November 13, 2007

Alternate Structures: Assemblies

One underlying thesis of this blog is that the play structure—as we commonly understand it—is an advanced, highly-specialized cultural artifact. Like the torque wrench or the toaster, it has been designed so effectively for its intended purpose—in this case, making the audience care about its characters—that it’s virtually useless for anything else. Were we a Pharaonic culture where art forms persisted for millennia, such an achievement might guarantee centuries of near perfect plays. But our cultural artifacts, like our toasters, tend to wear out quickly, and once the possibilities of "the play" structure approach exhaustion, theatre itself begins to feel exhausted. No further tweaking of the structure or its variants can revive the art form. New structures—and I would argue, new expectations for the nature and purpose of theatre—will be required.

But these things take time. By the middle of the last century, after several decades of mainstream representational playwriting (at least in the Old and New Worlds) and the intermittent dazzlements of a lunatic fringe, much of the energy of innovation had turned to souping-up, stripping down, boring out, chopping, channeling and candy-striping the old fleet of '97 Chekhovs and Ibsens. Lord knows, the craftsmanship was impressive. Yet by the three-quarter mark, it was hard to avoid the growing realization that underneath the hood, these were still used cars (which would make our current American non-profit theatre a virtual Little Havana!) At roughly the same time, though, an American avant-garde schooled in visual art theory began making full-length works with little or no apparent literary structure; Performance Art erupted as a rapidly mutating virus; and a smaller group of American "language" playwrights began generating scripts which sought to set aside as much of the baggage of play construction as possible. Such was the broth out of which new forms would finally emerge.

One of the most powerful, flexible and (to use the term from software development) extensible of these new structures is the assembly. In its simplest terms, an assembly is any theatrical piece composed of two or more recognizably different1 texts. For this one condition automatically subverts any illusion of textual authenticity (on which all further representational illusions turn out to depend), and shifts the author into the role of editor, selecting and shaping source "footage" to build up sequences of elements (which I call "threads") that can be further layered and intermixed. Distinguishing the shape, size and relation between disparate threads and elements is therefore critical, so the different pieces of an assembly are typically given distinct edges and joins. Other elements may even be purely formal—marking beginning and end points within the text—or function as arbitrary "markers" (see below for discussion). Narrative is often the connective thread within a source or sequence, yet no one narrative typically prevails as the story (or "meaning") of the overall piece2. The tolerance for textual variation, and the presentational stance vis-à-vis the audience make it easier to introduce expository material alongside traditional dramatic scenes and situations; indeed, assemblies are far more effective than plays in presenting flights of ideas or examinations of history and fact.

Charles Mee's Iphigenia 2.0, Carson Kreitzer's Flesh and the Desert, and Jason Grote's 1001—all seen recently in New York—represent variants of the assembly process which I've labeled Interpolation, Tessellation, and the Looped Stack.


The title alone—Iphigenia 2.0—tells the audience that the piece will be a myth reworked—and for those ignorant of mythology, copious notes were provided in the program and mounted in the theatre lobby. Indeed, the beauty of Mee's method was its simplicity: using some memory of Euripides as a framework within which foreign elements could replace their equivalents in the original. Every element in the piece was either part of the "original" story or an interpolation, and the figure/ground relationship between original and interpolation was in every case indicated by emphasizing the boundary between—most often by deliberate anachronism (as in the excerpt below, when "car," "Palestinian" and "RPGs" all appear in the same sentence.)

I was wrong.
I made a mistake.
I can't
do this.

I understand the difficulty this puts you in.

Do you?

But what sort of leader do
you pretend to be?
One who can make a decision
as it were, from the
but not when you imagine seeing
face to face
decision means in fact?

Is this the first you've heard what
dangers men
face in battle?

The time a car came towards us,
when, just five
minutes before, another car had come
and there were
four Palestinians in it
with RPGs
and they killed three of my friends.
So this new Peugeot comes
towards us,
and we shoot.
And there
was a family there--

Unlike the tightly-integrated, complex-story structure of the play, where everything is of-a-piece and fits, assemblies present an audience with the overall challenge of making sense of a set of disparate elements over the course of the piece, along with the secondary, more immediate challenge of maintaining one's place in the structure (continuity) from moment to moment. Paradoxically, the use of strong boundaries—which at first glance might seem disruptive—makes it much easier to distinguish different threads (i.e., there are no RPGs in Ancient Greece); and once a distinction has been made between elements, it is much easier to drawing inferences and connections between elements (i.e., that there may be similarities in the two situations).

Mee further simplified the task of orienting his audience by adopting the schematic (and, in a sense, already interpolated) structure of Greek drama—scene, chorus, scene—each scene consisting of an encounter between two (or sometimes, three) principals, interspersed with choral "odes" in the form of extended dance or movement sequences. In both versions, this clear alternation underscores the events or elements which constitute the main story, and those which are elaborations or digressions.

Interpolation differs from other assembly constructs in that there need be no necessary connection between successive interpolations. Continuity is only maintained through the spine or principal thread; the interpolations need not (and for clarity, probably should not) constitute a second thread. The lines “What do soldiers want?/They ask for almost nothing,” spin off into a catalog of modern products (“Oreo cookies. Canned tuna. Saltines. Salami.”). Or, when the bridesmaids enter and say they have been celebrating the upcoming wedding, the first bridesmaid continues with “My friend Dana/had this bachelorette party at the Beverly Hills Hotel?”

The extreme and obvious textual disjunction reinforces the internal boundaries essential to all assemblied. In an interpolation, the boundary lets us know at once that we have gone "somewhere else," a distinction maintained through the text (i.e., content) of the interpolation. For additional clarity, the relation of interpolation to the content in the principal thread it has replaced is typically one-to-one, and associative ("this thing is like that thing.") In these examples, we understand that the modern equivalent has replaced whatever Euripides might have written. Not only is the forward motion of the story halted—and its time scale suspended— but the location of the action shifts to some other place and time; yet these extreme displacements are not disruptive because the return point is also implicit. We expect the main story to resume where it left off, unaffected by (one wants to say "unaware of") its suspension. Indeed, it is not clear—to me at any rate—whether interpolation is even possible without a main thread whose broad outlines are known a priori; nor whether a second thread can be introduced without confusion. To put it another way, the problem that arises once interpolations acquire internal consistency is really the problem of working with multiple threads.


How can a second thread be introduced into a performance structure? The traditional solution, of course, was the sub-plot: a second, usually overlapping, story line within the superset of characters.3 Assemblies, on the other hand, do not require tight integration, and the number of sustainable threads is limited only by the author's ability to maintain a sense of overall structure while moving between them. Tessellation (arranging small shapes in a pattern) is a generic term for assemblies in which multiple threads have been broken into smaller parts and recombined.

Flesh and the Desert interweaves a set of very different narrative (or semi-narrative) threads having little in common beyond the fact that they all take place in Las Vegas. One principal sequence recreates an extended interview with an actual nightclub musician and his singer wife (Carter and Barbara); another is a realistic drama about a boy and girl who meet by chance at night in the desert; a third is a fantasy on Bugsy Siegel and his moll, who appear as ghosts moving through a presentational landscape of history; others include a series of vignettes of Vegas performers (e.g., Liberace, Siegfried and Roy; Elvis impersonators) and behind-the-scenes personnel. Finally, there are recurrent scenes of showgirls crossing the stage which serve as a kind of ornament or punctuation.

Here, in schematic form (and with a few simplifications, for clarity) is the opening sequence:

As the performance unfolds in real time, the first three scenes are self-contained, with only incidental clues (mentions of "slot machine" and "Flamingo hotel") suggesting how they might be related. The common connection (“it happened in Vegas”) is only revealed in the fourth scene (The Panorama of History), and then reinforced in the fifth by having the characters refer to the city:


A Brief History of Vegas.
First there was the boiling earth and molten lava. Ocean. Believe it or not. All underwater once.
Then the seas parted and left for parts more pleasant.
Plant life developed patience and a long root system.

Nevada territory.

State of Nevada. Not a whole lot of difference.
Everything’s legal, because there aren’t enough people to bother legislating too much.

Lights reveal Bugsy

Benjamin Siegel.

Bugsy lights a cigarette

Known to those who don’t actually know him as Bugsy.

Not my name.
I’m gonna let you off with a warning this time, but—
People been killed.
People been killed for that.

The gangster so crazy he saw Las Vegas shimmering before him, like a heat Mirage. Nothing but cactus, nothing but joshua trees. Nothing but the endless Sands. And he saw electricity flowing from the yet-unfinished Hoover Dam and he saw an Oasis out here in the Desert. . In(n) the beginning it was like a collection of cardboard flats, but it became a CIRCUS CIRCUS, like CEASAR’S PALACE in the roman days, where there had been nothing but FRONTIER. Now here we have this GOLDEN NUGGET shining in the DESERT, INN the dark. the light of a thousand suns.

The strip.

Benjamin Siegel saw it rising up out of the silent wavering sand.


What do you want to know?


Everything? You don’t want to know everything.

BARBARA enters

Well, I don’t know where I put those pictures. I’ll look again later.

Did you see that one of Carter on the bandstand? Of course, that’s in the War. The ones from Vegas are all in that box…


With its method and connective principle set forth, the play can begin to move forward by moving along and between threads; the structure may be loose but it is no longer random. The Barbara/Carter interview is interrupted, for example, by a second brief showgirl scene, easily understood as a merely interpolation with the swift resumption of the Carter thread. Siegel himself is mentioned in passing, then reappears in the next scene—a second history vignette, this time with showgirls—followed by the return of the scientist narrator who introduces a Liberace module, after which the Carter thread resumes yet again.

Once the framework and method have been well established, new threads can be introduced, letting the play unfold as a kind of structured ramble. Moreover, once the extent of the principal threads has been suggested, it is easy to bring in short, self-terminating elements (e.g., the wordless showgirl sequences, little more than a cross stage left to right) which serve as ornament; their information affects the structure or rhythm of the piece, contributing relatively little to the content. Indeed, a principal attraction of the tessellation is that it’s easy to build and easy to read. Like the chained or episodic narrative, its pleasures derive in large degree from the tension between continuity and discontinuity, between the surprise of branching out on a new tangent and the anticipatory satisfaction of the eventual resolution of any interrupted thread.

At the same time, the relaxed and discursive structure makes it easy to follow simultaneous stories. The forward motion of the piece derives from content; there is very little emphasis on pattern (which is to say that, unlike the Stack, the order in which threads follow each other is not strictly enforced). One is always curious to know what will happen next, yet there is little need to keep track of what has already happened. Some of these threads have a strongly delineated time scale, but others do not: we assume for example that the interview proceeds as it "really happened," but it wouldn't matter if the order were changed. Furthermore, the only thread with a conventional time-dependent narrative (Boy & Girl) does not define the time structure of the overall piece. Thus, by extension, the plot dynamics characteristic of the tightly-integrated play (foreshadowing, climax, reversal, etc.) are actually hard to deploy, because the overall structure of the piece is essentially flat.

All threads of a tessellation are functionally equivalent. The only way to establish even a crude hierarchy of primary and subordinate threads would be to regulate their lengths and frequencies—and, in the process, turn the assembly into a "badly-made" play.4 Instead of a hierarchy of "importance", tessellations encourage variations between threads in performance style, frame of reference, time scale and so forth--aspects of the dramatic framework which in the conventional play are typically kept constant.

Of course, replacing the expectation of unified "story" with a system of separate and distinct threads makes ending problematic; indeed, the truly cumulative resolution one expects in a play may not even be possible. At best, individual threads can be given clear endings, those ending can then be arranged to occur one after the other, and one or more threads may even be allowed to converge (e.g., by having similar actions, by introducing "connections," etc). But the true interpenetration of multiple threads requires one of two different structures: Stacks (discussed below), and a multi-threaded structure with frames or "windows" (which will have to await a subsequent post).

The Looped Stack

Like the assemblies previously discussed, Jason Grote's 1001 begins with a prolog, which itself begins with the word “Scheherazade”—who then appears—and, as in The Thousand Nights and One Night, sets off a series of stories within stories within stories. Here, for example, is the boundary between the initial framing story (The One-Eyed Arab) and the principal story of Shahriyar and Scheherazade:

B sits as THE ONE-EYED ARAB, as if in front of a tent in a bazaar. He may or may not be Middle Eastern, and clearly has both his eyes. During this, ensemble members dress C (still in bed) as SHAHRIYAR, a fearsome warrior king. Perhaps they remove the covers to reveal his costume.
Come! Come, sit! You like tea? Boy! Two teas. I see the lady is looking at my lace. For you, special price. Three dinar. I see you are looking at my missing eye. By Allah, do not be ashamed. There is a story behind the missing eye, but you are busy, we save this for another time. You like boys? Tea! Two boys. Come, sit. I shall share with you a story, bism'Allah, the story of the great Persian king and his famous wife Scheherazade. Come, sit.

Behind him, SHAHRIYAR jerks awake. Sounds of a porn film. SHAHRIYAR seethes, watching it.
There was a great king, Shahriyar, whose queen had been adulterous with a filthy blackamoor slave, hideous of visage, his lips like an open pot (etc.)

… and so the story continues, through a succession of scenes (helpfully given projected super-titles, like “THE TALE OF SHAHRIYAR, PART THREE”) , until at last, as we have known she would, the character of Scheherazade begins telling her first tale:

(all in one breath)
Begging your pardon my king my sister tells me stories each night before we sleep I beg mercy of your greatness for I know that after tonight I shall no longer hear the glorious stories of my sister Scheherazade and I beg you to allow her to tell me one last story for I shall never more hear her tell of them was that correct sister?

Yes, Dunyazade. A final mercy. Your "majesty." Before you lie with me.

Okay. And tell her she can take that

(loudly, as if D is deaf or foreign)
You can take that thing off!

And now, a story.
Stand in front of me now king, and look closely at the pieces of my face. Look at my full red lips, my wet mouth, my nose, and each eye. Gaze like the Sufis do when they spin and eat their hashish and stare into the endless black sky of the desert night. Do you see what they see?

Yes it's like a. Like a screen saver or.

She picks up the tome and opens it. It as if the storm she describes is rising from the pages.
A distant colorful storm, from an infinity away: is it a mirage? Watch as it comes closer, an exploding vortex, spinning forward, revealing itself to you, eating the sky, consuming all in its jagged green flames. Sit, o king, as I reveal to you: The Tale of Yahya Al-Husayni Amongst the Dead

B and E enter as FEMALE SLAVES, and begin to undress and make over D, who is now THE PRINCESS MARIDAH. F enters as PRINCE YAHYA, watching obsessively. B and E remove D's burqa, and dress her in a jilbab (a sort of Syrian pantsuit), deep blue, the same fabric as Shahriyar's silk.
There once was a Syrian prince, by Allah, who from his boyhood was madly in love with his own twin sister

… and this second tale continues for a while, until Shahriyar interrupts in the middle and Scheherazade launches into a third tale before concluding the second:

You have to tell me what Uh happens. It sounds like something I. But. Like a movie I uh. But.

If you must hear it. But I do fear the ending would surpass your understanding.

What is that supposed to uh. Mean?

I mean to cast no aspersion, majesty. I meant only that what the Emir Ghassan did then may only be understood if one first hears another story.

I don't uh What do you call it Care. Tell it.

(she's got him where she wants him)
Very well then. I bring you: The Tale of Alan in his Labyrinth

As SCHEHERAZADE narrates, ensemble members dress C as Alan, a modern-day hipster type who has been in a subway tunnel for some time. He has a head injury; perhaps he wears his dress shirt, the same color blue as the cloth in the earlier scenes, around his head. Dried blood should be visible through it. He wears work clothes of the type an arty temp or web programmer etc would wear. He is filthy.

There once was or someday will be, insh'Allah, one Man Hat... (etc)

… and so on. Even if you didn’t know the 1001 Nights, the structure would be obvious: Each new story is, as it were, added to top of a stack and—unless another is thrown on top of it—must be completed before the stories underneath is be resumed, in reverse order. The crucial difference between a tessellation and a stack is that a time-dependent hierarchy has been introduced across the entire assembly. The very possibility of anticipating the order in which the stories will be resumed establishes a meta-structure structure and even permits a degree of dramatic tension (in this, and only this, a stack also resembles an interpolated assembly).

Unfortunately, though, stacking is essentially a delaying tactic, which is fine if you don’t want your head cut off but in the theatre runs the risk of becoming tedious. Stories at the bottom of the stack cannot simply be abandoned without undermining the very principle of stacking; if early threads are allowed to wither away, the end result is little better than a sloppy (most unsatisfactory) tessellation. Yet to avoid tedium—how nice if only there were a second principle, whereby stories could be stacked to the point of maximum tension, and then collapsed, to allow accelerated motion without breaking the rules.

The solution turns out to be opening a thread which leads back to an earlier thread in the stack, a tunneling process I call "looping", and which in 1001 takes the form of literally moving the character of Alan through the subway tunnels of "Man Hat" until he runs into the One-Eyed Arab, thereby bringing the narrative action back to the top-level thread. At this point things get very interesting.

The earlier threads are not erased—they can be thought of as running "in the background"—and indeed the story will return to them soon enough. But the looping-back creates a kind of "knot," establishing an endpoint for one sequence and the beginning of another. The One-Eyed Arab leads Alan "on" to an entry-point into yet another thread, set in modern Manhattan, which is labeled "Alan and Dahna Part 4." Together in their apartment, Alan and Dahna then re-enact a later scene from the story of Shariyar and Scheherazade (the characters they have previously portrayed), wherein Scheherazade begins the story of Flaubert in Egypt—the novelist eventually meeting a "horrible monster" who starts up the tale of Alan and Dahna from the beginning...

Would a diagram help make this clearer?

Grote's structure, in effect, moves through two dimensions. The first is the "flat" world of multiple threads, which it shares with tessellations and interpolations. The second though is a dimension of loops and tunnels in which the sense of forward motion (what would traditionally be called "action") derives not from narrative but meta-narrative. Introducing this second dimension frees the structure from having to traverse all the points in a thread; the thread need only establish certain points, which can be traversed in non-narrative, a-temporal.5 sequences. (Grote's piece actually introduces a refinement of this notion by returning, at the end, to a variant of Alan & Dahna Part 4—maybe a third dimension??) Looping and tunneling circumvent the requirement that stories cannot be abandoned. The experience of the piece depends less on the information in the threads themselves than on the pattern of their traversal.

As we have seen, abstract structures which emphasize pattern over content also favor formal devices such as repetition and return, substitution (whereby one thing—be it object or actor—"stands for" or becomes another thing), and the introduction of "markers:" purely formal elements (objects, attributes, phrases, etc.) which recur, typically with variations, to create a further sense of internal consistency within the work. Thus a piece of blue cloth “the color of the desert sky at dusk” reappears twice and then, at a later point, a character standing at the top of a minaret, “could see all of everything , laid out before him against the deep blue canvas of the desert sky at dusk.” Or the actor playing the “One-Eyed Arab” (who clearly has two good eyes) reappears later as an Israeli with an eye patch. Like the shorter “decorative” elements of a tessellation (‘thread fragments”), such markers contain content of little inherent value. Their value is entirely structural, as further evidence of pattern. They can even imply causality—“Aha!” one says to oneself, “So that’s how he lost his eye!”—but of course, like so much else in theatre, it’s only a trick, leveraging our expectations to advantage.

(1) Even conventional are routinely composed of various sections of material; sometimes even writings from the same session can seem like the work of different authors. The critical term here is "recognizably;" once an audience recognizes that texts must have come from different places or, if you will, are in deliberately differentiated hands, the piece can no longer be presented as a transparent window on some virtual reality.

(2) The exception to this rule is the interpolated assembly, which I will argue necessarily requires one and only one principal thread. Despite the seams whenever the thread is interrupted or resumed, the continuity provided by the principal thread makes it easy to (mis)read the piece as the telling of that one story.

(3) The threat to internal consistency posed by the introduction of subplots is generally resolved by ensuring that the characters in all plots inhabit a common world, even if they do not necessarily interact. There can also, of course, be thematic connections, but the simplest solution is just to suggest that everyone in the play could theoretically meet each other.

(4) Once relationships are established between threads, it's hard to suppress expectations of eventual closure, and closure is perhaps the principal device of the conventional play. But closure alone is not sufficient to turn an assembly into a tightly-integrated complex story. To put it another way, identifying one thread as "principal" (and strictly speaking, there can be only one) immediately raises questions about its relation to all other threads. Interpolation side-steps this problem by ensuring that the interpolations clearly "do not belong" and do not, collectively, comprise a true second thread. And, as we have previously noted, the corresponding risk is that the principal thread will be read as the "story" of the entire assembly, thereby making the tension generated by the interpolations problematic.

(5) "A-temporal" only in the sense that the timeframes of the threads are disregarded. In fact, the threads in 1001 not only have their own internal time-scales, but are implicitly arrayed within a common time-scale wherein fictitious events might be imagined to happen on a plane parallel to historical time (this is a large part of the function of the Prolog). The looping, then, primarily moves through a multi-dimensional matrix of time.