Chapter 10 — Forms

Long form and the Harold

One suggestion, half an hour, and a structure designed less to be performed than to teach a group of people how to think as one.

Long form is the other half of the tradition. One suggestion is taken at the top and never again; what follows is a single connected piece, typically twenty-five to forty-five minutes, in which scenes recur, characters return, and something established early pays off late because somebody remembered it.

The comedy comes from pattern and recognition rather than from a rule imposed on a scene, and the pleasure for an audience is partly architectural — the sensation of watching a structure assemble itself out of apparently unrelated pieces.

The canonical long form, and the one every training theatre teaches, is the Harold.

Where it comes from

The Harold is associated above all with Del Close, who developed it over decades, first with The Committee in San Francisco in the 1960s and then, from the early 1980s, at the Chicago theatre ImprovOlympic — later iO — which he ran with Charna Halpern. The book Truth in Comedy, written by Halpern, Close and Kim "Howard" Johnson and published in 1994, is the closest thing the form has to a founding text.

The name means nothing. The story usually told is that when the group needed something to call the new structure, somebody answered with a joke name, and it stuck — an appropriately improvisational origin for a form built on accepting whatever you are given.

The shape

A Harold, in its textbook form:

• The opening. A group exploration of the suggestion, several minutes long, generating themes and images rather than scenes.
• Three first beats. Three unrelated two-person scenes, each establishing its own world and finding its own game.
• A group game. The whole cast, playing with an idea or pattern that has emerged.
• Three second beats. Return to each of the three scenes — later in time, heightened, or with the same characters somewhere new.
• A second group game.
• Three third beats. The threads push furthest, and typically begin to collide — characters from one scene appearing in another, or the three worlds turning out to share a theme.

That is the diagram. It is worth internalising and then holding loosely, for reasons we will get to.

The opening is not a warm-up

Beginners treat the opening as throat-clearing before the real show. It is the most important few minutes in the piece, because everything downstream is built from what it produces.

Common openings include word association, where the group free-associates from the suggestion; the pattern game, a more structured associative form Close favoured; a monologue, in which one performer tells a true personal story; and the invocation, where the group explores a single word through escalating layers of abstraction.

What all of them are doing is the same: generating a pool of specific images, phrases, tensions and themes that scenes can be drawn from. A good opening leaves the group with eight or nine things they are all now thinking about, which is why the scenes that follow can feel connected without anyone planning them.

Beats and callbacks

A beat is one visit to a thread. The first beat finds the world and its game. The second returns and heightens — often jumping forward in time, which is free escalation because the audience fills the gap themselves. The third pushes furthest, and this is where a Harold either becomes a piece of theatre or stays three sketches in a trench coat.

The convergence in third beats does not have to be literal. Characters meeting is the obvious version and often the weakest. The stronger version is thematic: three scenes that turn out to be asking the same question about loyalty, or debt, or the things people inherit from their parents, arriving at three different answers within a minute of each other.

That effect cannot be planned, only permitted — which is what all the listening in chapter three was for.

Group mind

The phrase you will hear constantly at iO-descended theatres is group mind: a state in which a cast is paying enough attention to each other that they start making the same choice simultaneously, without discussion.

It sounds mystical and is not. It is what happens when eight people have all absorbed the same opening, are all listening rather than planning, and are all willing to abandon their own idea in favour of the one already forming in the room. Two people initiating the same edit at the same moment is the visible symptom.

Where Harolds go wrong

• Performing the diagram. A group that hits every structural landmark on schedule while the scenes have nothing to do with each other has produced a Harold-shaped object rather than a Harold. Audiences cannot see structure; they can only see connection.
• Ignoring the opening. The most common failure. Everyone free-associates for four minutes and then starts a scene about something else entirely.
• Saving material. Holding a good idea back for your third beat means playing your first two beats worse. Spend everything immediately; more will arrive.
• Forcing convergence. Dragging characters across threads because the form says so, when no thematic link exists, reads as tidying up rather than as revelation.

The point of a structure you will not use

Here is the thing worth knowing before you spend a year on it: most experienced groups do not perform pure Harolds. They perform looser forms, or their own, or something that begins as a Harold and abandons the shape in the second half.

This is not a betrayal of the form. The Harold was always primarily a training structure — a scaffold that forces a group to practise callbacks, thematic listening, group games and returning to scenes, until those become instincts that no longer need the scaffold. Close himself was reportedly impatient with people treating it as a rulebook.

Learn it properly. Perform it until the habits are automatic. Then put it down and keep the habits.

The structure is not the show. The structure is the thing that taught eight people to notice the same thing at the same time, which is the show.