TOWARD THE END of The Presentation of Self in Everyday Life, Erving Goffman writes, “The claim that all the world’s a stage is sufficiently commonplace for readers to be familiar with its limitations and tolerant of its presentation, knowing that at any time they will be able to demonstrate to themselves that it is not to be taken too seriously.”1 The object of this essay is to ask, how seriously is too seriously? Since one perfectly plausible answer to that question is offered both in The Presentation of Self in Everyday Life and in Frame Analysis—an answer that reminds us that what goes on on the stage is not real—some justification is needed for making more fuss than that about so-called dramaturgical explanation. If we know why we should not take the dramatic analogy too seriously, do we need to take it seriously at all?
The justification for doing so runs something like this. A large part of social science requires us to explain why actors act as they do in the situations in which they find themselves. To explain their behavior is to reconstruct what Karl Popper terms “the logic of the situation.”
This, again, can plausibly be represented as explaining their behavior in relation to their doing what is, in some sense yet to be explained, “the thing to do.”2 It is a commonplace of recent sociology that the concern of traditional sociologists with social structure, with invariant relationships rather than varying personnel, often led to the assumption that “the thing to do” and the actor’s belief that it was indeed “the thing to do” could be read off from some universally agreed description of the social role occupied by the actor. Hence, one mode of explanation by situational logic would simply be that of discovering an actor’s role and inferring what it required him to do on a given occasion. To understand what an agent did would simply be to understand what role he was filling. There are, no doubt, far more questions raised than answered by such a claim. It tells us nothing about why a given society recognizes the roles it does; it does not suggest any way of discriminating between those bits of activity that are role fulfilling and those bits that are not. All this is commonplace.
Appeals to roles leave out questions of motivation in particular. That is, at one level, we may suppose the question “Why did Smith shoot himself after he had been slapped by the pork butcher?” is answered when we are told that Smith was an officer in a crack regiment. To be slapped by a social inferior with whom it was impossible to settle the matter by dueling would be to be dishonored; only suicide could restore his honor and that of the regiment. But what we do not know is why he was so committed to the code that he was willing to accept death rather than dishonor. This suggests that if we want to know why someone does what he does, we must look to the goals he is seeking and to his beliefs about their rewardingness to himself. The assumption underlying the demand is the belief that men do what they find it most rewarding to do—even if, to paraphrase G. C. Homans, they find the damnedest things rewarding.3
This assumption provides the general structure of an answer to the question about an actor’s motivation for doing what, as we say, “the role requires” him to do. The answer to the question of why Smith shot himself is that he found death more rewarding than dishonor, or that, of all the actions available to him, death by his own hand was the most rewarding. This still leaves unanswered the question of why he finds death the most rewarding of the available options; nonetheless, we should have no trouble, in principle at least, in providing a genetic account of how he came to feel that he would rather die than live dishonored. The structure of explanation would be to assume that what was done was the most rewarding thing to do, and then to answer questions about how it can have been such in the individual’s socialization. Sometimes we are unlikely even to be tempted to invoke the role a person occupies when explaining his behavior in maximizing terms. The man who is dying of a painful disease and who therefore shortens the process by taking an overdose of sleeping tablets is quite readily represented as calculating that if the returns on being dead are zero, the returns on dying slowly and painfully are negative, so the rational choice is to be dead sooner rather than later.
There is some tendency for writers to oppose role-filling and returns-maximizing models, as if the fact that one account talks about “officers” and the other about Smith, Jones, and Wilkinson placed them in competition. Of course, this is not so. There is much to be said about an excessive reliance on the concept of a role or, relatedly, on the concept of a norm, and about the dangers of assuming that men will just do what they think they ought, or what their role requires; and there is much to be said about the dangers of tautology in the returns-maximizing model, as well as more sophisticated things about which elements of a theory of motivation we can safely agree to be tautological and which elements we must at all costs preserve as vulnerable to empirical test. Still, the immediate point is simply to observe that there are both actors and roles, that Smith is an officer and that an officer is what Smith is.
It is this insistence on preserving both aspects of the role-filling and returns-maximizing model that lets in the third view of how to determine what “the thing to do” is. We accept both premises of the previous models; there are roles with rights and obligations attached to them, and rights and obligations do provide reasons for action and can, therefore, feature in the explanation of why an actor did what he did: but there are also agents who are individual, flesh-and-blood creatures who are not bundles of rights and duties and who have goals, wants, wishes, purposes of their own. Explanations of social behavior must accommodate both persons and parts, just as an explanation of a game of chess needs to accommodate not only the rules that lay down what moves each piece can make, and what counts as victory and defeat, but also the aims of the players, their skills, their temperaments and so on. The sociologist may, perhaps, rest content with giving a structural description of a society’s role structure—and no one reading Siegfried Nadel could be in any doubt about the importance and the difficulty of so doing.4 But, if we want to explain what happens during some strip of social interaction, we need both roles and actors, both the recognized rights and obligations and the hopes, fears, skills, and weaknesses that the actor brings with him in filling his roles.5 It is an insistence on this that seems to mark out the “dramatistic,” “dramatic,” or “dramaturgical” approach to the study of social interaction.6 (I am not sure that anything hangs on which of these expressions one adopts.)
I now broach the question that will haunt what follows, even though little of what follows tackles it explicitly. This is the question whether the dramaturgical model insists on the theatricality of social life merely in the sense of insisting that people fill roles just as persons act parts in a play; that they can, therefore, cease to fill them, can at all times fill them with more or less commitment to the standard goals and purposes that the roles appear to presuppose; that they can, at least sometimes, stand back from their own performance of the role and ask themselves why they are bothering to go on with it when they could be doing something more interesting or more agreeable; that they can take advantage of the role to indicate to others that they are not quite the people they seem from the performance they are putting on.7 That is, it is the question whether the crucial element in the dramaturgical picture is that cluster of insights that goes under the general heading of “role distance.” Or is the crucial element the claim that style is the man? That is, that we leave out the aesthetic dimension of social life at the risk of misunderstanding much of what goes on; that when people put on a good performance of some social part, this is not simply a device for leaving in other people’s minds the impression that they are peculiarly estimable or admirable, nor necessarily to plant in other people’s minds false beliefs about the terms on which they are prepared to cooperate in the projects of others, to exchange benefits with them, or whatever. Is it, perhaps, the claim that the only way in which we can accommodate the twin pressures of social life, on the one hand, and of our private selves, on the other, is by turning that process of accommodation into a work of art, into a project that is better understood by poets, playwrights, and novelists than by a cost-benefit analysis or the functionalist analysis of roles?
Antoine de Saint-Exupéry was very fierce about a young man who pulled on a pair of white gloves before blowing his brains out.8 Was he complaining of the “theatricality” of the gesture? If he was, are we to infer that there is a distinction between serious matters and matters susceptible to dramatic renditions; or are we to infer that serious matters are susceptible to dramatic rendition precisely because theatricality is a vice that destroys good drama and only good drama?
So much by way of introduction. I now want to say something—though I do not know how to say enough—about the peculiarities of rational explanation and about the role of reconstructions of “the thing to do” other than the role of explaining an action or series of actions. To do this, I take a brief look at the perhaps overworked problem of accounting for voting behavior in the terms proposed by Anthony Downs’s An Economic Theory of Democracy.9 All this leads on to the suggestion that if the world both is and must be rather short of people who are rationally maximizing values—either returns to themselves or some other value10—it is no wonder that social scientists should turn their attention to some of the nonmaximizing activities in which people are engaged, such as staging social dramas or planning their lives as works of art. This remains nothing better than a suggestion, for reasons that become sufficiently obvious when we turn to some recent accounts of the phenomenon of suicide (and of Émile Durkheim’s Suicide).
It would be pointless to adduce as an explanation of an action the good reasons for doing it unless we accepted something like “the principle of rationality,”11 which Popper calls “false but indispensable” to the social sciences. Yet as Popper’s characterization of it suggests, it is a strange principle; it can hardly be a generalization from particular cases, for all the familiar reasons about the impossibility of identifying the explanandum in particular cases without already presupposing the principle. Thus, we seem to be left to read the principle as a conceptual truth or as a heuristic principle. That is, we might construe the principle of rationality not as saying that by and large men do the thing to do, but as the conceptual claim that what a (rational) action is is what an agent does because it is (in some sense or other) the thing to do; to read it thus would be to take it in something like the way Kant takes the principle of universal causality. Or we might treat it as a heuristic principle, which tells us not to give up trying to find reasons for a piece of human behavior, not to rest content with any explanation that does not show how what happened was at least subjectively the thing to do. People who are skeptical about heuristic interpretations of such principles might reflect on the work of, say, R. D. Laing, in which the principle is adopted in very unpromising circumstances in order to try to render the behavior of schizophrenics intelligible.12 The attraction of this is that it brings back into the fold of human actions events that would have to be taken otherwise as things which merely happened to the schizophrenic patient.
If the rationality principle is not a conceptual truth or a heuristic principle, it is hard to know what else it might be. As an empirical generalization about men generally doing the right thing, it is pathetically false, in view of the unfortunate human proneness to error. If it is to be read more narrowly, as it would be, for instance, by many sociologists when they claim that rational behavior is common in economic matters but not in social life generally, then “rational” is clearly being interpreted much more narrowly than it is here—either in such a way as to fulfill Max Weber’s account of Zweckrationalität or in some sense as equivalent to Vilfredo Pareto’s account of logical action.13
What this means is that, in one view, to show that an action was in any sense “the thing to do” just is to show that it was the rational thing to do, for showing that it was the thing to do is rationally to explain it. In the other view, to show that it is the thing to do is to render it intelligible; to show that it is the rational thing to do is to go one step further. There is no easy way to tell what hangs on the choice of a more or less restrictive account; if we want to insist on the significant differences between explanations in relation to people’s goals and knowledge, on the one hand, and straightforward causal explanations, on the other, presumably we shall have no objection to the wider usage; if we want to insist on how differently the artist, poet, or priest shapes his means to his ends from the way the entrepreneur and engineer shape theirs, we shall presumably want to narrow the notion.
Downs’s Economic Theory of Democracy illustrates some of the problems we are faced with here. Downs’s book expands on a suggestion implicit in Joseph Schumpeter’s Capitalism, Socialism, and Democracy to the effect that political parties and voters can be understood as if they were firms on the one side and consumers on the other.14 Voters spend their votes to secure the greatest possible flow of utilities from the policies that political parties will implement if they gain office. Political parties are assumed to have only one aim, that of securing power by selling their policies to as many voters as possible. Initially, Downs credits all parties with much the same omniscience that the theory of perfect competition assumes in orthodox economics, but the analysis gets exciting only when real-life phenomena analogous to brand loyalty and information costs are allowed to intrude. The best-known crux, however, arises when we make only the assumption that voting is an activity with opportunity costs; because once voting is allowed to have costs, Downs sees that it is very hard to explain why people vote at all. The rational voter operates with some estimate of what difference it will make to his overall utility if one party rather than another should win; what he then has to work out is whether the chance that his vote will make all the difference between the “right” and the “wrong” party winning, multiplied by the difference it makes for the “right” party to win, is a larger sum than the cost of voting rather than doing any of the things he would rather do. Now, it is evident that in a large electorate, the chances that any given individual’s vote will make all the difference are tiny; the value of the voting act, therefore, is tiny also, since it is unlikely that the party differential is so enormous that it would make a serious difference to the result. Hence, if voting has any cost at all, the voter will not vote—not if he is rational, that is.
Downs begins his book by suggesting that the only point of a theory is to predict what happens, and therefore commits himself to the view that the only proper course of action when we are faced with such a divergence between what the theory demands voters should do and what voters actually do do is to junk the theory. One would suppose, therefore, that Downs would simply throw out the hypothesis that voters are acting to maximize their utility flow. In fact, he does not do so. So strong is the grip of the theory that he looks round for something other than the party differential to supply the voter’s payoff. Thus, he suggests that the voter is not just voting for the party that offers him the better deal, but also voting to sustain democratic government too. If fewer than a certain number of voters turn out, democracy will collapse; the voter, therefore, has to consider the costs to himself if the whole system were to collapse. This, in essence, is a way of pushing up v (the utility of an outcome) to compensate for the smallness of p (the probability of its occurring under specified conditions). And like other arguments addressed to self-interested persons in order to induce them not to free-ride, it is bound to fail if the odds against their individual contribution being decisive one way or the other are very high. The odds against any single voter’s participation or abstention being decisive are so enormous that the self-inflicted costs of abstention are bound to be very low, no matter how much the individual may dislike the prospect of the collapse of democracy.15
Now, here two avenues open up that we ought to keep distinct. The first is that which leads us to investigate what decision rules people do use, in order to see why they make decisions that pure utility maximizers might not make—supposing they were infallible in their estimate of the probabilities attaching to the various outcomes. In essence, we should stick to the view that people act self-interestedly, but look more closely at how they do so, in order to see whether they are, perhaps, satisficers rather than maximizers, or whether they are intendedly maximizers but have curious views about probability. Now, along these lines there are many interesting results we might want to investigate. For instance, the phenomena of football pools and insurance companies suggest two things—that when we are faced with the prospect of really large losses, we overestimate the danger of their occurrence and thus accept odds that a true utility maximizer would not, and that when we are faced with a really large gain, we are prepared to gamble a small amount at odds that the utility maximizer strictly ought to reject. Downs might, along these lines, have argued that voters will overestimate the possibility of system collapse, or else that there is a threshold below which they do not count costs at all and would happily vote in the same frame of mind as they would spend fivepence on a lottery ticket. But Downs takes the second line, which is to treat as sacrosanct the assumption that men are utility maximizers, and to then look round to see what yields them a utility. Downs tinkers with the payoff to the voter by suggesting that the pleasures of a satisfied conscience make it worth the voter’s while to go and do his duty and vote. It is, of course, an empirically well-confirmed view that people’s disposition to vote at all correlates much more closely with the strength of their sense of “citizen duty” than with their perception of the differences between the parties. This certainly suggests that the first version of Downs’s theory is less likely to hold water than the second version.16
But, of course, once we allow “doing one’s duty” to feature in the desires of the agent, we run into exactly those troubles that Thomas Babington Macaulay so enjoyed pointing out to James Mill.17 The theory that men maximize their flow of utility loses its explanatory stuffing. We begin with the belief that when rational behavior is thought to be utility-maximizing behavior, we can find some simple constraints on the sort of thing that will yield men a utility—such as money income. We then find ourselves baffled, because if men were acting to maximize money returns, say, they would not vote at all. Now, when we stand by the notion that they must be maximizing something or other, we end up with a different theory and a different empirical task. We do not have a predictive, falsifiable, and, indeed, falsified theory; what we have is an analytical framework into which we try to fit a rather different predictive theory, if we can find one.
Faced with the failure of utility-maximizing models to do as much as they promised to tidy up the explanation of choice, one can see why people have been attracted to alternative models, such as the dramaturgical. But it can hardly be said that the model has received what one might take to be a definitive statement, or to have been spelled out in such a way as to achieve consensus on what its scope and limits are. The situation is not helped by the fact that the writer with whom the model is most frequently associated, Erving Goffman, has more than once done his best to get out from under its shadow. That is, a good deal of Frame Analysis is devoted to denying that everything is a piece of theater; Goffman insists that events in everyday life have real consequences in a way that events depicted on the stage do not, and he spends the bulk of a particularly engaging chapter elaborating the devices by means of which we make sure that what happens on the stage is not taken to be part of real life.18 The topic is not an easy one to keep clear; but a small start might be made if we were careful to distinguish between pretending and depicting. It is important that the actor who plays Othello does not pretend to be Othello in anything like the sense in which the con man pretends to be a gas-meter reader; his playing of Othello is not at all intended to mislead anyone into supposing that he is Othello, whereas it is the essence of the con man’s activity that he should induce us to believe that he really is the gas man, and just as it is an essential part of the fraudulent claimant’s activities that he should be taken to be the very same person as the legitimate heir. It is an important consequence of Goffman’s analysis of, for instance, the occasion when a member of the audience leaped onto the stage and began to attack the “hero” of Look Back in Anger that whatever sort of suspension of disbelief it is that the theater requires, it cannot be one that requires us to think that we are really seeing Othello strangle Desdemona.19 The proper response to Othello’s strangling Desdemona would be to stop him.
Now, it is certainly true that in the world that Goffman shows us there is a great deal of pretending going on. People are constantly pretending to be acting in good faith when they are not, pretending to competences they do not possess, and pretending to commitments they do not feel. At any rate, one aspect of the notion of role distance—though not, perhaps, the most important—is its reminder of the way we often pretend to be more at ease in a given task than we are, the way we try to mislead others about our ability to perform the task at hand with something to spare. To that extent, it may be thought that the element of deception is crucial and that “putting on a show of doing” is the central notion in dramaturgical analysis. This, I think, is to chase too hard in only one direction. For the notion of depiction, which is more obviously dramatic in its connotations, has a role to play. We do not need to suggest that a man who plays the role of surgeon is only pretending to be a surgeon; what he really is is a surgeon all right. But, we might well want to say that what he does qua surgeon also depicts what a surgeon does, in spite of the obvious awkwardness of saying any such thing. The awkwardness is, of course, that it seems odd to say that a thing might be both itself and a picture of itself; but there is no particular awkwardness in saying that a given piece of behavior might be both that piece of behavior and a picture of pieces of behavior of that sort. A cricketer who plays an admirable cover drive both plays that particular drive and displays an image of the good cover drive. So in this view, we ought not to be unduly alarmed by the suggestion that a given activity is both the real-life activity of saving lives by surgery and an exemplary picture of what it is like to do the job conscientiously, efficiently, skillfully, and so on.
Nor does the argument end there; we might go on to claim that it is just because the activity can be seen as an image of that sort of activity that it allows room for considerations of style, for an aesthetic dimension. That a man fills the role at all is not usually a question of style; to be a surgeon at all is mainly a question of ability, or qualifications, or what he usually does to the patients confided to his care. Doing the job is a technical matter; but the surroundings in which the job is done offer the chance to do it in style rather than merely. In something like surgery, style is very much the man—bound up with how an individual manages the demands on him; but it is also an element in the role, in the sense that an account of the style in which a role can be filled is one of the things we would want to know about any role before we felt we understood it. It would be a thin knowledge of cricket that did not include the insight that Wally Hammond was, and Ken Barrington was not, a stylish player.
This is by no means all there is in Goffman’s accounts of “strips” of social interaction, and a worrying aspect of Goffman’s habitual offhandedness about his theoretical commitments should be broached now before we push on to take a closer look at dramaturgy. This is the ambiguity between a disposition to see “what is going on” as something played as a drama, and a disposition to see it as a game in the game-theoretical sense, a game in which we are standardly beset by something like prisoner’s dilemma problems and in which we are constantly trying to get as much as possible for ourselves out of an essentially fragile cooperation with others. In this second reading, the importance that Goffman attaches to the control of information about ourselves has a much more obvious explanation than it would on a more committedly dramaturgical view. It is the object of each of us to convey an impression of himself that will make the terms of cooperation more favorable to himself. On this view, the surgeon who does his job stylishly is not concerned to make the terrors of dying and of causing death more bearable by distancing them in an aesthetic framework, but to secure his dominance over the rest of the surgical team, to induce in them the belief that he is vital to the success of the operation and that to secure his cooperation, they have to yield to him. Or, more generally, what the actor is up to is trying to rig the social terms of trade by engendering the appropriate beliefs in everyone else. This, of course, requires what Goffman always insisted on, the essential gap between agent and role. Everyone in everyday interaction knows that social life is possible only if everyone is kept sweet enough to be willing to do his bit in the social undertakings on which we all depend; so everyone is watching out for signs that other people are reliable or unreliable, competent or incompetent. A feature of Goffman’s work that is apt to offend people is the apparent implication that we spend so much time either sneaking a look at the shortcomings of others or trying to tart up the image we ourselves present. A good many of his readers protest that they rarely, if ever, think about putting themselves over to an audience and never, or at least very rarely, entertain doubts about the sanity, sexual normality, good temper, or whatever of their associates. If they are outraged and are also attracted to one of the contemporary versions of Marx, they may insist that these anxiety-ridden phenomena are typical of a society whose values are corrupted by commodity production, and where, therefore, we worry incessantly about what price our virtues will fetch and about whether we shall be sold a pup by others. I am not at all sure that anything so sinister ought to be read into it all; though it does appear that there is a genuine puzzle about the generalizability of Goffman’s work, a genuine puzzle whether what he says is true of Americans and not of the British, as Michael Banton suggests,20 or whether it is true of advanced industrial societies and not of simpler, preindustrial societies. The only question at issue here is whether we should treat his account of social interaction as a story about how we rig the market or as a story about how we engage in putting on a good show.
In The Drama of Social Reality, Stanford Lyman and Martin Scott suggest that they have something of the same anxieties about Goffman’s theoretical commitments.21 But they regard themselves as thoroughly committed to the view that life really is theater. The question then arises to what are they thus committed, seeing as G. H. Mead and Freud are invoked along with Goffman to explain what the dramaturgical standpoint is, even though Goffman is then dismissed as not a full-fledged dramaturgical theorist. The obvious considerations on which they draw are such things as our ability to rehearse in foro interno (literally, “in the inner court”) what we are going to do before we do it—but that in itself scarcely seems enough, since the mere fact that I work out a calculation in my head before I write it down will qualify as an instance of rehearsing what I am going to do without tempting many of us to see it as the rehearsal of a drama. Again, they seem to suggest that in all aspects of social life we follow a script; but for the reasons suggested above, it is not clear whether this is a script or a game plan, whether it is a script containing technical instructions or the lines in a tragedy. I am following a text when I follow the instructions in the car’s handbook and change a wheel; but Goffman’s insistence that there is a difference between changing a wheel on the road and changing a wheel in a play still makes one want to say that even if the two activities looked like each other, we are miles from showing that all life is theater.
It might be said that changing a wheel is not a social interaction, being a matter merely of the driver and inanimate matter locked in combat; but merely introducing another character to assist one in following the instructions does not seem to alter the analysis much. More plausibly, we might say that only where we do what we do in front of an audience, real or imagined, is there room for the theatrical analogy to operate. Thus, where it is not simply me, or me and a mate, wrestling with the beastly wheel, but me representing myself to some audience or other in the guise of “one who is wholly incompetent, but through no fault of his own,” we might say I am putting on a little drama. But at just this point the ambiguity we noticed above comes back: can we not equally well say that what I am doing is telling people that my imposing upon them the burdens of assisting me is not the result of a flaw in my character or of mere idleness? If we can, we are back where we were with the question of how we are to distinguish an insistence on social life as a drama from an insistence on attending to the ways in which people set out to secure a definition of the situation, either cooperatively or in competition with each other.
There is one hint that might suggest that the attractions of the emphasis on aesthetic explanation rest on something other than the false claim that all the world is really a stage. The suggestion stems from Freud’s concern with art, and at best it surfaces briefly in the analysis offered by Lyman and Scott. The hint is that the enjoyment that we get from art, which lies at the root of all aesthetic activity, is to be explained by the role of artistic transformation in making the pains and anxieties of life more bearable. In Lyman and Scott’s account, it sometimes looks as if it is Freud’s theory that is the drama, with id, ego, and superego playing parts in a tragicomedy;22 the more interesting notion is that within the theory, ego is to be recast, not as an economic calculator trying to maximize psychic gratification in a dangerous world, but as a poet or dramatist making life bearable by giving it a unity that is not itself explicable in economic terms. That is, even preserving the claim that in some sense the task of the ego is to adjust the dealings between the organism and the environment, the claim would be that the nature of this adjustment was essentially rather than contingently to be understood by invoking considerations appropriate to the evaluation of a work of art. It is a striking claim, and it goes much beyond anything to be elicited from, say, Mead’s observation that behind the social Me there lies the personal I; for it is a claim, however inexplicit, about how the I shapes what happens to the Me, and about why the I is prepared to put up with all sorts of misadventures to the Me so long as they are organized in the right way.
Still, this is no more than a speculative throwaway. If we go back to our earlier anxieties, there is one further issue to settle before ending with a confession of bewilderment. It is sometimes said that dramaturgical explanation must be parasitic on other sorts of explanation in the sense that what goes on in the theater is already parasitic on what goes on in the nontheater. There are two rather different points to be made. The first is that some versions of everyday behavior depend for their effect on being copies of what goes on in a film or in a play or whatever; the surgeon who puts on the manner of Gary Cooper for some particularly grim operation is plainly borrowing from the realm of the theatrical in the strictest sense. But this is not to say that the dramatic qualities of real-life behavior depend on our lifting the script from a drama—what goes on in plays, films, and the rest is a heightening and a concentration of what goes on elsewhere. But, second, there is certainly a sense in which dramaturgical explanations are parasitic on other explanations. In any case in which a dramatic explanation is in place, there is some other explanation in place as well. The woman who commits suicide and represents herself as having been truly brokenhearted, therefore, may be said to be performing the last lines of a tragedy of blighted love.23 But we have two explanations pinned to one action here. The reason for committing suicide might be to make a coherent ending to an otherwise intolerably muddled existence; but the misery that needs resolving is a misery in its own right. The woman may be engaging in a dramatic rendering of “I am really brokenhearted,” and the half-intended impact of her death may be to ensure that the blame for it falls squarely on the man who jilted her; but none of this would make any sense unless being brokenhearted was already understood to provide adequate reasons for putting an end to oneself. If we invoke the analogy of the drama only to suggest that her suicide was a saying as well as a doing, the passing of a message as well as an act of self-destruction, the burden of the explanation falls as much on those motives to which she is drawing attention as it does on the dramatic presentation of them. To get extra mileage out of the dramatic analogy, we have to suggest an explanation of why only a drama will put across the message; otherwise, the drama is not merely, as it must be, parasitic on the other account of motivation, but also thoroughly uninformative.
This point is important and general. When Lyman and Scott apply their talents to the analysis of Shakespeare’s most famous plays, the result is oddly unimpressive. They may be right to suggest that what Antony and Cleopatra is in part concerned with is the rise of bureaucratically rational administration; but what we want to know is why should Shakespeare write a play about a phenomenon that is handled so differently in, say, Talcott Parsons’s The Social System?24 The distinction this rests on, between the message and its vehicle, is one that is crucial to the anxieties I have expressed about separating out the information-passing account of behavior from a genuinely dramaturgical account. When Durkheim attends to moments of social effervescence and explains how religious rituals enforce, symbolically, the moral relationship between individuals and society, one feels inclined to ask why this way of telling the individual these things is the way to do it. Even in Weber’s account of wertrational behavior, something of the same difficulty recurs; a given action is thought to be the only possible way of expressing a particular value or of driving home the point of a given way of life. And we are left anxious about the way the values and the expression of them are connected, and left to look for some account of how a mode of expression is or is not adequate to what it has to express.25
A last look at three sorts of suicide, in the light of the anxieties of what has gone before, may sum up the doubts of this essay, even if it will not lead to any conclusions. The three cases we might contrast are a man killing himself to avoid dying of a painful, wasting illness; a captain going down with his ship, having been forced to scuttle it; and the jilted woman who drowns herself. It does not seem at all difficult to account for the first in value-maximizing terms; the only awkwardness is posed by the values of the man’s society and the extent to which they restrain the pursuit of self-interested goals of this sort. The man certainly has good reasons, in self-interested terms, to take his own life, but according to the local norms, he may have good reasons not to do so. I do not have anything to say about when a man will take as his reasons for acting one set of reasons rather than another, but it is obviously a topic of central importance in accounting for rational explanation.26
The second case is analyzable so simply that it explains why role theory has been the object of some derision; that is, we might simply adduce in explanation the convention or norm that disgraced captains go down when their ships go down. If we want to know why Jones chose to drown, we simply point out that he was a captain who had been forced to scuttle. We have, as was suggested at the beginning of the essay, a problem that is the converse of that which the first case presented. That is, the first case raises the question of how norms constrain the acceptance of self-interested reasons; here we have the question of how self-interest constrains the acceptance of altruistic norms. The captain has a good, selfish reason for not going down with his ship, but he does not take it as his reason for action.
Although, as suggested just a little earlier, it is the third case that raises the question of how to apply dramaturgical concepts, it might be argued that they are at home here too. For what is it to go down with one’s ship except to end one’s career in style, to leave the stage in the grand manner? This is not, of course, to say that there is any room here for such a notion as role distance, for the thing about the captain is he does not distance himself from the role—he identifies himself with it so completely that he never thinks of doing anything but going down with his ship. But the suggestion raises the question whether the role of captain is sustained by the opportunities it offers for displays of heroism and so on.27 And one then has to decide how an act’s heroism refers to its “desirability characteristics”: is heroism to be explained simply as a function of altruism, endurance of pain, and so on, or as a feature of the show a hero can put on in front of an appreciative audience?
In the third case, there is no temptation to suggest that the woman played a role, so long as roles are thought of as bundles of rights and duties, though if we want to run the dramaturgical model for all it is worth, we may well say that “Sheila, the jilted woman” is a part, a dramatis persona. But just as we worried earlier about the tendency of dramaturgical images to boil down to stories about how behavior conveys information—getting a certain definition of the situation established and so on—here we may end up feeling that she has not put on anything interestingly analyzable as a drama, but has done something, part of the point of which was to establish why she had done it. What she was up to was, in part, telling us why life seemed to her to be so intolerable.
All of this raises far more questions than it answers, and this in spite of the fact that innumerable issues have not been raised at all, although a full analysis would require their resolution. Thus, nothing has been said about distinctions between rituals, dramas, different kinds of games, interior monodramas, and so on; nothing has been done to resolve the doubts of those who think that all this suggests that the only sort of theater ever devised is the naturalistic drama of recent western European culture. Again, nothing has been done to spell out the differences between those theorists who insist on our need for an audience to appreciate our activities and those who insist on our need to retreat “backstage” and put on consolatory shows in private. Indeed, the upshot of what has gone before is perhaps this only: first, to raise the question whether, in appealing to the dramatic or stylistic component of social behavior, we have invoked a distinctive kind of explanatory framework for activities otherwise misunderstood or merely noticed what was noticed before the Flood—the importance of impression management; and second, to raise the question of the indispensability of particular symbols, particular modes of passing information, and so on—the question, perhaps, whether in this area, the medium and the message are distinguishable.