancesof the theorem in print did not include the usual
academiccitationsof its source.This chancedfeatureof its primarydiffusion
from membersof the invisible college to the largercommunityof sociologists
became serially reproduced in the course of secondary diffusion through
textbooksand correlativewritings.Thus theredevelopeda special eponymous
variantof the social mechanismknown in the sociology of scienceas "obliteration by incorporation"or by the ultimatebrevity,OBI,an acronymthat stands
for Obliterationof the source of ideas By Iticorporationin currentlycanonical
knowledge (Merton1968:27-28,35-37;Garfield[1975]1977).
This is described as a "special eponymous variant"inasmuch as OBI
ordinarilyinvolvesobliterationof boththe author(s)andthe originalsource.This
unplannedsocialmechanismconsistsin "usersandconsequentlytransmittersof
the particularbit of knowledge[having]becomeso thoroughlyfamiliarwith its
origins that they assume this to be true of theirreadersas well. Preferringnot
to insult their readers'knowledgeability,they no longer refer to the original.
32
At the time Youngwas seizing upon the theoremfor his epigraphs,he also wanted to do
he
a biographicalpiece on W.I.who emphaticallyrefusedto give him leave. Characteristically,
responded to Young's urgent request in this vigorous fashion: "I don't regard myself as
important.I don't want to be noticed. I don't care whether a word appearsabout me in print,
living or dead." Unhappily, Thomas's wish remains fulfilled: there is still no full-scale
biography of this founding American sociologist. Manuscriptletters by Kimball Young (30
April, 1930) and by W.I. Thomas (4 May, 1930). I am indebted to David L. Sills for making
these letters available to me. Although KimballYoung'snewly published oral memoir (1995)
is at times fallible when unchecked by supporting documents, this deftly edited account
provides much detail about the W.I. and the DorothySwaine Thomassociocognitivenetwork.
33 None of the other five articles in that June 1948 issue of TheAntiochReviewcontained a
citational footnote; there were five discursive footnotes all told. The next (September)issue
consisted of 17 articles,with a total of six citationalfootnotes.


source" (Merton1993:218-20)as is the case, for example,with such diverse
knowledge-unitsas the theory of games, secondaryanalysis,revealedpreference, opportunitycosts, or latentstructureanalysis.
- such
At the tail end of the OBImechanism,manyconcepts-and-phrasings
as charisma,stereotype,lifestyle,significantothers,double-bind,androle-model
among countless others - enter the vernacular,with rareif any awarenessof
their sources in the social sciences (Merton 1982:100-106).Obliterationby
incorporationin the generalcultureset in long beforefor much-quoteddicta
and concepts such as FrancisBacon's "knowledgeis power' (1597),Joseph
Glanvill's"climateof opinion"(1661),and JohnAdams's"governmentof laws
and not of men" (1774).But plainly, in those cases where a scientificcontribution has been eponymized,as with Boyle'sLaw,Halley'scomet,Le Chatelier's
principle, the Rorschachtest, Gini coefficient,or Thomas theorem, only its
originalbibliographicsourceand not its authoror authorsbecomesobliterated.
I hazardthe hypothesisthat when the new idea is not eponymized,its source
is more probably,more rapidly,and more extensively,deleted.
Thus,in his fine accountof "theChicagoSchoolof Sociology,"LesterKurtz
(1984:34)in effectexemplifiesboth types of obliterationof publicationonly
and of authortoo - as he reportsthe differingfates of Thomas'suneponymized and eponymizedconcepts:
The most persistentof [W.I.]Thomas'sspecificcontributionsis his emphasison
interactionand situationsin the study of the subjectiveside of sociallife. His conceptof
the 'definitionof the situation'has become one of those conceptsso widely used in
sociologicalanalysis that it is often not explicitlyattributedto Thomas.His situation
analysisis a result of the influenceof pragmatism,and much of his laterwork can be
encapsulatedin his phrase 'If men define situationsas real, they are real in their
consequences' Janowitz1966,p. xl). Muchhas been made of the formulation,calledthe
'Thomastheorem'by Merton(1968,pp. 475ff.),includingeffortsto link it with dramaturgicaland ethnomethodological
perspectives. . .
To be necessarily obvious about it, when authors fail to provide a citation
to the original source of a concept or quotation, this cannot directly lead their
34 Along with providing apt examples of the two degrees of obliteration - complete
obliterationof the source of the genericconcept(definitionof the situation)and inevitablyonly
partial obliterationof the source of the eponymoustheorem- this brief passage provides a
singular array of well-identified pattems in the transnmssionof knowledge. Thus, the
thoroughlyknowledgeablescholarKurtzelects to cite a secondaryratherthan primarysource
of the theorem,taking care to abide by an only slightly institutionalizednorm of citation by
citing that mediating source in precise detail. Degrees of obliterationin the transmissionof
knowledge are also exemplifiedas Kurtzdraws upon Janowitz'sobservationthat"muchof this
later work [by W.I. Thomas on what 'he called situationalanalysis']was encapsulatedin his
[n.b.]phrase,'If men define situationsas real, they are real in their consequences,'as stated in
The Childin America."(janowitz 1966:xl).But then we find that authoritativescholar of the
Thomas corpus, Moris Janowitz, not pausing to give a full citation indicating where the
Thomas sentence is to be found in that expansive 583-page volume. And as a final irony in
Kurtz's meticulous passage, we note that though he cites only a mediating source for the
consequentialsentence itself, he carefully cites the exact primarysource of its eponym, "the
Thomastheorem!'


readersto thatsource.Such omissionsmakefor the type of serialdiffusionthat
merelyreproducesthe patter of an uncitedconceptor quotation.Failureto cite
a source need not, of course, result only from chanicecircumstancessuch as
those at work in the early history of the Thomastheorem.It may also result
froma conflictbetweendiversesocial and cognitivefunctionsof citation.As the
rewardsystem of scienceevolved, payingpeer respectby citingone's predecessors became essential to it, but plainly it would be highly inefficientwere
scholarsrequiredto specify the originsof every bit of incoxporatedknowledge
every time it was put to use. As Kaplan(1965)and Garfield(1995)have in effect
shown, the norms governing citation practice are still neither sufficiently
detailed nor standardizedto solve this problem of conflicting functions of
citation.
At any rate,the Thomastheoremsuffereda deficitratherftiansuch excess
of precisecitation.This, owing to the chanceof its first quotationsby members
of the Thomasinvisible college having appearedin the form of epigraphsand
eponyms thatfailed to signal its precisesource.Thatomissionwas reproduced
throughlaterserialdiffusionamongauthorswho had themselvesevidentlynot
come upon the originalsource in 7heChildin Amenica.
Despite the laterprecise
citationsof the theoremby membersof the Thomasinvisible college (starting
with the Volkartanthologyof W.I.Thomasin 1951),thatoutcomeof obliteration
exceptfor the eponym is, to judge from the Smith (1993)survey,still typicalof
textbooksintroducingstudentsto sociology.Obliterationby incorporationin the
transmissionof knowledge remainsa largely intact process until subjectedto
exogenous historicalor sociologicalexamination.
But if these social mechanismsof OBIand serial diffusionaccountfor the
widespreadfailureto cite the book sourceof the theoremat all, they plainly do
not accountfor this double case of the partialcitationphenomenonwhich has
W.I.Thomasbeing solely creditednot only with the theorembut also with the
book in which it appeared.Eachof these types of partialcitationderivedfrom
distinctthough mutuallyreinforcingsocial mechanismsin the transmissionof
knowledge and so requiresseparateanalysis.
PrimaryGroup Ascription of the Theorem to W.L
There was nothing problematicalabout the origin of the theorem among
membersof the Thomassociocognitivenetworkand particularlyamong those
who at the time the theoremappearedor soon afterwardswere in close touch
with the central figures of that network, W.I. Thomas and Dorothy Swaine
Thomas., Membersof the network who have entered this account as early
35 Much tacit and explicit knowledge was exchangedin such networks then as now, Those
networkswere, of course,less numerousand less specializedbackin the 1920s and 1930swhen
the membershipof the AmericanSociologicalSociety numberedsome 1200, less than a tenth
as many as the membershipof the ASA in the 1970s and 1980s. (I am indebtedto ValeriePines
of the ASA for this information.)One-timemembersof the Thomasnetworksare of course in
increasinglyshort supply; my own relativelyyouthful engagementwith the Thomasesdid not
begin until W.I.came to Harvardas a visiting professorin the mid-1930s,the year I becamea
newmade instructor.


disseminatorsof the theorem - KimballYoung, Howard P. Becker,Willard
Waller, and myself - had ample cause to regard the theorem simply as a
memorableformulationof a new aspect of W.I.'slongstandingparadigmatic
idea of definition of the situation. Which we know was also the case with
Dorothy Thomas herself since, without any promptingon my part, she had
volunteeredas much in her smoking-gunletter (cf. Exhibit6).
However,what was transparentlyobviousat the time to Thomascolleagues
close-by and at-a-distancehas evidently become problematicin this time of
politicalcorrectitude- at least for those who define the sole ascriptionof the
theorem to W.I. Thomas not as a matter of historicalfact but simply as an
expressionof sexism. It may thereforebe useful to collate a few more bits of
evidencefromboth public printand privatearchivethatbearupon this current
definitionof this particularsituation as furtherprelude to an analysis of the
widely diffused partialcitationof the book and the theorem.
DS1fS REIROSPEC'IWEON The Child in America
SometwentyyearsbeforeDorothySwaineThomashad crisplyand emphatically describedW.L'sand her own distinctivecontributionsto the book in the
smoking-gunletter, she had taken the occasionof her PresidentialAddress to
the AmericanSociologicalSociety to do so in much greaterdetail. An incomplete private archive yields an almost instant response to the Address that
impatientlyaimed to spread the word just a bit more quicklyin a way typical
of sociocognitivenetworks.
26 September1952
DearDorothy:
Thinkingback,I find thatrm glad to be a memberof the Council.I am sureI would
have seen nothing of you at the AtlanticCity meetingsif I hadn't been among those
fortunatefew who spent almost three whole days in your privatesuite. rve told Paul
Lazarsfeldaboutyour presidentialaddressand if you have a sparecopy of it, couldyou
send it on for him to read so that he will not have to wait for publication?...
The responsefollows soon:

October3, 1952
DearBob,
I too enjoyedthe threedays of Councilmeetings.Thanksforyourkind remarks.I do
not have an extra copy of the PresidentialAddress, but it will be published in the
Decemberissue of the Review.Underseparatecover,I will send you, within a few days,
a copy of a very good pictureof W.I ...
What,then,does DorothyThomas(1952:665)have to say aboutour subject

in that Address by the first woman to become president of the then ASS?
Herewith,a few snippets from this autobiographicalpiece on the life of the
mind (thatstill bears re-reading):
The frameworkof The Child in America was W.I. Thomas's famous situational
approach... which definedthe 'totalsituation'as alwayscontainingmoreor less of the
subjective... It is always dangerousto try to reconstructthe separatecontributionsof


collaborators,but I am reasonablysure that the designationof subjective,documentary
materialsas the 'as-yetunmeasured'and the emphasison 'transmuting'moreand more
factors'into quantitativeform'were mine and that the very positiveevaluationof the
behaviordocumentper se was W.I. Thomas'.For when I joined the staff of the Child
DevelopmentInstituteat TeachersCollegein 1927,I was stillsomewhatdistrustfulof the
subjectiveand the 'as-yet unmeasured'as materialsfor scientificinvestigations.I still
preferredto work with the objective,definedin almostmechanistictermsand to count,
measure,sample,fit curves,correlate,test for reliability,validityand the significanceof
quantitativedifferences,ratherthan to utilizedescriptivematerialsor life histories,case
records,and other types of personaldocuments.I hoped, indeed, that the series of
observationalstudiesof socialbehaviorwhich I directedthereand continuedduringthe
1930'sat the YaleInstituteof HumanRelationsmightyield 'dataas objectiveas the best
were dealing.And althoughI gave verbal
of thosewith which the statisticalecononmsts'
recognitionto the value of case histories,diary records,and what I called 'merely
descriptive'accountsof behavioras 'hypothesis-forming
materialfor furtherstudies'I
made slight use of these materials,on the groundthat they 'obviously[would]not yield
data appropriatefor statisticalanalysis'(D.S.Thomas1952:665citingD.S.Thomas1929:
19-30,passim).
This public avowal with its emphasis on Dorothy Swaine Thomas's

commitmentto 'objective'statistical analysis inevitably brings to mind her
statementin the 1973 smoking-gunletter that "the statisticalportions [of 7he
Childin America]were mine."Thatspecializedrole was in effectreaffirmedby
W.I. (who insisted on declaringhis technicalignoranceof statistics)when he
concludedone of the only two papersin which he reproducesthe theoremby
statingthat "DorothySwaineThomas,of YaleUniversity,is responsiblefor the
items relatingto statisticalprocedurein this article."
DSrS TACIT APPROVALOF VOLKARTS 1951 ASCRITON TO W.L
At almostthe same time,DorothyThomasis providingbehavioraltestimonyto

the division of scientific labor in the book and to virtual ascription of the
theoremto W.I.Fouryears afterW.I.'sdeath in 1947,she is first among many
who are troubled by the lack of access to his works "in the field of social
behavior"thatwere out of printor had neverbeen published(Thomas1951:xi).
She draws upon three colleagues, near and far, in her own sociocognitive
network- DonaldYoung,ThorstenSellin,and HerbertBlumer- to servewith
her as a committee of the Social Science ResearchCouncil to oversee the
collectionof W.I.'swritingsthatwas to be broughttogetherby EdmundVolkart
(Thomas1951).Volkartof courseelectsto reprinttheconsequentialfinalchapter
of The Childin America(duly cited as written by the two Thomases)and
observesin his introductionto it that "theimportanceof subjectiveexperience
to a science of behavioris still emphasized."Thatcontinuingemphasison the
subjectiveis manifestlybeing ascribedto W.I.and surelynot to DorothySwaine
Thomasfor, as we know, she is just then reportingin her presidentialaddress
her (since relaxed)"distrustof the subjective"at the time the book was being


written.36Volkartgoes on to note, as a matterof course,that "Inthis connection Thomas' [n.b.,not "the Thomases"']discussion of the life history as a
sourceof researchmaterialshouldprove especiallyusefulto studentsof culture
and personality."As a member of the committee supervising the volume,
Dorothy Thomas evidently did not object to this reading of the chapter
containingthe paragraphon the value "of the highly subjectiverecord... for
behaviorstudy" which culminatesin the theorem.Once again, we are put in
mind of her smoking-gun letter which resonates with that reading of the
chapter when she writes that "the statisticalportions were mine and I am
sending you under seperate[sic] cover Volkart'sbook which makes this clear.
The conceptof 'definingthe situation'was strictlyW.L's."
WrrS SMOKING-GUNASCRIPTIONOF THE THEOREMTO HIMSELF
W.I. evidently agreed. For long before, he had anticipatedDorothy Swaine

Thomas'sprivate smoking-gunletter in a public smoking-gunascriptionto
himselfwhile addressinga panelof socialscientistsappraisingBlumer'scritique
of ThePolishPeasant;this, in the course of yet again statinghis methodological
case for "the behavior document, whether autobiography,case record, or
psychoanalytic exploration."37In a matter-of-factvein (here italicized for
emphasis,no doubt needlessly),he remarksthat "I quote what I said in this
connection"and then proceedsto quote the passagefrom TheChildin America
we have come to know so well, the one that closes with the now symbolically
historic sentence: "If men define situations as real, they are real in their
consequences"(W.I.Thomasin Blumer1939:85).
DOCUMENTARYCONFIRMATIONOF THE DSr AND WIT SMOKING-GUNASCRIPTIONS
Even in the absence of these archival and published documents by both

Thomases,intellectualhistorianswould have little difficultyin ferretingout the
36 I
write "since relaxed distrust of the subjective"inasmuch as the presidential address

concludes, with typically Thomasiancandor: "On the behavioral side, I have not found it
profitableto proceed as if all behaviormust be or even can be 'transmuted'into quantitative
terms. And whereas I still push the statistical aspect of all studies to the limit, I no longer
relegate the subjectiveand the descriptiveto secondarypositions"(Ihomas 1952:669). It was
that kind of candid public retrospectivethat elicited this fan letterdrawnfrom my own private
archive: "I salute our out-going President - this being said as a sociologist and referring
thereforeto the Chief Executiveof the A.S.Sratherthan the U.S.A.The particularoccasionfor
my drinking to your good health is the appearance, in print, of your salty, meaty and
otherwisenourishingpresidentialaddress.It stands up on the printedpage as it did beforethe
collected audience.Most of us never try to make sense of the life of the mind we have led, and
of the few who do try fewer still succeed.Yourswas a complete success, all the more solid for
being wholly unpretentious"(RKMto DST,27 December1952).
37 This, it should be noted, is a virtual quotationfrom a comparablepassage in ne Childin
Americawhich reads: "The behavior document (case study, life-record, psychoanalytical
confession..." (571)


highly probable authorship of the theorem. For neither before nor after
publication of The Childin Americadid Dorothy Thomas make sustained use of
the theorem or the more inclusive concept of definition of the situation whereas
W.I. Thomas devoted much of his twentieth-century worklife to what he
described as 'situational analysis." In 1917- Dorothy Swaine Thomas was then
17 and about to enter Barnard College - his influential paper on 'The
Persistence of Primary-group Norms in Present-day Society" observes that
'this defining of the situation is begun by the parents. .." and a few years later,
The Unadjusted Girl (Ihomas 1923) is 'mainly concerned with situational
analysis and the definition of the situation" Janowitzin Thomas 1966:xxvii).But
rightly enough, it is his presidential address to the American Sociological
Society, a year before publication of the for-us landmark book, The Child in
America,that almost wholly anticipates the formulation of the theorem:
A documentpreparedby one compensatingfor a feeling of inferiorityor elaboratinga
delusionof persecutionis certainlyas far as possiblefromobjectivereality.On the other
hand, this definitionof the situationis from one standpointquite as good as if it were
true.It is a representationof the situationas appreciatedby the subject,"asif' it were so,
and this is for behaviorstudy a most importantphase of reality(Thomas1927:7).
There it is: the essential idea, down to the detail of including the Hans
Vaihinger ([1911] 1924) phrasing of "as if" as this appears in the canonical
version of the theorem. This anticipatory version lacks only the pragmatic
element of consequences made explicit and the felicitous formulation that made
the theorem memorable.
In light of this cumulation of private and public evidence, it does not seem
extravagant to conclude that Dorothy Thomas and W.I. Thomas were probably
speaking truth about the origin of the theorem.
ZNANIECKI'S SUMMARYOF THE WIT AND FZ DIVISION OF SOCIOLOGICALLABOR
This accumulation of evidence would also seem to bear upon the suggestion by
Howard P. Becker that the theorem was "in content at least. . . probably
Znaniecki's." Beyond this evidence is Znaniecki's statement to that panel of
social scientists engaged in appraising Blumer's appraisal of ThePolish Peasant.
There he summarizes Thomas's and his own "previous results of comparative
analysis and generalization," thus:
Thomashad at the timealreadyformulatedseveralwell knownand originaltheories
in socialpsychologyand sociology,based upon an exceptionallygreatmass and variety
of significantdatacarefullychosenfrommanydifferentcultures;and in startingto collect
materialsconcerningEuropeanpeasantshe meantto applyhis theoriesto this new mass
of data.I had publishedseveralworks in generaltheoryof cultureand in epistemology
which eventuallyprovedto have somebearing,howeverabstractand indirect;the former
38
Thus, in an article anatomizingtextbooksin social pathology,C. WrightMills (1943:171)

notes that "About the time W.I. Thomas stated the vocabularyof the situationalapproach,a
social worker was finding it congenial and useful. In M.E. Richmond's influential Social
Diagnosis(1917)we gain a clue as to why pathologiststend to slip past structureto focus on
isolated situations. . ."

TheThomasTheoremand the MatthewEffect/ 415

upon the data of Polish peasant culture,the latter on the method of handling them.
(Blumer1939:87,90)
Thus, typical of both, Thomas takes the occasion to focus anew on his
conceptof "definingthe situation"in generaland on his theoremin particular
while Znanieckihas nary a word to say about conceptor theorem.Instead,he
straightforwardlydistinguishes"Thomas'swell known and originaltheoriesin
socialpsychologyand sociology"(which, of course, were centeredon his idea of
"definitionof the situation)from his own "theoryof cultureandepistemology."
In
this way, Znanieckiis also beingforthrightfor,like DorothySwaineThomas,he
too made no sustaineduse of the idea eitherbeforeor afterthe greatcollaboration.39
So much, then, about how it was that members of the Thomas sociocognitive network ascribedthe theorem solely to W.I. Given their first-hand
knowledge of its origin, no alternativecould possibly have occurredto them.
But,as I have emphasizedfromthe start,this articleis primarilyconcernedwith
examining the Case of the Thomas Theoremin an effort to understandthe
genericpartial-citationphenomenon,not with the specificmatterof adjudicating
its origin. And though early obliterationby incorporationand serial diffusion
may explain the partialcitationof the theorem
by the authorsof textbooksand
others who had no direct access to this first-handinformation,those social
mechanismscannot,of course, explainthe practiceof ascribingthe booksolely
to W.I.,since that was not a practiceappearingoften in the early diffusionof
the theorem.
However, it may be rememberedthat documentationsupplied in the
archivalletterto SSSbrieflyproposedthatthis specificcase of partialcitationof
the book may have resulted from the Matthew effect operatingas a generic
social mechanismin the transmissionof knowledge. That proposal warrants
furthertheoreticaland methodologicalscrutinyin light of the concernvoiced by
SSSthatascribingthe theoremto W.I.was sexistand the renewedinterpretation
by Smith(1993)thatthis practice"canalso be explainedin termsof a structural
issue - the genderizationi4of the disciplineas partof the processof professionalization.By not citing Dorothy Swaine Thomasthese authorshelp sustain a
view of sociology as historicallya male domain."
39 In strikingcontrast to Thomas who, in his last, massive, work, PrimitiveBehavior(1937),

draws upon the favored mode of situational analysis from its first page to its last. The very
first sentence in the book announces that every aspect of culture "can best be approachedin
terms of thedefinitionof thesituation"while the index lists 28 pages which explicitly deploy that
concept.
40 The "genderizationeof scientific disciplines has been described by Evelyn Fox Keller
(1985) as involving the domination of the sciences by men scientists at the expense of
recognition of women scientists. Keller (1991) has gone on to put her argument thus: "the
exclusion of the feminine from science has pertained to a particular definition of science:
science as incontrovertiblyobjective,universal, impersonal- and also masculine"(235).


The Partial Citation Phenomenon and The Matthew Effect
As the terms imply, partial citation phenomenondesignates a fact while the
Matthew effect designates a theory. There is no reasonable question that
incomplete citations of authors occur but it remains to be shown whether and
to what extent and under which conditions they are cases of the Matthew effect.
This is not the place to attempt a further systematic explication of the effect but,
as noted earlier, there has been a tendency to conflate the fact and the theory,
the phenomenon and its proposed explanation.
In cases of collaboration between scholars of notably unequal reputation,
the Matthew effect confers excessive credit on the better-known scholar(s) and
little or none on the other collaborator(s). Thus, the biologists R.C. Lewontin
and J.L. Hubby instructively describe the far larger numbers of citations
accorded one of a pair of their joint papers as apparently
a clearcutcaseof [the]'MatthewEffect'. . . In 1966,Lewontinhad been a professionalfor
a dozenyearsandwas well knownamongpopulationgeneticists,to whomthe paperwas
addressed,while Hubby's careerhad been much shorterand was known chiefly to
biochemicalgeneticists.As a result,populationgeneticistshave consistentlyregarded
Lewontinas the seniormemberof the teamand given him undue creditfor what was a
completelycollaborativework thatwould have been impossiblefor eitherof themalone.
(Lewontin& Hubby 1985:16)
This report holds immediate interest for us here since in this case, as in
countless others, the Matthew effect cannot be easily attributed to a difference
of gender. After all, their fellow scientists know that both Lewontin and Hubby
are males. The fact that partial or other forms of skewed citation most often
refer to author-sets of the same gender is enough to raise the interesting
theoretical question why, in the case of the Thomas theorem, failures to include
Dorothy Thomas in citations of the theorem or the book have been promptly
attributed to her being a woman and his being a man. Both W.I. and D.S.
Thomas occupied complex status sets.41 He was not merely another male
sociologist nor she merely another female sociologist. They had many other
differing statuses and distinctive social attributes. For example, the attribute of
their comparative standing in the discipline and in the field of social psychology
(in which Dorothy Swaine Thomas had not worked at all). At the time that one
memorable sentence appeared in the coauthored Childin Americaand diffused
as the Thomas theorem, W.I. was widely recognized as one of the most
consequential of American sociologists. (A quick and, for historians of sociology,
redundant indicator of his standing was provided by a study of the comparative
amount of space devoted to founders of the discipline in historical textbooks
that found him ranked first among living sociologists and sixth among the likes
of Durkheim, Comte, Spencer, Ward, and Max Weber [Palmore (1962) 1971].)
41
On the structuralconcept of status-set as the set of distinct socially defined positions

occupied by individuals at a given time, see Merton (1968:434-38);for its application to the
phenomenaof "sex typing" and structuralobstaclesconfrontingwomen in Americansociety,
see Epstein (1970:86-101;1988:101-28).


Thomas was esteemed in particular for his theoretical development and
applicationof situationalanalysis.At the age of 64, just a year beforepublication of The Childin America,he had finally been elected President of the
American Sociological Society (presumablybecause enough members had
abandonedtheir prudish attitudestoward his alleged notoriousadulterythat
had led to his dismissalfromthe Universityof Chicagoten yearsbefore).Along
with his exalted place in the social stratificationof Americansociology and
other statuses in his status set, such as his age, race, religion, ethnicity,
universityaffiliation,W.I.Thomasalso happenedto be a man.
In contrastto W.I.Thomas'sworldwidefame,DorothySwaineThomashad
ago (see Exhibyet to achieve her fame. And, as was noted a quarter-century
it 1), she was also "subjectto the double jeopardy of being a woman of
sociologicalscience and still in her twenties."
With such enormous differencesin the extent of obviously age-related
accomplishmentand reputationbetween the two coauthors,why should we
fasten onto the one status differenceof gender to explain this case of partial
citation?On what grounds should we assume that this one status determined
both the amply warrantedascriptionof the theoremand the wholly mistaken
ascriptionof the book to W.I.Thomasalone?To the extentthat gender-,race-,
age-,ethnic-,or otherstatus-influencedascriptionsdo enterinto particularcases
of the partialcitationphenomenonthatplainlyinvolve the Mattheweffect,they
involve overdetermination
(in the methodological,not thepsychoanalytic,sense
of havingmore determiningfactorsthanthe minimumnecessaryto bringabout
the outcome).All this raises a series of theoreticaland methodologicalquestions.42 How do we go about discovering whether and to what extent cog-
nitively irrelevantstatusesof authorsand of peer ascribersmakefor the partial

citationphenomenon?Do scientiststendto attributethe primerole in collaborative work by men and women to those of their own gender?If so, does this
practice differ by gender? Does it obtain irrespective of the comparative
standingof the collaboratorsgenerallyand in their special fields of investigation? In short, there is still much to be done by way of systematicempirical
investigationof the diverseworkingsof the Mattheweffectin relationto such
functionallyirrelevantstatuses.43
42
Thesegenerictheoreticaland methodologicalquestionsalso hold specificpersonalinterest.

For,as Eugene Garfield(1994:13)has recentlyreported,my colleague HarrietZuckermanand
I have long been subjected to a pattern of mis-citationreminiscentof the Lewontin-Hubby
experiencewith the skewed distributionof citations to their two joint papers. Paperswritten
jointly by "Zuckermanand Merton"(1971,1972) are often cited with the order of the authors
reversed. This, of course, carries its own irony, since the author who Garfield notes "had
identified, named, and harshly criticized" the Matthew effect thus becomes its dubious
"beneficiary."The irony becomes all the greaterin light of a statementinsertedin a reprinting
of the 1968 paper, "TheMatthew Effectin Science":"It is now [1973]belatedly evident to me
[RKM1that I drew upon the interview and othermaterialsof the Zuckermanstudy to such an
extent that, clearly, the paper should have appeared under joint authorship"(Merton[and
Zuckerman][1968]1973:439).
43 For inquiries into the workings of the Matthew effect, though without referenceto this
matterof gender-or otherstatus-influencedattributions,see StephenCole (1970;1992:chap. 6).


In their normativeaspect, the partial citation phenomenonand Matthew
effectmakefor injustices.Theyviolatethe basicnormof giving peerrecognition
of contributorsto the commonwealth of science and scholarship.And, as we
have seen, such normativeviolations evoke deep-seatedresponses.To extend
the normativeproblemin the case of collaborationamongstatus unequals,the
tendency toward this type of injustice is systemic. That systemic inequity
derives in no small part from there seldom being public evidence of the
respective parts taken by collaboratorsin a particularresearch since the
standardformatof the scientificpaper"presentsan immaculateappearancethat
reproduceslittle or nothing of the intuitive leaps, false starts,mistakes,loose
ends, and happy accidents that actually clutteredup the inquiry. Thepublic
recordof sciencetherefore
fails to providemany of the sourcematerialsneededto
(Merton 1968:4;italics
reconstructthe actual courseof scientflc developments"
inserted).44
Absent such detailed information, fellow scientists and scholars are
evidentlyinclinedto thinkit "reasonable"thatthe more accomplishedcollaboratorwith a histoxyof majorcontributionsto the field - i.e., the one with the
far better "trackrecord' as it is often put - has probablyoriginateda joint
work or contributedmore to it, - unless there is compellingevidence to the
contrary.This,even thoughsuch a probabilisticinferenceof coursetells us next
to nothing about the particularcase with certainty.
However, in the case of the Thomastheorem,the compellingevidence is
there and this time it is not to the contrary.
References
Bartlett,John. [1882]1992. FamiliarQuotations,edited by JustinKaplan.16th ed. Little,Brown.
Blumer,Herbert.1939. An Appraisalof Thomasand Znaniecki's'ThePolishPeasantin Europeand
America.'Social Science ResearchCouncil.
Broad,William, and Nicholas Wade. 1983. Betrayersof theTruth.Simon & Schuster.
Burchell,Howard B. 1985. "Thoughtson Eponyms."International
Journalof Cardiology
8: 229-34.
Cole, JonathanR. 1979. FairScience:Womenin the ScientiflcCommunity.
Free Press.
Cole, JonathanR., and Harriet Zuckerman.1984. "The ProductivityPuzzle: Persistenceand
Change in Patternsof Publicationof Men and Women Scientists."Pp. 217-58 in Advances
in MotivationandAchievement.
edited by P. Maehrand M.W.Steinkamp.JAIPress.
Cole, Stephen. 1970. "ProfessionalStanding and the Reception of Scientific Discoveries."
AmericanJournalof Sociology76:286-306.
. 1992. MakingScience:BetweenNatureand Society.HarvardUniversityPress.
Coleman,JamesS. 1990. Foundations
of SocialTheory.HarvardUniversityPress.
Coleman,JamesS., Elihu Katz, and HerbertMenzel. 1966. MedicalInnovation.Bobbs-Merrill.
Crane,Diana. 1972. InvisibleColleges.University of ChicagoPress.
44 It therefore seems only appropriatethat a paper based on private as well as public

knowledge of the respective roles of the distinguishedpair of collaboratorsin this case should
depart from the rigid pieties of the standardformat.


Cronin,Blaise. 1984. 7he CitationProcess.TaylorGraham.
Holmes.Doubleday.
Doyle, ArthurConan [1894]1953. 'Silver Blaze"in TheCompleteSherlock
Epictetus,[c. 110] 1926-1928.TheDiscoursesandtheEncheiridion.
Translatedby W.A. Oldfeather.
2 vols. Putnam
Epstein, Cynthia Fuchs. 1970. Woman'sPlace: Options and Limits in ProfessionalCareers.
University of CaliforniaPress.
-.1988.
DeceptiveDistinctions:Sex, Gender,and the SocialOrder.Yale University Press &
Russell Sage Foundation.
Garfield, Eugene. [1975] 1977. 'The 'Obliteration Phenomenon' in Science - and the
Advantage of BringObliterated."Pp. 396-98 in Garfield,Essaysof an Information
Scientist.
Vol. 2. ISIPress.
. 1982. "Moreon the Ethicsof ScientificPublication:Abuses of AuthorshipAttributionand
Citation Amnesia Undermine the Reward System of Science." Pp. 621-26 in Garfield,
Essaysof an Information
Scientist.Vol. 5. ISIPress.
1994. "Commentary."TheScientist.17 Oct., 1994.
1995. "Whento Cite." LibraryJournal.In press.
Garfield,Eugene,IrvingR. Sher,and RichardJ. Torpie.1964. TheUseof CitationDatain Writing
theHistoryof Science.ISIMonographs.Institutefor ScientificInformation.
and Competition
in Science:A Studyof the BritishHigh Energy
Gaston, Jerry.1973. Onrginality
PhysicsCommunity.
University of ChicagoPress.
Review39:1-18.
Hagstrom,Warren.1974. "Competitionin Science."AmericanSociological
James,William. [1890]1950. ThePrnciplesof Psychology.Dover.
Kaplan,Norman. 1965. Scienceand Society.Rand McNally.
on Genderand Science.Yale UniversityPress.
Keller,Evelyn Fox. 1985. Reflections
. 1991. "The Wo/Man Scientist: Issues of Sex and Gender in the Pursuit of Science."
Pp. 227-36 in The Outer Circle:Womenin the ScientificCommunity,edited by Harriet
Zuckerman,JonaffianR. Cole, and JohnT. Bruer.W.W.Norton.
Kurtz, Lester. 1984. EvaluatingChicagoSociology:A Guideto the Literaturewith an Annotated
Bibliography.
University of Chicago Press.
and Economic
Kuznets,Simon, and Dorothy S. Thomas(eds.). 1959-64.PopulationRedistnbution
Growth:UnitedStates,1870-1950.3 vols. AmericanPhilosophicalSociety.
Sciences43:16.
Lewontin,R.C.,and J.L.Hubby. 1985. "CitationClassic."CurrentContents/Life
Mead, George H. 1936. Movementsof Thoughtin theNineteenthCentury.University of Chicago
Press.
and Societyin Seventeenth-Century
Merton RobertK [1938] 1970. Science,Technology,
England.
Howard Fertig.
. [1938] 1973. "Scienceand the Social Order."Pp. 254-66 in Merton 1973. University of
Chicago Press.
. 11942]1973. "TheNormativeStructureof Science."Pp. 267-78in Merton1973. University
of Chicago Press.
-.
[1948]1982. "TheSelf-FuIfillingProphecy."Pp. 248-67in Merton1982, edited by Aaron
Rosenblattand ThomasF. Gieryn Abt Books.
. 1948a. "Discriminationand the American Creed. Pp. 99-126 in Discriminationand
NationalWelfare,edited by R.MTMacIver.Harper& Brothers.
. [1961]1973."Singletonsand Multiplesin ScientificDiscovery."Pp. 342-70inMerton1973.
University of ChicagoPress.
-.
[1961a]1973."SocialConflictover Styles of SociologicalWork Pp. 47-69in Merton1973.
-.
[1963]1973. "MultipleDiscoveriesas StrategicResearchSite."Pp. 371-82in Merton1973.
. [1949,195711968. SocialTheoryand SocialStructure.Free Press.


. 1973. TheSociologyof Science.Universityof ChicagoPress.
. 1976. "SocialKnowledge and Public Policy."Pp. 156-79in Sociological
Free
Ambivalence.
Press.
Traditions
from
.1980. 'On the Oral Transmissionof Knowledge."Pp. 1-25 in Sociological
Generationto Generation,edited by Robert K. Merton and Matilda White Riley. Ablex
PublishingCorporation.
. 1981. "OurSociologicalVernacular."Pp. 100-106in Merton1982. Abt Books.
.1982. SocialResearch
andthePracticingProfessions.
Editedby AaronRosenblattand Thomas
F. Fieryn.Abt Books.
. 1983. "FlorianZnaniecki:A Short Reminiscence."Journalof theHistoryof the Behavioral
Sciences10:123-26.
. 1984. "SociallyExpectedDurations:A Case Study of ConceptFormationin Sociology."
In Honorof LewisA. Coser.Editedby WalterW. Powell
Pp. 262-83in ConflictandConsensus:
and RichardRobbins.Free Press.
1987. "ThreeFragmentsfrom a Sociologist'sNotebooks:Establishingthe Phenomenon,
SpecifiedIgnorance,and StrategicResearchMaterials."AnnualReviewof Sociology13,1-28.
1988. "TheMatthew Effectin Science,II:CumulativeAdvantage and the Symbolismof
IntellectualProperty."Isis 79:606-23.
.1989. "Lemolteplici originie il carattereepiceno del termineinglese Scientist."Pp. 279-93
in Scientia:L'immagine
e il mondo.Scientia.
- [1965]1993. On theShoulders
of Giants:A Shandean
Postscript.Universityof ChicagoPress.
1995. "OpportunityStructure:The Emergence, Diffusion, and Differentiation of a
Sociological Concept, 1930s-1950s.' Pp. 3-78 in The Legacyof AnomieTheory.Vol. 6 of
Advancesin Criminological
Theory,edited by FredaAdler & WilliamS. Laufer.Transaction
Publishers.
. 1996. "De-Gendering'Man of Science':The Genesis and EpiceneCharacterof the Word
Scientist."In Sociological
Visions,edited by Kai Erikson.Yale UniversityPress. In press.
Merton, RobertK. [and Haniet Zuckerman].[1968] 1973. "The Matthew Effect in Science."
Pp. 439-59 in Merton1973. Universityof ChicagoPress.
Mills, C. Wright.1943. "TheProfessionalIdeology of Social Pathologists."AmericanJournalof
Sociology49:165-80.
Ogburn,William F. 1922. SocialChange.B.W.Huebsch.
Ogburn,WilliamF., and DorothyThomas.1922. "AreInventionsInevitable?:A Note on Social
Evolution."PoliticalScienceQuarterly37:83-98.
Palmore,Erdman.[1962]1971. "Sociologists'Class Originsand PoliticalIdeologies."Pp. 23843
in T7he
Phenomenon
of Sociology,edited by EdwardA. Tiryakian.MeredithCorporation.
Price, Derek J. de Solla. [1963] 1986. Big Science,Little Science. . . and Beyond.Columbia
University Press.
Reinharz,Shulamit.1989. "Teachingthe History of Women in Sociology:Or Dorothy Swaine
Thomas,Wasn'tShe the WomanMarriedto William I.?"TheAmericanSociologist,Spring
1989:87-94.
The Free Press.
Rogers,EverettM. [1962,1971, 1983] 1995. Diffusionof Innovations.
Sarton, George. 1931. Introductionto the History of Science.Vol. 2 in Two Parts. Carnegie
Institutionof Washington.
Translated by E.F.J.Payne.
Schopenhauer,Arthur. [1851] 1974 Parergaand Paralipomena.
ClarendonPress, Oxford.
Schutz, Alfred. 1962. CollectedPapers.Edited by MauriceNatanson.2 vols. MartinusNijhoff.
ScienceWatch.1995. "ReallyBig Science:MultiauthorPapersMultiplyingin the 1990s."Science
Watch6:1-2.


Sills, David L., and Robert K. Merton (eds.). 1991. SocialScienceQuotations:Who Said What,
When,and Where.Macmillan.
Smith, R.S. 1993. "DorothySwaine Thomas and the 'ThomasTheorem'."Paper presented to
the EasternSociologicalSociety.
__.
1995. "Giving Credit Where Credit Is Due: Dorothy Swaine Thomas and the 'Thomas
Theorem'."TheAmericanSociologist.In press.
Stigler,GeorgeJ. 1961. "TheEconomicsof Information."Journalof PoliticalEconomy69: 213-25.
Stigler,StephenM. 1980. "Stigler'sLaw of Eponymy."Pp. 147-57in ScienceandSocialStructure,
edited by ThomasF. Gieryn.Transactions
of theNew YorkAcademyof Sciences,series II,Vol.
39.
Tate, Leland.B. 1974. "WI. Thomas:PioneerSociologistand His Kin."In HistoricalSketchesof
SouthwestVirginia.Southwest VirginiaHistoricalSociety.
Thomas,Dorothy Swaine. 1938. ResearchMemorandum
on MigrationDifferentials.
Social Science
ResearchCouncil.
.1952. "Experiencesin InterdisciplinaryResearch."AmericanSociological
Review17:663-69.
Thomas,DorothySwaine and Associates.1929. SomeNew Techniquesfor
StudyingSocialBehavior.
TeachersCollege, ColumbiaUniversity.
Evacuationand
Thomas, Dorothy Swaine, and RichardS. Nishimoto. 1946. Japanese-American
Resettlement:
TheSpoilage.University of CaliforniaPress.
Thomas, Dorothy Swaine, Charles Kikuchi, and James Sakoda. 1952. Japanese-American
Evacuation
and Resettlement:
TheSalvage.University of CaliforniaPress.
Thomas,W.I.1917. "ThePersistenceof Primary-groupNorms in Present-daySociety and Their
Influence in Our Educational System." Pp. 159-97 in Suggestionsof Modern Science
ConcerningEducation,by Herbert S. Jennings,John B. Watson, Adolf Meyer, and W.L
Thomas.Macmillan
.1923. The Unadjusted Girl. Little, Brown.
. 1927 "SituationalAnalysis: The BehaviorPattem and the Situation."Publicationsof the

AmericanSociological
Society22:1-13.
. 1931. "TheRelationof Researchto the SocialProcess."Pp. 175-94in Essayson Researchin
the SocialSciences.The BrookingsInstitution.
_
1937. Primitive Behavior.McGraw-Hill.
1951. SocialBehaviorand Personality:Contributions

of W.I. Thomasto Theoryand Social
Research.Edited by EdmundH. Volkart.Social ScienceResearchCouncil.
.1966. On SocialOrganizationand SocialPersonality:SelectedPapers.Edited and with an
introductionby MorrisJanowitz.University of ChicagoPress
Thomas,W.I., and Dorothy Swaine Thomas.1928. TheChildin America:BehaviorProblemsand
Programs.Knopf.
Thomas,W.L,and FlorianZnaniecki.[1918-20]1927. ThePolishPeasantin EuropeandAmerica.
Five volumes in two. Knopf.
Vaihinger,Hans. [1911]1924. ThePhilosophyof 'As If.'Routledge-KeganPaul.
von Wiese, Leopold.1932. SystematicSociology,On the Basisof theBeziehungslehre
und Gebildelehre.Adapted and amplified by Howard Becker.Wiley.
Waller,WillardW. 1932. TheSociologyof Teaching.Wiley.
. 1938. The Family:A Dynamic Interpretation.Cordon.
An Analysisof SocialBehavior.Crofts.
Young,Kimball.1930. SocialPsychology:
. 1931. Social Attitudes. Hnemy Holt.
Presidentof
. 1995. On Sociologyin Transition,1912-1968:An OralAccountby the Thirty-fifth
the ASA, edited by Fred B. Lindstrom, Ronald A. Hardert, and Laura L. Johnson.
University Press of America.


Zuckerman,Harriet.1968. "Patternsof Name-OrderingAmong Authorsof ScientificPapers:A
Study of Social Symbolismand its Ambiguity."AmericanJournalof Sociology74:276-91.
. 1977. "DeviantBehaviorand Social Controlin Science."Pp. 87-137in Devianceand Social
Change,edited by E.A. Sagarin.Sage Publications.
. 1978. "TheoryChoice and ProblemChoice in Science."Sociological
Inquiry48:65-95.
. 1988a. "The Sociology of Science.' Pp. 511-74 in Handbookof Sociology,edited by Neil
Smelser.Sage Publications.
. 1988b. "IntellectualProperty and Diverse Rights of Ownership in Science." Science,
andHumanValues.Vol. 13, nos. 1 & 2:7-16.
Technology,
_. 1989."Accumulationof Advantage and Disadvantage:The Theory and Its Intellectual
edited by
Biography."Pp. 153-76 in L'Operadi R.K Mertone la sociologiacontemporanea,
Carlo Mongardiniand SimonettaTabboni.Edizioni CulturaliInternazionali.
. JonathanR. Cole, and JohnT. Bruer(eds.). 1991. TheOuterCirde:Womenin the Scientific
W.W.Norton.
Community.
Zuckerman,Harriet, and JonathanR. Cole. 1994. "ResearchStrategiesin Science,' Creativity
ResearchJournal7:391-405.
Zuckerman,Harriet, and RobertK. Merton.1971. "Patternsof Evaluationin Science:Institutionalization,Structureand Functionsof the RefereeSystem."Minerva9:66-100.
__.
1972. "Age, Aging, and Age Structurein Science."Pp. 292-356 in A Sociologyof Age
edited by Matilda White Riley, MarilynJohnson,and Anne Foner. Russell
Stratification,
Sage Foundation.


1995 - The Thomas Theorem and The Matthew Effect

Caricato da

Informazioni sul documento

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

1995 - The Thomas Theorem and The Matthew Effect

Caricato da

Copyright:

Formati disponibili

Social Forces, University of North Carolina Press

The Thomas Theorem and the Matthew Effect

This content downloaded from 79.175.121.210 on Mon, 27 May 2013 17:41:43 PM

I am indebted,once again, to HarrietZuckernan,Robert C Merton,CynthiaFuchs

Social Forces, December1995, 74(2):379-424

This content downloaded from 79.175.121.210 on Mon, 27 May 2013 17:41:43 PM

380 / Social Forces 74:2,December1995

3 As I have noted elsewhere (Merton1984:282),the designationThomas theorem"does not,

This content downloaded from 79.175.121.210 on Mon, 27 May 2013 17:41:43 PM

The ThomasTheoremand the MatthewEffect/ 381

5Forthe paradigmwhich maintains that the institutionaldynamicsof science derives from

This content downloaded from 79.175.121.210 on Mon, 27 May 2013 17:41:43 PM

382 / Social Forces 74:2,December1995

7ZAnticipationsa referto earlierideas, formulationsor findingsthat overlaplaterones but do

This content downloaded from 79.175.121.210 on Mon, 27 May 2013 17:41:43 PM

The ThomasTheoremand the MatthewEffect/ 383

9"Undevelopedideas" inasmuch as these earlier adumbrationsor anticipationswere not

This content downloaded from 79.175.121.210 on Mon, 27 May 2013 17:41:43 PM

384 / Social Forces 74:2, December 1995

10This statementwill surely not be taken to detractfrom Znaniecki'smajorcontributionsto

11Onthe phenomenonof unintended genderingand de-genderingof language, see Merton,

This content downloaded from 79.175.121.210 on Mon, 27 May 2013 17:41:43 PM

The Thomas Theorem and the Matthew Effect / 385

This content downloaded from 79.175.121.210 on Mon, 27 May 2013 17:41:43 PM

386 / Social Forces 74:2, December 1995

This content downloaded from 79.175.121.210 on Mon, 27 May 2013 17:41:43 PM

The ThomasTheoremand the MatthewEffect/ 387

To begin with, it will be noticed that SSS matter-of-factlyadopts the eponym

This content downloaded from 79.175.121.210 on Mon, 27 May 2013 17:41:43 PM

388 / Social Forces 74:2,December1995

This content downloaded from 79.175.121.210 on Mon, 27 May 2013 17:41:43 PM

The ThomasTheoremand the MatthewEffect/ 389

This content downloaded from 79.175.121.210 on Mon, 27 May 2013 17:41:43 PM

390 / Social Forces 74:2,December1995

This content downloaded from 79.175.121.210 on Mon, 27 May 2013 17:41:43 PM

The ThomasTheoremand the MatthewEffect/ 391

Of course, this could be condensed - or extended!

This content downloaded from 79.175.121.210 on Mon, 27 May 2013 17:41:43 PM

392 / Social Forces 74:2,December1995

ly unpublishedarchivalevidenceaboutthe respectiveroles of W.I.and Dorothy

This content downloaded from 79.175.121.210 on Mon, 27 May 2013 17:41:43 PM

The ThomasTheoremand the MatthewEffect/ 393

This content downloaded from 79.175.121.210 on Mon, 27 May 2013 17:41:43 PM

394 / Social Forces 74:2, December 1995

This content downloaded from 79.175.121.210 on Mon, 27 May 2013 17:41:43 PM

The Tomas Theorem and the Matthew Effect / 395

Youwill noticeat once the emphaticthoughprobablyunintendedway in which,the

This content downloaded from 79.175.121.210 on Mon, 27 May 2013 17:41:43 PM

396 / Social Forces 74:2, December 1995

ascribethe theoremto W.Laloneand absolvethe

This content downloaded from 79.175.121.210 on Mon, 27 May 2013 17:41:43 PM

The ThomasTheoremand the MatthewEffect/ 397

Exhibits4 and 5 fIhich bearrenewedwitnessto an longtimeeffortto deterthe

This content downloaded from 79.175.121.210 on Mon, 27 May 2013 17:41:43 PM

398 / Social Forces 74:2,December1995

Backto our immediatesubject,I also reportthat PartV of the same volume,THE

23 Earlyscholarlystudies of the reward system of science include Gaston (1971),Hagstrom

This content downloaded from 79.175.121.210 on Mon, 27 May 2013 17:41:43 PM

The ThomasTheoremand tiheMatthewEffect/ 399

Exhibit5 RWichreportsbutdoesnot documentthefact thatDorothyThomasherself

This content downloaded from 79.175.121.210 on Mon, 27 May 2013 17:41:43 PM

400 / Social Forces 74:2, December 1995

27R.S.Smith (1993:4-5) found that threeof the 40 textbooksreferringto the Thomastheorem