institutional theory and goffman’s dilemma


Recently I sat down to re-read Pamela Tolbert and Lynne Zucker’s (1996) insightful essay “The Institutionalization of Institutional Theory” (in this collection; a similar argument is made in this paper). I discovered two things: 1) this is a much better paper than I remembered reading years ago in grad school and 2) they make a point that I think has gone under-noticed in recent discussions of theoretical problems and prospects in institutional theory, but which I think deserves a much wider airing.

The problem begins with what they note is an “ambiguity” in the conceptualization of institutionalization, that emerges out of the classic Meyer and Rowan (1977) piece, and which is a creature of their creative yet incomplete adaptation of the notion of loose coupling (which in the original Weick formulation simply referred to a state of disconnection [or lack of integration] between the parts or components of an organization) to explain the alleged lack of tight connection between an organizations formal structure and its everyday (“technical”) activities in the context of embeddedness in an “institutionalized environment.” Notice the sleight of hand: from disconnected parts to a disconnection between institutionally mandated form and behavior. Not the same thing.

This is crucial according to Tolbert and Zucker, since it allows Meyer and Rowan to make use of the cognate notions of institution and institutionalization in two different senses, which carry wildly divergent empirical referents and theoretical presupositions. The first sense is the traditional “phenomenological”(from its formulation by Berger and Luckmann)/Durkheimian notion of an institution as a taken-for-granted, exteriorized and “typified” patterns of habitual action and cognition, which exerts a constraining force on behavior patterns and which is in fact the primary explanation for the actor’s behavior. The other is the notion of an institution as an externally mandated “facade,” which actors adopt for purposes of legitimacy and/or external mimicry but which carries little cognitive legitimacy and empirical weight in explaining the behavior of actors since (by definition) it is decoupled from their everyday activities.

Tolbert and Zucker note that institutional theorists cannot have it both ways: either institutions are seen as an important element in the explanation of an actor’s behaviors and thus externally constraining in the classic Durkheimian sense (the Berger-Luckmann conception), or it is seen as an externally imposed “frontstage” that actors adopt in order to secure important environmental goodies (resources, legitimacy, etc.), but which has little bearing on their everyday patterns of activity. What is more, if the loose-coupling idea is taken seriously as the impetus for a research program, then institutional theory collapses (in terms of empirical predictions) into the Pfeffer-Salancik resource dependency theory: organizational actors are seens as rational calculators adopting externally legitimated environmental accouterments as a way of getting the resources that they need from their environment.

From this viewpoint, the uniqueness of the institutionalist paradigm (its attention to culture as constitutive of actors in a strong sense) is lost:

The lack of theoretical distinctiveness in these studies results in part from the de-emphasis on a distinguishing feature of institutional theory, a focus on the role of cultural understanding as determinants of behavior..and on the normative bounds of rational decision-making. By shifting toward and emphasis on changes in ‘appearance’ and downplaying the internal consequences of institutionalized structure, treating structure as merely symbol and signal, we end up with the implicit argument that a structure can maintain its symbolic value in the face of widespread knowledge that its effect on individuals’ behavior is negligible. How such a contradiction in cultural understandings (i.e. that structures signify commitment to some action, and that structures may be unrelated to action) can endure poses and unanswered riddle in this approach.

It may surprise some that the “new” institutional approach inspired by the Meyer-Rowan paper (touted as subverting traditional notions of rationality and means-ends conceptions of behavior ["ritual and ceremony"]) ends up coming so close to its alleged rational-choice opposite. However, this may have something to do with a larger meta-theoretical issue, which revolves around the conceptual work that notions such as “loose-coupling” do in the theory. It turns out that Erving Goffman, faced a similar theoretical dilemma throughout his career, as could only be noted by Randall Collins in his essay “The Three Stages of Erving Goffman” in an old out of print book called Sociology since Mid-Century. In that essay, Collins noted how Goffman began to develop his microsociology as a straightforward application of Durkheimian ritual theory, where individuals performances to sustain “face” and to manipulate their micro-environment to project and “acceptable” self and to protect others from failing to do the same, were seen as motivated by an inherent desire to sustain the moral order of society. However very early on, the entire conceptual armamentarium that Goffman developed (frontstage/backstage, saving face, interaction as dramaturgy, etc.) was seen as fairly compatible with “cynical” and rationalistic manipulation of appearances to secure advantages in interaction.

Thus Goffman’s invention of the notion of “role-distance” can be seen as strictly analogous to the Meyer-Rowan notion of loose coupling. Not surprisingly, Collins notes, after the “Durkhemian” period of the early and mid 1960s, where social actors were seen as motivated by a primarily moral commitment to the interaction order, Goffman enters a “rational choice” stage in which he begins to take seriously the notion of micro-interaction and self presentation as a game (in the technical sense as the references to Schelling attest) under limited information conditions. Thus, with the introduction of “role-distance” (an external “social” imposition that does not bind the actor to behave in specified ways) a space is opened up for the original Durkhemian inspired theory of micro-interaction to “devolve” into a rational action oriented formulation.

Is there a lesson for institutional theory here?

About these ads

Written by Omar

September 25, 2006 at 7:46 pm

Posted in just theory, omar, sociology

19 Responses

Subscribe to comments with RSS.

  1. There is a third way that institution is used: as a post hoc rationalization/justification for organizational activities. For example, in one of Dobbin and Sutton’s papers on the HR movement, they describe how human relations divisions created corporate policies to deal with equal employment opportunity laws, and later justified those policies according to a dominant institutional logic – market efficiency. In this third way, institutions are not loosely-coupled but they don’t exactly constrain behavior either, as they become forceful (or meaningful anyway) after a behavior has already become systemic.


    September 25, 2006 at 8:30 pm

  2. Omar wrote: “Thus, with the introduction of “role-distance” (an external “social” imposition that does not bind the actor to behave in specified ways) a space is opened up for the original Durkhemian inspired theory of micro-interaction to “devolve” into a rational action oriented formulation.”

    I have always thought the dichotomy tween interactionists and rational choice is false. Melding into RC is not devolving, but rather a recognition that once the rules of the game are set, then certain behaviors are probably the equilibrium of the game defined by the rules.

    My opinion about the division of labor between cultural theories and rational choice theories is that institutional/cultural arguments seem to address how the rules are made and RC theories defines what happens after you define roles, rules and other aspects of the game. The RC imperialists would deny or simply ignore how the rules get made and the origins of identities and interests that create the game. Strangely, the interactionists would have us ignore the simple fact that rules or institutions emerging from interactions create incentives, which are amenable to economics analysis.

    This is just a long way of saying that yes, there are formulations of institutional theory and other interactionist theories that allow for a natural integration with rational choice theory. (Anyone wanna write it up and submit to R&S or Soc Theory? :) )

    Fabio Rojas

    September 25, 2006 at 11:15 pm

  3. Brayden.- Good point. You often see the paper on “Accounts” by Scott and Lyman (1968) cited by Meyer and others, as well as the C. Wright Mills paper on “Vocabularies of Motive” (1940). This makes me think that Tolbert and Zucker have an even better point than they thought; while you can think of shared post hoc rationalization schemes as “institutions,” this seems to be equally incompatible with the “institutions in the strong sense” phenomenological usage of the term (and I think the correct one), and equally vulnerable to the “impression management” critique (actors cynically draw on institutionalized justifications for their actions in order to secure the inflow of resources, legitimacy, etc.).
    Fabio.- You are absolutely correct. What worries me (and I think what worries Tolbert and Zucker) is that in light of the agreed upon fact that institutional theory does not have a well worked out model of “how the rules are made” (institutional genesis) then, at the level of empirics rational choice theory and institutional theory are equivalent (maybe Teppo would love this). In other words, for any institutional formulation that explains adoption or conformity based on strong “constitutive” accounts pointing to culture, schemas, scripts, etc. an empirically equivalent account of an organizational actor cynically manipulating his or her image in order to secure environmental resources could in principle be constructed (even if we allow that at the beginning of the process adoption is guided by technical reasons and at the end for “legitimacy” reasons; insofar as legitimacy is tied to resources, then the latter choice is “rational”).


    September 26, 2006 at 12:16 am

  4. Excellent post and great discussion!
    The problem I see with institutional theory is that the origins and change of institutions never get addressed – generally the theory is about taken-for-grantedness, persistence and diffusion (from higher levels), rather than the more interesting origins and change (where you get into specific mechanisms etc).
    So, briefly to Fabio’s point – I don’t then agree that institutional theory addresses “how rules are made” (and RC the subsequent operation of rules) – institutional theory assumes the pre-existence of institutions rather than explaining their origin. RC approaches seem to do a much better job at explaining origins (e.g., public choice perspectives in particular). So, the argument is reversed. You might be thinking specifically about game theoretic conceptualizations in terms of RC explaining how rules play out, but that is only a very narrow slice of the broader approach.


    September 26, 2006 at 5:07 am

  5. A few thoughts, as I make my way through your post:

    I think you commit two mistakes, possibly not for institutional theory as written but perhaps as it should have been written – although if you go back and read M&R again, its a more subtle arg than you might think.

    The first is perhaps kind of ticky-tack, but maybe not. It is that there is a notion, filtered through the new instit in economics associated with North, et al, to think about institutions as contraints. That is, agents would engage in all sorts of action if they could, but institutions provide rules of the game that make somet things possible and others impossible. But institutions enable as much as constrain – they provide both structure and cognitive tools to do so. That institutions operate at the supra-individual level doesn’t mean that they aren’t experienced as deeply individual. In this sense, I would think about gender as an institution, not just labor law or something like that.

    More problematic, I think, is to think about institutions as facades, behind which the ‘real’ technical stuff happens. This is, I think equally problematic in thinking about Goffman. In Presentation of Self, Goffman is pretty straight about it (though it’s interesting that we don’t believe him here):

    “In this report the performed self was seen as some kind of image…While this image is entertained concerning the individual, so that a self is imputed to him, this self itself does not derive from its possessor…A correctly stated and performed scene leads the audience to impute a self to a performed character, but this imputation – this self – is a product of a scene that comes off, and is not a cause of it. The self…is not an organic thing that has a specific location, whose fundamental fate is to be born, to mature, and to die; it is a dramatic effect arising diffusely from a scene that is presented…” (252-253).

    Same with M&R, I would say. Empirically, from my own research, I find that financial analysts think art specialists effectively make shit up in order to determine art prices, telling me with a straight face that their analyses are based on economic fundamentals. That the former is seen as a facade masking a status order while the latter is seen as technical, means-end decision-making does not necessarily make it so. For me, separating out the ‘real’ from the ‘ceremonial’ is not the project, understanding how institutional environments make ontology of price knowable is…

    Of course, RC is not particularly compelling to me, and I was born and raised on cultural institutionalisms of varied stripes, so this may be more than a little self-serving.


    September 26, 2006 at 12:57 pm

  6. I don’t think M & R are as subtle about the technical versus institutional dichotomy as you make them out to be, since they pretty much take that for granted. In fact, the entire paper is premised on the fact that one of the “scope conditions” for the theory that they present is that it only applies to “institutionalized organizations,” (and not purely “technical” organizations) where the quality of technical outputs is hard to judge and where connections between technical activity and results is ambiguous (thus, the theory does not apply to toilet paper manufacturers). It was only later as noted by Scott in his “adolescence of institutional theory” paper in ASQ and with the later work of Fligstein, Dobbin, Sutton, Edelman, etc. (and the rise of economic sociology) that the entire technical/institutional dichotomy crumbled, when it was shown that “technical” activities were also subject to “institutional effects.” Although an argument can be made that the seedlings for this development where already in the original M & R paper, that might be a bit of a “rational reconstruction” of the history of institutional theory.

    Did not mean to give new life to the entire notion that Goffman’s dramaturgy implies a “pre-constructed” self that deviously manipulates appearances. Collins is adamant that this is a clear mis-representation of Goffman’s early neo-Durkheimian project (in which the self was a product of the interaction order [the stand-in for society with a big “S”) as you show in the quote from the Presentation of Self. However, Collins’ interesting point is that this was not the end of Goffman’s intellectual development and that toward the middle part of his career (before “Frame Analysis”) he did seriously experiment with Schellingian strategic interaction theory, which actually moved him closer to the “RC” position (with really big scare quotes). However in the end, Collins argues that Goffman abandoned both his early Durkhemian theory (the self as a “sacred” product of the modern moral order) and the mid-career flirtation with neorationalist strategic theory and developed a “Godelian” self-referential Durkhemianism with his multilevel theory of frames. However this last development has never been digested by mainstream sociologists (who continue to think of Goffman as a “symbolic interactionist”).


    September 26, 2006 at 3:09 pm

  7. “Godelian” self-referential Durkhemianism



    September 26, 2006 at 7:21 pm

  8. Ok, Omar, you assert a substantial break between Presentation of the Self (PoS) and Frame Analysis (FA). How real is this distinction? My reading of these texts is usually that PoS is a very egocentric text that talks about individual manipulation of situation. It’s very much a text about signalling identity and waiting for people to respond.

    In contrast, FA, as you say, deals with multilevel analysis of situations (which by the way is not “Godelian” – just my pet peeve: inaccurate appropriations of math by social scientists). FA is about how people negotiate the attribution of meaning and the acceptability of utterance and behavior.

    These two activities (signalling in a context vs. negotiating meaning) seem to complimentary. If your history of Goffman is correct (which I assume it is), that what really happens is:

    a. Goffman reads Durkheim and wants to explore how ritual is built into daily life, which leads to:

    b. (PoS) The idea that individuals construct faces and identities through daily interaction, which leads to:

    c. The idea that there can be advantageous signalling/construction of identities in daily settings – e.g., the “keeping the line” concept (the quasirational approcah), which is rejected and he moves to:

    d. (FA) The exploration of how meaning is attributed to objects and people via accepted semantic schemes.

    So in my reading, Goffman seems to be expanding on the idea that interaction is ritual and that rituals can be used to advance individual interests as well as be an expression, or a method for creating, social identity. Then you can go micro and explore how rituals work for individuals (PoS, the Schelling approach) or go slightly “macro” and explore the preconditions for rituals (FA).

    Maybe from Goffman’s perpective this seems like radical jumps in thinking, but if you take a step back, I think it’s all of the same cloth.

    Fabio Rojas

    September 26, 2006 at 10:13 pm

  9. Fabio .- This is a great rational reconstruction of Goffman! It makes more sense than Collins’, which is much more of a descriptive stage analysis which gives up on coherence. Maybe that’s the paper that should go to Soc Theory…


    September 26, 2006 at 11:55 pm

  10. Fabio:

    I have always thought the dichotomy tween interactionists and rational choice is false. Melding into RC is not devolving, but rather a recognition that once the rules of the game are set, then certain behaviors are probably the equilibrium of the game defined by the rules.

    Neither strong Rat-Choice people nor strong (i.e. ‘phenomenological’) institutionalists will be satisfied by this parceling out of tasks. Think about the question of institutional design or reorganization. For the RC crowd there will be some rational way to do it given what we want to accomplish, and they will say they can bootstrap a design — or, more ambitiously, explain present arrangements that were not consciously designed — by asking how rational agents would respond to various rules or incentives. By contrast, the strong institutionalists will want to show that all the rationality talk is just ritualized practice — i.e. that it has no traction on its own terms.


    September 27, 2006 at 5:24 am

  11. With reference to Kieran’s comment – the ‘division of labor’ approach to different (more often than not, opposing) theories, has always bothered me. Also, the reinterpretation or reconstruction via some other perspective of a theory (intended as something completely different) also seems problematic (see e.g., Elster’s [1985] RC reconstruction of Marx).


    September 27, 2006 at 5:51 am

  12. I don’t know that it’s as much a “division of labor” as it is a problem of irreconcilable views about human behavior.


    September 27, 2006 at 1:26 pm

  13. [...] institutional theory and goffman’s dilemma [...]

  14. Omar sez: “Maybe that’s the paper that should go to Soc Theory…”

    I think this post and the comments should be published unedited in a special issue of Soc Theory.

    Fabio Rojas

    September 27, 2006 at 5:46 pm

  15. So Goffman is essentially a micro-institutionalist. Interesting thought. It underplays the extent Goffman was inspired by Burke’s dramatism (that also inspired Mill’s paper on the vocabulary of motives) but it certainly explains Goffman’s over-socialized understanding of agency.

    Re Meyer & Rowans piece: I have always understood it as a functionalist interpretation of ceremonial behavior. Put another way, it is a way to rationalize apparent slack. Not sure that it breaks with Weick’s understanding of loose coupling. First, Weick is famously loose and vague when it suits his argument, and, second, he introduces loose coupling as a way of understanding slack as a resource rather than as waste.

    Dan Karreman

    October 3, 2006 at 10:59 am

  16. On another issue: the idea that “rational choice” is secondary to cultural framing is pretty unoriginal in anthropological circles. See also Bourdieu, Pierre for a twist where “rational choice” is the cultural framing.

    Dan Karreman

    October 3, 2006 at 11:17 am

  17. [...] drew on Schelling’s strategic model to analyze self-presentation in social interaction (see here)–noted that contextual “awareness status” “deviants” (their example [...]

  18. [...] is something that is formally equivalent to Tolbert and Zucker’s (1996; see previous posts by myself and Brayden) discussion of institutional and resource-dependence theories–any purported case [...]

  19. [...] and Zucker are clearly knocking close the reflexivity problem.  This is what I called “Goffman’s dilemma.”  The reason why institutional theory becomes indistinguishable from “resource [...]

Comments are closed.


Get every new post delivered to your Inbox.

Join 1,070 other followers

%d bloggers like this: