Review essay: The importance of being Erving—Erving Goffman, 1922 to 1982


  • This article originally appeared in Sociology of Health & Illness, 1983 vol 5, issue 3, pages 345–355. We are grateful to Sociology of Health & Illness for permission to reproduce the article in full here.

What then was Goffman's contribution? For medical sociology, the answer seems too obvious to need much examination. The author of Asylums (Goffman 1968a) and Stigma (Goffman 1968b) requires no obituary: the books themselves and the revolution in institutional care to which they contributed are sufficient monument. Add to these, “The Insanity of Place” (Goffman 1971) a classic if much neglected article, and one gets a brilliant career within the field. Yet these three pieces form only a small, somewhat atypical, part of his work. His major contribution to the area has yet to come. Medical sociology like almost every other branch of the discipline, has still to assimilate his study of the microstructure of interaction; the relationship of both individuals and society to that structure; and of the methods he used in that study.

This other work, or some of it at least, has not gone unnoticed. That he began his career with an article on social class (Goffman 1951), may come as a surprise but most students are likely to have heard of The Presentation of Self in Everyday Life (Goffman 1969) and to have read one of his easier pieces such as “Cooling the Mark Out” (Goffman 1952). Nevertheless, for all the ready fame that these enjoyed, their theoretical and methodological impact has been small. It is still too easy to dismiss Goffman 's main work as amusing, interesting but minor; as applying only to our own society and era, not to other places and times; as the product of a light essayist, not a scientist; as dealing with micro-trivia rather than macrostructure; or, most seriously of all, as fundamentally immoral, as taking a cynical, manipulative, and ultimately destructive view of humanity.

There is therefore a discrepancy between his reputation in various sub-disciplines and his overall status as a sociologist. Within fields such as medicine and deviance he has a clear and solid fame. Here even his radical critics have acknowledged the importance of his work (Sedgwick 1982).

The same can hardly be said for the rest of his output. What, then accounts for the differences? If you take him, as I do, to be one of the major figures in the history of the discipline, the discrepancy is puzzling. Being a fan, I find it hard to see why others are not. Perhaps, after all, that work is limited and unworthy of much serious attention?

There are several reasons for thinking otherwise, but before spelling these out, it may be useful to reflect on the rather odd nature of Goffman's main work. In neither its style nor its content does it fit the disciplinary norm, and many of the problems in its reception may be traced to its academic oddity.

That oddity has several components. First, it seems no accident that the two most readily accepted parts of his work, Asylums and Stigma, are the only parts which focus directly on contemporary “social problems.” Moreover, they do so from what is currently the conventional moral stance within the discipline. We can see at a glance that Goffman dislikes the way that mental patients can be treated, is sympathetic to the plight of the stigmatized, and fully aware of the manner in which the powerful may exploit the unfortunate. But in his other work, Goffman's moral stance is not quite so obvious. His detached observation of the minutiae of our lives; his delicate hint that all of us are systematically corrupt and that human society is predicated on this; his amused tone: all these may suggest, at least to some, that far from being with the angels, Goffman now stands at the right hand of the Prince of Darkness.

If his moral stance produces one set of problems, his principal topic creates another. Asylums and Stigma concentrate on institutions and persons and, though they do this in a novel fashion, the subjects are familiar enough. But the matter on which he has mainly focused has been, very broadly speaking, that of manners and this is an unusual topic for a sociologist, and one with a slightly decadent feel to it, at least in contemporary society. The notion that “manners maketh man” may have been commonplace in Medieval, Renaissance, and even Victorian times, but it seems odd to think so now when “mannered” is a term of abuse.

Goffman's version of manners is of course different from that of William of Wykeham or Baldessare Castiglione (1964). His aim is not to instruct us how to behave. Nor is he interested in the origins and development of “civilized” manners—the theme of the other major sociologist of these matters, Norbert Elias (1978). His interest instead lies in the following notions: that there are manners appropriate to every type of social occasion; that occasions may be distinguished by their distinctive set or sets of manners; that each set of manners generates an internal structure, a framework of interactional possibilities for that occasion; that the set of capacities needed to utilize manners in an appropriate fashion is one of the central human competences and their development an essential task in socialization; that as individuals we are both shaped by and in turn shape manners, being on the one hand, grossly deformed by them, if the manners and occasions are to our disadvantage—as in Asylums and Stigma—but, on the other hand, possessing the capacity to use them to our advantage, as in Strategic Interaction (Goffman 1970); that, finally, the social world we live in is mediated primarily through manners—each set is heavily influenced by the fundamental historical processes that move and shape the world but at the same time, each has its own reality, its own measure of independence, however slight.

In summary, types of occasion and the manners appropriate to them constitute one of the essential levels in the hierarchy of human social organization, a level which can neither be reduced to the individual nor absorbed into the dominant macrostructures of social life. Manners and occasions are all about us, so close to us indeed and so familiar that normally we see right through them to the individuals and structures which lie beyond. Goffman's task has been to make large that of which we normally make little.

Not only is his topic unusual and his morality uncertain, but his professional style is most peculiar, even deliberately perverse. Consider in turn, his chosen discipline, genre, personal style, view of science, method, and faction. To what profession or discipline does he actually belong? Is he primarily an anthropologist or a sociologist, a social psychologist or an ethologist, or is he perhaps really a student of linguistics? All may equally claim or disown him. What does he himself claim? Very little, it would appear. Not only does he acknowledge no obvious disciplinary allegiance but he systematically refuses to tell us quite what he is up to. There are no formal retrospectives, replies to his critics, critiques of the works of others (well, perhaps one; Goffman 1981a) and hardly any reviews, either of the literature or of books. Most of the normal ways through which academics try to state their position and claim a particular patch as their own are ignored. All one normally gets are glancing asides and deflationary footnotes. He believed instead “that a writer is the last person to say what he is up to and that to do that one needs a literary kind of thematic analysis by another. (Note Weber's fatuous review of what he takes to be his own underlying concepts.)” (E. Goffman, Personal communication).

This refusal to link his work systematically to other academic traditions may well explain his enormous appeal outside the discipline, but for a sociologist the position is unusually modest and exceptionally single-minded. (Some, of course, might term it devious.) At any rate, for whatever reason, Goffman systematically refused to take part in the inter-galactic paradigm-mongering which conventionally passes for really serious sociology. This refusal had two somewhat related parts. On the one hand, he never located himself within the great tradition or chain of sociologizing, and, on the other, he never spelt out his own world view. Thus, while he identified a major level of social organization which others had previously neglected, he contented himself with its ever more detailed internal exploration. How this might revise or complement the work of previous authors and what its implications were for more macro- or more microlevels—these were matters he left to others (Goffman's letter was in response to a copy of a review of Forms of Talk (Strong 1982) which I had sent him. Unfortunately I never met him).

This highly personal and, to some, irresponsible style was matched by his genre: the essay. Most of his best work was done in essays, and his books are either collections of essays or else extended versions of the form. The essay is a term which derives from Montaigne (1958) who used it to describe his “attempts” or “tries” at reflecting in great detail on various aspects of his own life, thereby hoping to illuminate those of others. The essay is a systematically personal enterprise, rooted in everyday experience. In it, the author seeks ardently for a slippery truth, but makes no claim to having necessarily found it. It is an exploratory and essentially open form in which one may use data from anyone: friends, journalists, novelists, and social scientists alike. No one within it has greater authority than any other. Finally, precisely because it is so personal, the essayist is free to develop his or her own style, to make jokes, to be whimsical, to digress, to employ both the tragic and the comic modes; to use, that is, all the literary devices which the writer of the scientific article can, at best, only smuggle in surreptitiously. Those who proclaim scientific truth must dress in sober apparel; essayists may wear whatever they choose.

This lack of normal professional manners would seem, by all accounts, to have been matched by a marked absence, on occasion, of conventional personal manners. In both his public and his private life he refused to play the game. The serious student of manners could not himself afford to possess them, or, more accurately, to be possessed by them. Since manners were for Goffman as much an object of study as of practice, being with Erving could be unnerving. Gouldner, not apparently one of Goffman's best friends, has made the distinction between living for sociology and living off it. Only a dedicated and selfless few, according to Gouldner, lived for it; the rest of us lived off it. For Goffman, however, the distinction fails to hold. His style of sociology involves both. To treat one's entire life as data is at one and the same time to dedicate oneself entirely to the discipline; to relentlessly combat “that touching tendency to keep a part of the world safe from sociology” and to treat the whole of life, including sociology, its works and homilies, as a resource for intellectual exploitation. Data was what Goffman lived for, not sociology.

The reason that Goffman lived for data, made light of the professional trappings and chose the essay form, is that he held unusually modest views about the state to which we had currently got in micro social science. For him, we are only at the beginning of real science. To engage solely in complex theoretical debate or highly focused research is to mistake the overwhelmingly primitive nature of what we can currently achieve:

I do of course tart up what I write self-indulgently and also to get read. But science, of the naturalistic kind, is exactly what I think I'm trying to do. My model for the device I use—the essay—comes in part from my regard for those of Radcliffe-Brown … supported by an old aphorism of L. J. Henderson … that a single conceptual distinction could constitute a substantive scientific contribution. I am impatient for a few conceptual distinctions (nothing so ambitious as a theory) that show we are getting some place in uncovering elementary variables that simplify and order, delineating generic classes whose members share lots of properties, not merely a qualifying similarity. To do this I think one has to start with ethnological or scholarly experience in a particular area of behavior and then exercise the right to dip into any body of literature that helps, and move in any unanticipated but indicated direction. The aim is to follow where a concept (or a small set of them) seems to lead. That development not narrative or drama is what dictates. Of course nothing gets proven, only delineated, but I believe that in many areas of social conduct that's just where we are right now. A simple classification pondered over, worked over to try to get it to fit well, may be all that we can do right now. Casting one's endeavor in the more respectable forms of the mature sciences is often just a rhetoric. In the main I believe we're just not there yet. And I like to think that accepting these limits and working like a one-armed botanist is what a social naturalist unashamedly has to do (E. Goffman, Personal communication).

Goffman, then, believed in the oracle's basic message—don't expect too much. This unusually modest view, both of the achievements of social science and of his own contribution, led him to adopt a rather different method from that in conventional use in micro-social science. Such was its mystery but apparent power that one of the most baffling questions in contemporary sociology has been to know how he did it. There is even a story that someone applied for an SSRC grant to watch Goffman for a year and bring the secret back to Britain. The answer in fact is relatively simple. Its essence is contained in his statement above but the lessons are sufficiently important to be worth spelling out in more detail.

First, there is the belief in naturalism. If we want to find out what the micro-social world is like we should go out and look. Somewhere he accuses psychologists of standing too close to their instruments to be able to see anything else and the same goes for many sociologists too. Here we are, surrounded by the data of human social life, yet somehow we don't see it. All we see instead are our experiments or surveys or whatever. You could never guess, for example, at least from their published work, that most organization analysts had ever worked inside an organization, yet presumably, somehow, they manage to hold down jobs; it's just that they don't look at them. So the second precept is this: you too can treat your own life as data. Each one of us is a natural control group1; if our splendidly universal theories don't even apply to our own lives, there must be something wrong with them.

The third rule is pretty standard: get a good initial data-base. But the way Goffman applied it is unusual. Naturalistic data comes in all shapes and sizes. Most of its analysts, however, have extraordinarily narrow preferences. Some cluster round tape-recordings, others around pictures and yet others around the printed word; some, notebook in hand, spy on lives or institutions, others will only watch video; some swear solely by interview, others believe in nothing but observation. Perhaps, looking on the bright side, such choices reflect nothing more than that minute division of labor which affects any large and growing area of human endeavor. But they also smack of, and certainly seem to generate, a bitter sectarianism. Thus each area of naturalist enquiry produces its own brand of fundamentalism; the belief that only austere and single-minded dedication to this chosen area of investigation, whether texts or telephone calls, can truly guarantee a science of human conduct.

Goffman himself doesn't appear to have bothered with any of this. It would be wrong to say he scorned it, for that would imply some minimal attention on his part. He simply proceeded on his own way. That way was to use whatever sources struck him as rich and interesting. Thus, at various times he used as his initial database: detailed ethnographies of particular communities and organizations (Goffman 1968a, 1968b); a large collection of photographs used in advertisements (Goffman 1979); and a record of humorous mistakes made by radio announcers (Goffman 1981b). He might even, oh horror, invent his database where none was immediately to hand. Thus, his crushing response to conversation analytic fundamentalism (Goffman 1981a)—“the approach of the communication systems engineer”—was based, not on the tape-recordings of actual speech that are held so dear by the devotee, but on examples of possible conversations that he himself had invented. Manufacturing one's basic data is not advisable as a general rule, but the rule does not always hold. Our linguistic resources are so rich and our current formal data-base so impoverished that it is still legitimate, in order to make an argument or rebut another, to create examples that we can all recognize. For linguistic philosophers to trade solely in these is clearly a mistake, but there is nothing to prevent us from using them now and again.

We can therefore discern a fourth Goffman rule of naturalistic enquiry: don't get uptight about a single data-source. There are of course some advantages in specialization; each source has distinctive features which the complete novice will do well to heed. But it is mere adolescent fantasy to pretend that their scientific analysis has advanced so far that no one, bar the life-time specialist, should dare pronounce upon them. We would instead do well to follow the example not just of Goffman but of historians and look at film, literature, life, letters, photos, pictures, tapes, and videos—wherever, that is, our particular topic leads us. Most aspects of most phenomena are still unexplored, at least formally; we are all, as yet, relative amateurs.

Such thoughts lead straight to the next rule, that of comparison. In one sense there is nothing special about this. The comparative method is the very essence of science. But it is a method which most naturalistic enquiries seem to ignore or, at best, pursue half-heartedly. Through focusing so intently on a single data source most of us commit exactly the same sin as the experimental social psychologists. We stand so close to our sources that we can't or won't see anything else. The only comparisons we can make are internal. Since our sources are normally so tiny, the power of our theorizing is normally trivial. For Goffman, however, there was no point in analyzing a single data-source. What is for the most of us the end of our enquiry was for Goffman merely the beginning. Having done his ethnography or collected and analyzed his initial data, he would then scour a host of other more readily assimilable data sources to see what light they threw upon the matter in hand. And since he, like the rest of us, could only engage in primitive forms of analysis, he was not fussy about the purity of these other sources. Novels, biographies and newspapers: all these were ransacked along with the academic literature.

To take one example, instead of rushing to publish his PhD thesis “Communication Conduct in an Island Community” (Goffman 1953), a brilliant study based in the Shetlands, he waited. By waiting, by testing his hypotheses on a range of other much wider sources, he was able to produce a more general and more valuable set of conclusions: The Presentation of Self in Everyday Life. These were, and are still, highly tentative but they gain their theoretical power from being based on extensive as well as intensive research. Adequate theory can be based only on comparison. In micro-social science very few make the attempt or even know that it is worth making. Goffman shows us how to do it.

To close this long list of professional peculiarities, let me give a brief note on Goffman's position in the perennial debate over process versus structure. Most social scientists favor one rather than the other; Goffman favored neither. Because he had read Mead, opposed Parsons and was a leading North American practitioner of ethnography, many have assumed that he was a symbolic interactionist, a process man. To this rather naive point of view, Gonos's (1977) detailed analysis of the major structuralist emphasis in Goffman's work is a most valuable corrective (E. Goffman, Personal communication).

Gonos lays special weight, rightly on “Fun in Games” (Goffman 1972a), Goffman's single most important theoretical statement. Yet, though Gonos does well to stress the profound influence of Durkheim, he is surely wrong, as we shall see, to class Goffman with extreme structuralists such as Althusser.

Goffman's refusal to commit himself to any particular discipline or paradigm was matched by and, indeed was part of a refusal to simplify the complexities of human beings and human social organization. We are free and determined, there is structure and process. Individuals and societies have their separate existences as well as their mutual comminglings. Goffman refuses to deal in either/or. Both sides of these dichotomies have their special truths. At some points he stresses one, at others another. In Interaction Ritual, Goffman (1972b) toys with the notion of people as puppets, as mere devices for the enactment and maintenance of pre-existing and superordinate structures. In other places, he demonstrates the reverse truth: the capacities, techniques and margins of freedom that allow individuals space for their interests, purposes, identities and manipulations (Goffman 1972c). Structure is always powerful; never totally so.


What, then, was the importance of being Erving? The verdict must ultimately rest on the usefulness of the conceptual and methodogical tools which he provided. And this is an empirical matter which cannot be settled by a priori theorizing. We shall just have to wait and see. For those who cannot bear to hang around so long, one or two guesses may, however, be in order.

A good many of Goffman's peculiarities seem important. The current division of the social sciences is clearly barmy (Strong and McPherson 1982), as is our steadfast refusal to acknowledge the work of the other picturing trades such as novelists, biographers and journalists (Strong 1983). Goffman's blithe capacity to ignore our petty demarcation disputes is one we should all copy. His modest view of social science, his reverential but non-empiricist attitude towards data, his profoundly comparative viewpoint: all these are worthy of emulation. None of them, however, whether collectively or individually, would render him the subject of undying fame. What puts him in with a chance is his attempt to pin down the basic micro-social forms within and through which we live our daily lives. He was the first since Simmel to recognize their importance as a level of social organization and the analyses that he produced are, far and away, the most powerful that anyone has yet achieved in the area. Goffman was the Linnaeus of interaction, the man who first began to bring serious conceptual order to the domain.

Not that there is anything very neat about most of the distinctions Goffman made. Faced with the extraordinary diversity of the human social realm, a realm every bit as varied as the natural order and a good deal more flexible, he tried out an extraordinary range of classifications. There is a seemingly endless conceptual proliferation in Goffman's work; something which, for all his style, often makes him hard to read and even harder to summarize. Even his most famous concepts tend to disappear in his next book emerging elsewhere with a new name and a cunningly revised definition. His principal legacy is in one way therefore rather un-Linnaean; not so much a neat classification, more an armful of concepts and methods through which others may carry on the work and improve upon his own initial attempts.

Where that may eventually lead is not clear. What is obvious, however, is that none of the traditional micro theories will do any longer. The versions of the individual and of human action that underlie the macro theorizing of Marx, Weber, Durkheim and Parsons have all been superseded. The same is true for our models of what lies beneath strata, institutions and organizations. These too ultimately rest on images of action and interaction, individuals and individual capacities, even if these in no way wholly determine them. A fundamental shift in our mode of conceptualizing micro-order and micro-process has important implications for the kinds of model we produce at a macro level. Goffman himself never tried to make that link and, so far, those foolish enough to attempt what their master avoided have done so from a fundamentally micro point of view (Barth 1966; Collins 1975, 1980; Strong 1979; Strong and Dingwall 1983). Major reappraisal will most likely have to wait until macro theorists too have assimilated Goffman; a process which may take time. That, however, we may reasonably expect such re-assessment is prompted by the following consideration of Mulkay's when reflecting on the nature and development of science:

to a very considerable extent, scientific knowledge has developed by the identification and detailed investigation of phenomena which have not been known to exist previously or which have not been studied before in any depth. The typical pattern of growth, then, in science is not the revolutionary over throw of an entrenched orthodoxy but the creation and exploration of a new area of ignorance. (Mulkay 1979)

This account provides a better version of the likely impact of Goffman's work on micro-social science than the more conventional language of paradigm-shifts. Something like the latter has certainly occurred in some areas for the vocabulary and the basic models have changed. But what in the long run may be far more important is the reason why this is so. Goffman's model has proved appealing, not so much because it neatly re-ordered some jumble of facts that were already within the discipline—though it certainly did this—but because, for the first time, it allowed social science to encompass a range of phenomena that everyone knew about but which were somehow omitted from scientific consideration. It is the sheer obviousness of the phenomena, once pointed out, that makes Goffman seem likely to join the immortals. He will do so, however, only if they turn out to be more than trivia. And finding this out is the task of his successors. If normal science does ensue, Goffman will no longer seem odd.


  1. 1

    A phrase of Janet Askham's.


  • P.M (Phil) Strong (1945–95) was one of the leading UK commentators on Erving Goffman's work. Educated at the universities of Oxford and Essex, he worked successively at the Universities of Aberdeen and Oxford, and at the Open University, before moving to the London School of Hygiene and Tropical Medicine. His book, The Ceremonial Order of the Clinic, is one of the most systematic applications of Goffman's ideas, in this case to analysing communicative interactions in pediatric development clinics in the UK and the US. This paper originally appeared as an obituary following Goffman's death in 1982.