The Etiquette of Improvisation

When I used to play piano in Chicago taverns for a living, I dreaded the nights when guys who had been playing dances would come in, after their jobs had ended, to sit in with our quartet. In a traditional jam session, we would play well known tunes and everyone would have a turn to solo, improvising on the chords of the song. Why did I dread it? If there were, say, four horn players sitting in, in addition to our own, every one of them would play the same number of choruses. If the first player played seven choruses of “I Got Rhythm,” the other four would all play seven; I would have to play seven, whether I felt like it or not; the bass player, if his fingers held up, might play seven, and the drummer too; then people might start trading four bar phrases ad infinitum. That could easily add up to sixty thirty-two bar choruses of a song whose harmonies are not very rich (I was fond of songs, like “How High the Moon,” that had what we called “interesting changes,” harmonies that changed frequently and departed from the original tonality). Remember that the pianist mainly plays accompaniment for all these choruses and you can see how someone who had already played for several hours might feel like falling asleep as the procession of choruses—not very interesting ones, usually—went on interminably.

It wasn’t always that bad. Once every several months, a lot of things, varying more or less randomly (although my colleagues and I often went in for theories that involved phases of the moon), would come together right, and the results would be extraordinary, we thought and felt. Usually that didn’t happen, and everyone involved was bored, not only listening to the other players’ choruses, but even to their own.

Why was that? For one thing, most improvising was not quite so inventive as the language we used (and that most people still use) made out. In one way, it was in fact spontaneous, created at that moment, and not exactly like anything anyone had ever played before. But, in another way (as Paul Berliner (1994) has amply demonstrated), every one of those seven chorus solos was basted together from snippets the players had played hundreds of times before, some they had come on themselves, many slight variants of what they had heard on records (of Gillespie or Parker or Getz); among these collages, especially when it was late and we had heard it all over and over again already that night, one of us might do something that sounded to our ears really different and original, even though it might well be something we had spent a week working out in privacy rather than something invented on the spot.

Why did we torture ourselves and each other that way? The etiquette of jam sessions required it. This very strict etiquette told us that the number of choruses the first player played set the standard others should follow. To play more would be rude, pushy, self-aggrandizing; to play less hinted that the first player had gone too far and, worse, that the following players who played less had less to say. (It usually happened that the first soloist played too many choruses, hoping to get something going even though he had started slowly.)

No one taught us these rules, nor had we read them in an etiquette column in Downbeat. We learned them by quietly observing, as youngsters, what older players did, and noting what happened when someone (usually a novice or some other unsocialized type) failed to obey these rules. The grossest examples I ever saw of someone failing to follow them came years later when groups of sociologist-musicians played together at sociology conventions. A few of them had not had the years of playing in such sessions the rest of us shared and would break in on each other’s choruses before they had finished their allotment and stop before they had finished their own. We never knew what to do with such people, believing that if someone didn’t know any better than that by now it was too late to try to teach them.

A rudimentary sociological theory suggests that etiquette is a way of providing for the systematic, formal expression of recognized and accepted relations of rank. That includes unequal relations, as when children are told not to speak until they are spoken to or when southern blacks were required to get off the sidewalk for whites or poor people were enjoined from wearing clothing that was “beyond their station.” It also includes relations of equality: fictitious, as when good party manners require people to pretend that everyone present is of the same social class whether they are or not; or real, as in the obligations of friends to one another. Etiquette is particularly important when people think that everyone involved in some situation ought to be equal but really isn’t.

That, of course, is exactly the case in the musical situations I have just described. Jazz is aggressively egalitarian, probably on the premise (common in many of the arts) that there is no telling where lighting might strike and what unlikely person might turn out to be really good even though they had never showed it before. Jazz players were, for instance, leaders in changing the etiquette of racial interaction in the United States (not saints, but leaders), although not particularly so with respect to gender. But not all players play equally well. In fact, in the unimportant places I worked in, all sorts of untalented and inexperienced people showed up to have another whack at it, in hopes of improving or gaining the needed experience. Some were wonderful, most were ordinary, a few were awful; they all got to play their allotment of choruses, in homage to the built-in occupational myth of equality.

Allowing the bad to play with the good, in deference to professional etiquette, has real consequences. Collective improvisation—not like Keith Jarrett, where one man plays alone, but like the more typical small jazz group—requires that everyone pay close attention to the other players and be prepared to alter what they are doing in response to tiny cues that suggest a new direction that might be interesting to take. The etiquette here is more subtle than I have so far suggested, because everyone understands that at every moment everyone (or almost everyone) involved in the improvisation is offering suggestions as to what might be done next, in the form of tentative moves, slight variations that go one way rather than some of the other possible ways. As people listen closely to one another, some of those suggestions begin to converge and others, less congruent with the developing direction, fall by the wayside. The players thus develop a collective direction which characteristically—as though the participants had all read Emile Durkheim—feels larger than any of them, as though it had a life of its own. It feels as though, instead of them playing the music, the music, Zen-like, is playing them.

If, however, the participants are not courteous to each other that way, do not listen carefully and defer to the developing collective direction, the music just clunks along, each one playing their own tired clichés. If the players are experienced and the clichés professional, the result won’t sound bad, although it won’t excite anyone either; if those things aren’t true, it will be a painful experience.

Etiquette might not even be the right word to describe the kind of attentiveness, care, and willingness to give ground and take direction from each other involved in that kind of improvisation at its best, although it is not a bad one. It calls attention to the granting of equal status to everyone’s ideas and intuitions. But there is something else going on that needs to be approached from another direction, something which involves the relations between the audience and the improvisers as well as among the improvisers.

When people improvise together they usually do so, though not always and not necessarily, by taking some elements of what they will do as given, not subject to change during the course of the improvisation, leaving others as what it will be O.K. to vary and work with as they perform together. Thus, for musicians of my era, improvising meant picking a tune, from among the ones we all knew, as the basis for the collective work. Picking a tune meant picking a melody and the accompanying conventional harmonies or conventional variations on those harmonies; and picking a tempo. Then we created variations, often forgetting the melody entirely though not its harmony or basic organization (where the cadences came, for instance), but never varying the tempo (that was known as “not keeping time” and signaled serious incompetence). The agreement to keep some things fixed and vary others made it possible for a group to sound like it knew, collectively, what it was doing: to not get lost, to have some idea of what might be coming next, to interpret what the others did as hints of a direction the collective effort might take. Equally important, agreement on these matters made it possible for a knowledgeable audience—whose members knew the songs the players used as a ground for improvisation and the limits of current performance practice—to listen to what was played and appreciate it, understand how the players were creating variations within a body of rules and known elements.

It might be stretching the use of “etiquette” unreasonably to say that, when improvisers stick to the tradition of what will be improvised and what will be fixed their audience already knows, they treat the audience as equals. Or that the failure to do that is a form of impoliteness that signals contempt for inferiors. Nevertheless, jazz fans usually did know the melodies and harmonies of the tunes jazz players used, knew where the band was in the song at any moment of the improvisation, and understood in some ways as well as the musicians what risks the players were taking, when they were playing safe, when they were coasting in familiar territory and when they were doing something strikingly original. Not all fans. Some knew a lot more than others and drew on a greater repertoire of history and past performance. But when musicians respected those simple rules, treating their colleagues and their audiences as knowledgeable equals, the music had a good chance to be both appreciated and worth appreciating.

A strict observance of these formalities would mean that nothing much would change. Such conservatism makes life easy but doesn’t create much art, although quite minor variations may be enough to sustain the interest of performers and audience, then and now. Both groups typically take changes to be much greater than they seem to be in the view of later observers. Audiences and players, confronted with people who start varying what had been taken as fixed, often treat that as a sign of disrespect. Performers confronted with such change often see that they will now be expected to learn how to do things that they don’t know how to do but others do, while what they do know how to do that others can’t do has been devalued. Audiences often take offense at these variations, suspecting that the performer who, for instance, changes tempo in the course of an improvisation either doesn’t know what he’s doing and thus insults them by pretending to be better than he really is, or knows that they won’t know what he’s doing and thus insults them by creating the possibility that he is just fooling around with them (a common suspicion of audiences confronted with art they don’t understand). In either case, the failure to abide by what we might call the etiquette of not-going-beyond-what-is-commonly-known suggests to sensitive audience members that they are no longer being granted equal status, but are instead being treated as non-participants in the important parts of the interaction.

In time, audience members—flexible members of the old audience or new recruits who are not bothered by people who change what used to be fixed—learn to play their interpretive part in the new effort. The process, which I have made sound like stages and steps, probably occurs in very small increments, so that relatively few people are inconvenienced. The very substantial differences between the playing of Roy Eldridge and Louis Armstrong, for instance, did not occasion the outrage among players and audience members that the change from Eldridge’s style to Dizzy Gillespie’s did, although those substantial changes were probably no more difficult to make than the earlier ones. It is probably not the fact of change itself but the degree to which it is surrounded by open signs of lack of respect for the rules of etiquette that makes the difference.

When do audience members (or even fellow musicians) interpret forays into unexplored areas, changes in what had been taken as given, as signs of disrespect? They could, after all, understand them simply as experiments which might or might not work; fellow players probably most often interpret them that way. Fellow players have a chance to hear all the tentative steps that suggest moving in the direction of the experiment; they have seen it coming and see the logic of that possibility. They may even have seen that same performer try other experiments with more or less success in the past, so that they know the experiment is sincere, not a put-on. Audience members have seldom had that opportunity. The organization of the performing arts is such that audiences come and go, seeing each performance once even though it is done nightly, while the performers remain together, thus able to know what their fellows are up to and see things in a longer and more differentiated perspective.

What I have been describing is improvisation in its most common and routinized form, as it occurs among professional entertainers performing for a lay audience. The opposite extreme occurs when the conditions are different: when, for instance, there is no audience immediately present and instead the improvisers are trying to solve a problem or perform a feat for its own sake or their own sake, because it is there to do and they have agreed to devote themselves collectively to doing it. Under those circumstances, the performers do not interact in a way that respects the conservative etiquette I have just discussed. If the rudimentary theory linking status and etiquette is correct, that means that participants in such a venture have decided not to act out the myth of equality embodied in that code. Instead, they act according to another set of rules, according special status to people who make special contributions to the collective effort, or are thought likely to. They agree, implicitly and collectively, to give priority to what, in their collective judgment, works and to give short shrift to what doesn’t, and not to be polite about it.

When people improvise this way, they become their own primary audience. Even though they may intend the results of what they do to be presented to some audience later, and even though an audience may be present while they work, they give no weight to putting on a good show or keeping up appearances. They concentrate on getting the job they have set themselves done. The image of a jazz group performing in public is not apt here; a better example would be actors improvising scenes in hope of discovering something they can use in a show or scientists discussing a difficult problem in hope of finding a lead to a solution.

Often enough, interaction in such situations simply recognizes a hierarchy of ability already well-known. Whatever Charlie Parker or Charlie Mingus wanted done got done, just because they were who they were. Similarly, scientists known for having solved important problems get a respectful hearing not accorded to people who haven’t proved themselves. Often enough, too, that reputation proves a false guide. Parker’s reputation gave his suggestions in a group’s improvisations a weight they often did not deserve, and many scientists who once solved an important problem make, for the rest of their lives, pedestrian contributions.

Leave that possibility aside. Consider what happens when all the participants ignore the past, ignore reputations, ignore everything but the contribution people make to the immediate collective effort. The rule in conventional improvisation is to treat everyone’s contribution as equally good. The rule in these situations is to treat everyone’s contribution as potentially better than all the others. Whenever anyone does something clearly better, everyone else drops their own ideas and immediately joins in working on that better idea. People do not move gingerly, gradually converging on some sort of amalgam of hints and implications, thus respecting the fiction of equality.

To be able to do that, everyone involved must have much the same idea of what the better looks like, a common criterion for knowing it when it appears. That suggests that they will have shared a past in which those criteria and their application have been worked out and applied in many cases, which in turn suggests their long-term participation in an organized world in which the kind of activity they are improvising is common, probably a professional or quasi-professional world in which ties of occupation bring people together in joint projects.

Likewise, people must have a real shared interest in getting the job done, an interest powerful enough to overcome divisive selfish interests. In an improvising musical or theatrical group, for instance, no one must be interested in making a reputation or protecting one already made. Scientists trying to solve a problem must care more for the problem than for their organizational positions or prerogatives. It is an interesting question—to which I have no answer—as to when people become unselfish in that way. But it is a common observation that, even in settings not likely to produce such a commitment (such as large-scale industries), people do act that way.

In this sort of interaction, great changes can occur. Not only do people respect and follow the lead of whoever comes up with something good, they may also collectively change their notion of what is good as the work progresses, adopting a new criterion, ending with a result that could not have been foretold from anything they knew and were used to doing before they started.

A good example of this on a more routine basis is the improvised theater game, in which people use a loose set of rules—something on the order of a children’s game—to organize a dramatic scene. They might, for instance, play “Exits and Entrances,” in which each player tries to maximize the number of legitimate exits and entrances he or she makes, legitimacy loosely defined as whatever everyone agrees “works” dramatically and helps the forward motion of the scene. The object is to produce a dramatically interesting scene, and whoever seems to be helping with that task is followed, when the game is well played. The example suggests that the occasions on which this kind of more radical improvisatory experimentation occurs need not be rare or require people of extraordinary talent or psychological abilities (e.g., tolerance for looseness), but can rather be as routinized as the more formal and polite versions considered earlier.

Clearly, we could identify a whole range of kinds of situations, varying between the two poles of those which work on the basis of an etiquette which recognizes and maintains a formal ideology of equality of status and those whose etiquette requires recognition of differentials in the contribution made to the collective effort. These differences are embedded in and supported by differences in larger social organizations, which create the conditions making one or the other possibility more likely.

In talking this way, I have gone beyond the face-to-face situation in which improvising typically occurs and is appreciated to the larger movements and organizations in which those situations are embedded. In so doing, I want to avoid the error often made in analyses of face-to-face interaction, the error of neglecting what is fixed in the situation under study by the influence of these larger entities. The relations between the dynamics of improvisation—the constant attentiveness and shifting focus of the participants, and the sensitivity to all that of an experienced audience—and the limits placed on those dynamics by tensions of the organized distribution of knowledge and the rewards connected to it suggest some of the dimensions still to be explored.


[“The Etiquette of Improvisation” was the basis for a symposium in the journal. Several people wrote comments, and I was invited to respond to them. What follows will not make complete sense unless you refer to the comments themselves. It appeared in the same issue of the journal, pp. 197-200.]

Examples and Generalizations

Kathryn Pyne Addelson has criticized philosophers—specifically, those who deal with the ethical problems involved in abortion—for their use of hypothetical cases to “clarify” issues:

The most obvious thing to say about this method is that although bringing up hypothetical cases may clarify our understanding of concepts and principles, everyone knows that the selection of hypothetical cases also biases understanding. This bias may be (unintentionally) systematic. For example, Judith Thomson [discussing abortion via such an example] gives a case where Jones faces a frosty death because Smith owns the coat. Why not, instead, use a case where men, women, and children face poor diets, poor housing, and loss of dignity because the owner of a mill decides to move it out of one region into another having cheaper labor and lower tax rates. Philosophers may say the second example is too complicated, but the selection is not a trivial matter of simplicity. The coat example ignores an essential distinction in kinds of property ownership, which the mill examples reveals. (Addelson 1991)

So, for instance, an argument that seems straightforward when you are talking about women who might hypothetically have been impregnated by seeds that floated in through the window loses some of its force when the case is Mrs. Jones, married, with six children, and dirt poor.

This problem occurs with respect to conceptual problems in general. We speak in abstractions which allow us, as social scientists, to compare cases and arrive at generalizations. But our concepts typically have in mind a specific case or sort of case, and the abstractions have features peculiar to this case or type. The abstraction gains complexity and power when new cases, which force us to add new features to the model, are introduced.

I think something like that is involved in this discussion. (What follows is based on my reading of the commentaries available to me, which are those of Cash, Sawyer, and Ruhleder and Stoltzfus.) My remarks on improvisation were based, as Cash says, on a specific situation (she might have said that I had a bad sample): late nights in the 1940s and 50s, in small Chicago taverns where players who had finished dance jobs would come to sit in with the house band, the owners not caring much what went on at that time of night. The players were not such bad players—in fact some of them were occasionally quite good, though none were famous then and few ever got to be famous—but the music was, for the most part, as undistinguished as my discussion suggested. Banal enough to bore the rhythm section that had to accompany all of them.

On the basis of this example, I suggested a small theory of improvisation that took off from the informal “rules” that seemed to govern what we did. The theory identified two aspects of what these musicians were doing. On the one hand, they treated each other as equals, in the sense that each player was entitled to a turn, and to a turn of the same length, as the others, and to the same kind of attention as the others. On the other hand, sometimes something special would occur: someone would play something that seemed extraordinary to the others and they would defer to that player and follow the lead that had been provided. Maybe not extraordinary, maybe just different enough to get everyone’s attention and provoke them to follow the lead and see where it would go. So the person who had proposed that change (proposed it in the sense of playing it and thus offering it to the others as a possibility) would for that time be treated as more equal than the others, as someone whose idea merited special consideration not given to others.

“Extraordinary” might be too strong a word for what was going on. It might be better to speak of small variations in standard routines, or even interesting mistakes. Ingrid Monson describes the latter situation in a recording of a conventional twelve bar blues by Jacki Byard. The bass player, George Tucker, somehow gets two beats ahead of the other players in the third chorus and, by the fifth chorus, is six beats out in front. The other players eventually decide that, despite their hints as to where the choruses are supposed to end, Tucker is not hearing them, and they eventually make an innovative leap and adjust to his mistake. Monson says:

When a soloist makes a mistake in keeping the form, the accompanying members will generally try to articulate the structure more clearly so that the soloist can get back on. When . . . a soloist remains in a different spot in the time cycle over several choruses, however, a sensitive band will adjust to the soloist, even if he or she is technically incorrect within the context of the original time cycle. (Monson, 1996, pp. 161-162)

That kind of intersensitivity among the members of a group is what I’m talking about.

This is a modest enough theory. I thought it of interest as a way of understanding some situations that jazz players found, and probably still find, themselves in. But I also thought, of course, that it might be relevant to larger questions involved in thinking about improvisation (otherwise, I would have offered it to jazz players, and not to an academic audience). Improvisation seems to describe so much of what goes on in other areas of social life than jazz that many people have used the term to describe other forms of cooperation in quite different settings: conversation, informal work routines, and the like. So I offered my theory as potentially applicable to all those other areas.

I did not imagine that the theory really would account, unmodified, for all these other phenomena. I knew, from Addelson’s explanation, that it bore all the marks of the example my more abstract statements generalized about. What I did imagine is something like what has happened here: that others would take the idea, apply it to situations they knew, identify differences between their example and mine, and use those differences to complicate what had been said, making the kinds of distinctions that only become visible through such comparisons. In doing such work, the more various the cases you compare, the more you get out of the exercise. Cash talks briefly about contact improvisation in dance, Sawyer talks a little more about improvisational theater, and Ruhleder and Stoltzfus give an extended analysis of master classes in choral conducting. These situations, like the one I described, are all about artistic activity, which means we don’t get the good of a comparison to, say, team sports or to a work situation.

We still get plenty. For instance, these cases suggest that the norms or rules which govern what goes on in the improvisatory situation can come from a variety of places. My description took the etiquette that governed sitting in as a given, presumably the product of the autonomous activity of people like the ones I described, now passed on as “how we do things.” In the choral conducting class, conventions come out of the classroom—the teacher is, after all, teaching people how to conduct—and the concert hall—because the student conductor is, after all, conducting a musical performance. Both those settings have well-known, historically legitimated patterns of interaction that everyone involved respects and within which the described improvisation occurs.

Cash says that contact improvisation occurs in settings that are unusual for dance performances—outdoors, industrial settings—but which neverthess contain audiences. She intimates that the dancers involved are creative enough to operate without relying on conventional structures at all: “. . . inventing new problems on the spot and pushing themselves to answer their own questions in ways that stressed and valorized the unexpected.” She doesn’t describe what the dancers do or how audiences respond in the kind of detail the other commentators do, so we can’t really add the case to our analysis.

Sawyer’s description of improvised theater emphasizes the “combinatorial explosion,” the myriad of possible continuations that open up when an actor proposes a first action for others to respond to. I think (and he’s right, I lived in Chicago and knew people in that business and am not unfamiliar with the genre) that there is somewhat more in the way of “culturally available” convention than Sawyer allows. “Bus drivers” were standard characters in Second City skits, as were psychiatrists, University of Chicago professors and students, and many other familiar social types. Perhaps more to the point, the collection of games contained in Viola Spolin’s Theater Games (Spolin, 1963), the “bible” from which thousands have learned to improvise on stage, provides a variety of frameworks on which improvisations can be hung.

Combining all these examples, we see that the jazz example defines one position on a dimension we could call “the source of structural conventions,” a dimension that ranges from no such conventions present (in the case of contact improvisation), to conventions developed within the world of action in which they are now applied (as in jazz and improvised theater), to conventions borrowed from neighboring worlds, as the choral conducting class borrowed the structures of the classroom and the concert hall.

Another problem that interests all the commentators is one that I have not given as much attention as they think appropriate. “Creativity” is often invoked as the counterweight to “structure,” the fresh, the good, and the new opposed to the stale, bad, and old. Who doesn’t prefer “creativity”? As Monson (p. 214), talking about the work of jazz rhythm sections, says: “The image of improvisation appears often in postmodern and cultural studies, functioning as a metaphor for hybridity, for the blurriness of boundaries, and for liberation from the inhibiting confines of both hegemonic ideologies and structuralism. Yet this usage is only partly apt.”

One reason the usage is only partly apt is that creativity and its opposites are terms of praise and blame, not analytic categories, and don’t work well in the latter capacity. “Creativity” is clearly not the same as “innovation” and neither are the same as “difference.” George Tucker, in Monson’s analysis of his “mistake” cited above, had done something different by introducing some extra beats into a very conventional twelve bar blues structure. That could have been an innovation, if the others had followed along and incorporated the difference into succeeding choruses (though it’s not clear that Tucker had any idea that he had done anything different). Is a mistake that others then build on a creative innovation? Does it become creative when others pick it up and go with it? I’m not trying to settle the question of what creative “really is,” but rather to say why I think it’s not a useful idea to work with.

I won’t pursue the many other comparative possibilities the commentators have opened up here, but leave that for the further improvisations this exchange might provoke.


Addelson, Kathryn Pyne. 1991. Impure Thoughts: Essays on Philosophy, Feminism, and Ethics. Philadelphia: Temple University Press.

Berliner, Paul F. 1994. Thinking About Jazz: The Infinite Art of Improvisation. Chicago: University of Chicago Press.

Monson, Ingrid. 1996. Saying Something: Jazz Improvisation and Interaction. Chicago: University of Chicago Press.

Spolin, Viola. 1963. Improvisation for the theater; a handbook of teaching and directing techniques. Evanston, IL: Northwestern University Press.