Comments on "What Theory is Not" Author(s): Paul J. DiMaggio Source: Administrative Science Quarterly, Vol. 40, No. 3 (Sep., 1995), pp. 391-397 Published by: Johnson Graduate School of Management, Cornell University Stable URL: http://www.jstor.org/stable/2393790 Accessed: 01/09/2009 03:24 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=cjohn. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].

Johnson Graduate School of Management, Cornell University is collaborating with JSTOR to digitize, preserve and extend access to Administrative Science Quarterly.

http://www.jstor.org

Comments on "What Theory is Not" Paul J. DiMaggio PrincetonUniversity

Sutton and Staw displayan unerringeye for the dodges we authors use as substitutes for theory. But these ruses are easier to acknowledge than to do without. Why is this the case? Sutton and Staw suggest that the problem lies in a combinationof education (social science faculty don't train our students adequatelyin theory construction)and talent (not enough of us have that ineffable something that makes a good theorist).Withoutdenying the importanceof these factors, especially the first, I would suggest that three additionalissues renderthe problemof theory even more complicatedthan Sutton and Staw suggest. 1. There is More Than One Kind of Good Theory There are at least three views of what theory should be, and each of them has some validity.Each of them also has limitations.

? 1995 by CornellUniversity. 0001-8392/95/4003-0391/$1.00.

Theory as covering laws. A familiarposition, which Sutton and Staw implicitlyreject, is that theories should consist of covering laws: generalizationsthat, taken together, describe the world as we see (or measure) it. The fact that most social scientists to some extent embrace this approach renders Sutton and Staw's argument more radicalth, i it sounds, for they reject some key tenets of behavioral science as it is usuallypracticed:a focus on explaining varianceratherthan regularities;the view of scientific progress as a kindof R2 sweepstakes; and the image of a world in which variablesexplainone another-all parts of the perspective that Abbott (1988) has derided as"ordinarylinear reality."At the limit-a limitreached by economists who admit that they don't care if their assumptions are implausible,so long as their R2s are high (Friedman, 1953)-this view providesthe "what" of theory that Sutton and Staw argue is insufficientunless accompanied by the how and why. Theory as enlightenment. A second view of theory, especially prominentin those neighborhoodsof the social sciences influencedby the humanities,is as a device of sudden enlightenment. Fromthis perspective theory is complex, defamiliarizing,richin paradox.Theorists enlighten not throughconceptual clarity(a postmodernist once told me that to define what she meant by "postmodernism"would be unfaithfulto the theory), but, like R. Crumb'sZen master Mr. Natural,by startlingthe readerinto satori.The point of theory, in this view, is not to generalize, because many generalizationsare widely known and ratherdull. Instead, theory is a "surprisemachine" (Gouldner's,1970, unflatteringassessment of Parsons' system), a set of categories and domainassumptions aimed at clearingaway conventionalnotions to make room for artfuland exciting insights. Theory as narrative. A thirdperspective on theory emphasizes narrativity:theory as an account of a social process, with emphasis on empiricaltests of the plausibility of the narrativeas well as carefulattentionto the scope conditionsof the account. The minimalistversion of this approach(Collins,1981, on "micro-translation") simply requiresthat hypotheses detailingregularitiesin relations among variablesbe accompanied by plausibleaccounts of 391/AdministrativeScience Quarterly,40 (1995): 391-397

how the actions of real humans could produce the associations predictedand observed. More assertive versions argue that theory begins with baseline generator, models-formal models of human behaviorthat specify principlesof individualor group action that throughcomputer simulationgenerate observed distributionsof outcomes (Fararo,1989). Sutton and Staw clearlyhave in mind a temperate version of the narrativeapproachwhen they speak of theory. Although I share their bias, I thinkwe must also make room for the other versions, as well as for hybrids(e.g., Cohen, March, and Olsen, 1972, which combined narrativemodellingwith Zen-likeparadox).And if we admit other approaches, we must tolerate their limitations.As Sutton and Staw point out, for example, variancetheories can inch toward process theories, as when they treat sequences as dependent variablesor specify scope conditions in ways that call attentionto interactioneffects that illuminateprocess. But, ultimately,from the covering-lawperspective, the point of theory is to explainthings, and explanationmeans accountingfor variance: Inthis view, the distance between hypothesis and theory is vanishinglysmall, and if you need a lot of hypotheses to explaina lot of variance,then so be it. Similarly,because enlightenmenttheories are often intuitive, they may employ references or diagramsor graphic presentations of data as rhetoricaldevices to elicit epiphanies.

2. Good Theory Splits the Difference One can go beyond simply recognizingthe diversityof useful and plausibleapproachesto suggest that many of the best theories are hybrids,combiningthe best qualitiesof covering-law,enlightenment,and process approaches. One reason that theory constructionis so difficultto teach is that these approaches, as we have seen, are drivenby different purposes and embody differentvalues. Consequently,the researcherwho tries to combine them faces not a list of brightlinestandards,but a set of vexing choices. Clarityvs. defamiliarization. By defamiliarization,I refer to the process of enablinga native-of a society, an organization,or an academic discipline-to see his or her world with new eyes. Arguably,good theory should accomplish this. But it must not go too far. The conventional justificationfor neologisms is that the old words carrytoo much baggage to convey new ideas or perspectives. At the same time, too many neologisms rendera theory too strange for people to grasp. Similarly,it is often necessary to frame a theory in paradoxicalterms in orderto get readers to pay attention.Arguably,all good theory has a germ of paradox.Some of the best theory has no more than a germ: Hannanand Freeman's (1977) originalpaper on ecology was powerful because it effected a tiny but crucialshift in the reader'sfocus from change in survivingorganizationsto patterns of birthand death. Other exemplarytheories-e.g., Cohen, March,and Olsen's (1972) exposition of the garbage-canmodel-are awash in paradox.Too much paradox,however, and an interestingnew theory begins to sound preposterous.The trickis in the balance. 392/ASQ, September 1995

ASQ Forum/DiMaggio

Focus vs. multidimensionality. Most graduateprograms highlightthe importanceof focus in theory: Take a strong position or a new model and push it as far as it will go. Sutton and Staw endorse this when they argue that hypotheses, especially disparatehypotheses drawn from differenttheoreticaltraditions,do not constitute theory. Graduateprogramsalso highlightexegesis and teach students to pick aparta paper or study for the factors or variablesit omits. Some theorists even view "multidimensionality"-theextent to which a theory includes reference to agency, culture,structure,and several other abstract categories in its rhetoric-as a decisive criterionof its adequacy (Alexander,1982). Alas, one person's multidimensionalityis another's goulash; one author'sfocus, another's crude reductionism.Again, I side with Sutton and Staw in their general orientationtoward what one might call "strategic reduction":abstractingaway enough of the world's confusion to develop pointed explanationsof organizationalphenomena. But where one draws the line is still more art than science. Comprehensiveness vs. memorability. Theories that are both enlighteningand focused tend to emphasize processes and associations that many readers find surprising.We are rewardedfor derivinglogicaldeductions from theoreticalfirst principlesthat generate surprisingpredictionslinking domains that are often considered separate; for example, demonstratingthat people who receive little autonomy on the job give their kids littleautonomy in the home (Kohnand Schooler, 1978). The trouble is that the most interesting causal factors are often not the most important.A few years after graduateschool, I gave a talk at the Universityof South Carolinaon the effects of network position on certain organizationaloutcomes. Duringthe question period,the late Bruce Mayhew asked why I had spent forty of my forty-five minutes talkingabout network measures, when organizationalsize explainedtwice as much variancein my dependent variables.If I reallycared about the outcomes I was tryingto explain,why hadn't I focused on size? In formulatingan answer, I realizedthat I had never thought of size as interesting: Didn'teveryone know that size would influence many measures of organizationalpersistence and effectiveness? Why talk about what everyone knew? At the same time, if our job is to explainthe world, ratherthan to note small but paradoxicalstatistical relationships,shouldn't we focus precisely on the measures and processes that explainthe most? Ourcollective preoccupationwith theoreticalnovelty often leads organizationalresearchers to overlookcrucialif banalpatterns in their data (sometimes even omitting "dull"variablesat the cost of misspecifying statistical models). Once again, one must find the point on the tightropeat which balance can be achieved. 3. Theory Construction is Social Construction, Often after the Fact Even if one constructs a careful hybridtheoreticalstrategy and finds the properbalance between the conflictingvalues theory may embody, that theory's fate will be that goodte determined in partby factors outside one's control. If the 393/ASQ, September 1995

productionof good theory requiresthe utmost care, theory's reception is ordinarilyhelter-skelter:a process of appropriationdrivenmore by resonance than by reason, in which complex arguments are reduced to slogans and relatedto one another along binarydimensions more redolent of Levi-Strauss'stribalcultures than of graduate theory classes. And not only is theory created by its readers as well as its writers-it is then recreated by the authors who employ it. The value of resonance. Here is an hypothesis (not a theory):The reception of a theory is shaped by the extent to which a theory resonates with the culturalpresuppositions of the time and of the scientific audience that consumes it. An example helps explainthis. Before going to watch sea lions mate on a centralCalifornianbeach several years ago, I read an account of the importanceof this beach as a site of contests between males for dominance. The winning males, so the story went, were rewardedwith the affection of female sea lions-and with it the opportunityto pass their genes to new generations, as natureselected for sea lion machismo. Shortlythereafter I met a woman biologistwho had been studyingthe sea lions and other species famed (among male biologists)for their dominance contests. Accordingto this scientist, the beaches attracteda large but finite proportionof the sea lions. While the big bulls blustered at one anotheron the beach, a sizable minorityof sneaky little male and female sea lions frolickedabout the outer islands, breeding happily,beyond the view of much of the scientific world. At first, such critics received little attention. But as feminist theory (e.g:, Haraway,1989) made scientists more sensitive to the ways in which humanculture influences how biologists understandand portraythe naturalworld, their voices penetrated scientific debate, enrichingtheories of biologicalevolution.The change was not so much in the theories availableto biologists, much less in the social lives of sea lions, as in the intellectualenvironment-an environmentshaped by the humanitiesand social sciences-into which evolutionaryarguments were released. I suspect that the same thing happens in organizational science, as culturalchange enhances or corrodes our capacityto see aspects of the organizationswe study, by limitingthe metaphorswe thinkwith. If one observes the progress of organizationtheory from the 1950s throughthe 1990s, it is intriguingto consider the relativeimpact on our theories of organizationalchange, on the one hand, and changes in broaderpreoccupationsand culturalrepertoires, on the other. To what extent did the appeal of political theories of bureaucracy(Cyertand March,1963; Allison, 1969) reflect the inchoate decline of politicalorthodoxyin that decade? To what extent did the popularityin the U.S. of theories that emphasized limitsto rationalcontrolof organizations(garbage-can,resource-dependence, ecological, institutionaltheories) in the 1970s bear an affinityto doubts and fears associated with a decline in America's international hegemony? To what extent does the popularityof realist depictions of organizationsas dense spots in multiple networks of relations(the so-called network organization 394/ASQ, September 1995

ASQ Forum/DiMaggio

debated in Nohriaand Eccles, 1992), ratherthan highly bounded bureaucracies,reflect actual organizationalchange, and to what extent does it mirrorthe antinomiantemper of the 1990s and the widespread feeling that every surface unity masks a more complex underlyingstructure? Theories into slogans. People read quickly.Unless their teaching or research leads them to attend to a paper or book with special care, they will pick from the field of ideas in any theoreticalwork those that resonate with preexisting expectations and assumptions and forget the rest. In many cases, they will furthersimplifythe ideas they retainuntil those ideas fit neatly into preexistingschemas (Fleck, 1979; D'Andrade,1995). The more widely a theoreticalpaper or book is read, the greater the proportionof readers who are not specialists in the subject matter it addresses. The greater the proportionof nonexpert readers, the greater the extent to which its reception is determined by a cognitive field differentthan that of its authors, and the greater the extent to which its arguments are refashionedand simplified. I became familiarwith this process when a paper that Woody Powell and I coauthored(DiMaggioand Powell, 1983), "The IronCage Revisited,"came to be assigned widely to graduateclasses in organizationtheory and sociology. Withinthe field of institutionaltheory it represented a relativelymaterialistvariant,emphasizing,as it networks in driving did, the role of interorganizational processes of imitationand diffusionthat tended to make organizationssimilarto one another. Somewhat to my surprise, I began receivingpapers that cited our paper as supportfor the propositionthat all organizationsbecome like all others, regardless of field. Somehow the network argumentthat we authors regardedas so central had been deleted in the paper's reception.Withina few more years, the paper had turned into a kindof ritualcitation,affirming the view that, well, organizationsare kindof wacky, and (despite the presence of "collective rationality"in the paper's subtitle)people are never rational. Relatedto this is a tendency for the field to classify theories on the basis of primordialantimonies ratherthan coherent and multidimensionalanalyticcategories. Forexample, students (and sometimes non-studentswho should know better) often dividethe world of organizationtheories between organizationalculture,garbage-can,institutional, and loose-couplingapproaches, on the one hand, and transactioncost, technicalfunctionalist,populationecology, and agency theory approaches, on the other. This classificationseems to owe more to intuitivenotions of "hard"and "soft" than to analyticpositions on the role of environments,open vs. closed systems, rationalaction, or other factors that providethe basis for systematic multidimensionalmappings. Post hoc theory construction. Theories are not just constructed, they are socially constructed after they are written. Theoreticalideas take on a life of their own. In some cases, sophisticated ideas are degraded. In other cases, half-bakedideas go back into the oven, coming out in 395/ASQ, September 1995

more satisfactoryform. To some extent, the qualityof a theory is a functionof the qualityof the people who employ it. In draftinga recent paper on polarization,I found that the most useful papers and books on the topic all cited Georg Simmel's (1955) famous essay on conflict, so I returnedto the original,which I had not read for many years. Much as I admire Simmel, I did not find the ideas that claim his paternityin very crisp form. My conclusion is that if your theory happens to be taken up by the likes of Lewis Coser (1956), Ron Breiger(1974), and Peter Blau (1977), your reputationis in very good hands. The supreme example of this phenomenon in organizationtheory must be Chester Barnard(1938). I doubt that The Functionof the Executive satisfies any formalcriterionof good theory construction. But for whatever reasons, it inspiredHerbertSimon, who ignoredall that was quaintand vacuous in the book and joined with James Marchto turn Barnard'sconvoluted prose into a crystallineand compellingpropositionalinventory (Simon, 1957; Marchand Simon, 1958). Fromthis promising beginning,the book's fame grew, untilseveral years ago, a belated festschrift featured the most diverse group of organizationaltheorists imaginable,each paying homage to Barnardand his work (Williamson,1990). One could point to formalqualitiesthat made Functiona plausiblecandidatefor such treatment: Barnard'sprose is often ambiguous and the book's argument is undisciplined,both of which mean that a contemporarytheorist seeking some sign that Barnard anticipatedone of her or his best ideas has a target as wide as an aircraftcarrier.But one would not recommend that the aspiringtheorist emulate these qualities. CONCLUSION I have suggested two modest revisions to Sutton and Staw's argument. First,good theory is so difficultto produce routinely,in part, because "goodness" is multidimensional: The best theory often combines approaches to theorizing, and the act of combinationrequirescompromise between competing and mutuallyincompatiblevalues. Second, theory constructionis a cooperativeventure between authorand readers: Theoryreception rides on much more than scientific potential;in the short run,we tend to reduce theories to slogans; and in the long run,brilliantexpositors can turn muddledtheories into canonicalmasterpieces. If the first set of points, on tensions within theories, highlights the need for theorists to exercise judgment and pluck,the second suggests the importanceof environmentand luck. Let me be quite clear, however, that these remarksare meant to qualifySutton and Staw's account, not to question their main lines of argumentor the usefulness of their superb descriptionof non-theory.Any readerswho find in my qualificationswarrantto disobey Sutton and Staw's injunctionsdeserve whatever the reviewers deal them.

396/ASQ, September 1995

ASQ Forum/DiMaggio REFERENCES Abbott, Andrew 1988 "Transcendingordinarylinear reality."SociologicalTheory, 6: 189-196. Alexander, Jeffrey 1982 TheoreticalLogicin Sociology, vol. I: Positivism, Presuppositions,and Current Controversies.Berkeley: Universityof CaliforniaPress. Allison, Graham T. 1969 "Conceptualmodels and the CubanMissile Crisis." AmericanPoliticalScience Review, 63: 689-718. Barnard,Chester 1938 The Functionof the Executive. Cambridge,MA: Harvard UniversityPress. Blau, Peter 1977 Inequalityand Heterogeneity. New York:Free Press. Breiger, Ronald L. 1974 "Thedualityof persons and groups." Social Forces, 53: 181-190. Cohen, Michael, James G. March, and Johan P. Olsen 1972 "A garbage-canmodel of organizationalchoice." AdministrativeScience Quarterly,17: 1-25. Collins, Randall 1981 "On the microfoundationsof macrosociology."American Journalof Sociology, 86: 984-1014.Coser, Lewis 1956 The Functionsof Social Conflict.Glencoe, IL:Free Press.

Cyert, Richard, and James G. March 1963 A Behavioral Theory of the Firm. Englewood Cliffs, NJ: Prentice-Hall. D'Andrade, Roy 1995 The Development of Cognitive Anthropology. New York: Cambridge University Press. DiMaggio, Paul, and Walter W. Powell 1983 "The iron cage revisited: Institutional isomorphism and collective rationality in organizational fields." American Sociological Review, 48: 147-160. Fararo, Thomas 1989 The Meaning of General Theoretical Sociology: Tradition and Formalization. New York: Cambridge .University Press. Fleck, Ludwik 1979 Genesis and Development of a Scientific Fact. Thaddeus J. Trenn and Robert K. Merton, eds., Fred Bradley and Thaddeus J. Trenn, trans. Chicago: University of Chicago Press. Friedman, Milton 1953 Essays in Positive Economics. Chicago: University of Chicago Press. Gouldner, Alvin 1970 The Coming Crisis of Western Sociology. New York: Basic Books.

397/ASQ, September 1995

Hannan, Michael, and John Freeman 1977 "Thepopulationecology of organizations."American Journalof Sociology,82: 929-964. Haraway, Donna 1989 PrimateVisions: Gender, Race, and Naturein the World of ModernScience. New York:Routledge. Kohn, Melvin, and Carmi Schooler 1978 "The reciprocaleffects of the substantivecomplexityof work and intellectual flexibility:A longitudinal assessment." American Journalof Sociology,84: 24-52. March,James G., and Herbert Simon 1958 Organizations.New York: Wiley. Nohria, Nitin, and Robert Eccles (eds.) 1992 Networksand Organizations: Structure,Form,and Action. Boston: HarvardBusiness School Press. Simmel, Georg 1955 Conflict.KurtH. Wolff and ReinhardBendix,trans. (Originally publishedin 1908.) New York:Free Press. Simon, HerbertA. 1957 AdministrativeBehavior,2nd ed. New York:Macmillan. Williamson, Oliver (ed.) 1990 Organization Theory:From Chester Barnardto the Present and Beyond. New York:OxfordUniversityPress.