Free forever·no account, no tracking, no third-party requests ·open data, free API
showstack / learn / empathy
Stage 6 of 7 · The person6 of 7 in this stage10 min read

Feeling it with them

An audience does not decide to be moved. It either is or it is not, and a surprising amount of what decides that sits in departments who would not describe their job that way at all — the temperature of the room, whether the exit is visible, how long the foyer queue was, and whether the PA lets a thousand people hear each other.

About the register of this page

Most of this site reports things that were measured. This page mixes two kinds of claim and it is worth saying which is which as we go.

Measured: entrainment and synchrony, the effect of shared attention, contagion of expression, the peak-end structure of memory. These have numbers behind them.

A lens: the holding environment, the ritual passage. These are frameworks from clinical psychology and anthropology — useful for organising what you already know, not findings you can cite as fact. They earn their place here by being operational: each one names something a technical department decides.

The precondition

Nobody feels anything in a room that has not held them yet

Before any of the mechanism matters, there is a gate. A person will not risk an emotion in a space that has not demonstrated it can be trusted, and the demonstration is almost entirely non-verbal and almost entirely technical.

The psychoanalyst Donald Winnicott called this a holding environment: conditions reliable enough that somebody can afford to stop managing themselves. Borrowed into a venue it stops being abstract very quickly. Can they see an exit. Is the temperature something they have stopped noticing. Did the effect that just went off get announced at the door. Is the sound level one they can stay in for two hours rather than endure. Can they hear the words. Is the seat one they can sit in without the seat being the thing they are thinking about.

None of that is atmosphere. Each one is a decision with an owner, and each one is a decision that gets made on the basis of something else — the SPL that suits the mix, the haze level the lighting wants, the seat count the budget needs. The point of naming the gate is that an audience whose attention is on the room cannot put it on the stage, and no amount of craft downstream recovers that.

rule
An audience that is managing the room is not available to the show. Comfort is not the goal — absorption is — but discomfort you did not choose is a tax on everything upstream of it.

The shape

An event is a passage, not a duration

The anthropologist Victor Turner described ritual as three phases: separation from ordinary life, a threshold state where the usual rules are suspended, and a return. What happens in the middle he called communitas — the temporary sense of being one undifferentiated group rather than a room of individuals.

Whatever you make of the anthropology, the shape describes a show accurately and it exposes something useful. The threshold is the part everybody designs. The two ends are where events are consistently weakest, and both are technical.

Separation is the queue, the foyer, the walk down the aisle, the pre-show state of the room, and the moment the house lights go. That sequence is doing the work of getting a person out of an ordinary evening, and it is usually inherited rather than designed — a playlist nobody chose, a house state left at whatever the get-out was, a house-lights fade whose length nobody has ever discussed.

The return is worse. A curtain call, a house state, a walk out into a street, a last train. It is the last thing that happens and therefore weighted heavily in what gets remembered, and in most buildings it is the moment the technical department is thinking hardest about the load-out. An audience dumped from communitas into a work-light state with the bar shutters coming down has had the ending designed by nobody.

The mechanism

How one person’s state gets into another

Once the gate is open, three things do most of the work, and they are not the same thing.

Simulation. Watching an action recruits some of the machinery involved in doing it. The strong versions of the mirror-neuron story got well ahead of the evidence and are worth treating carefully, but the weaker claim is solid: perceiving effort, strain and intention engages motor and affective systems in the observer. Which is why real effort reads and simulated effort often does not. A performer genuinely holding a position, genuinely out of breath, genuinely lifting weight, is legible in a way a mimed version is not, and audiences detect the difference without being able to say what they detected.

Contagion. Expression spreads. People unconsciously mimic faces and postures around them and the mimicry feeds back into their own state. In a room this compounds, which is the mechanism behind laughter being louder in a full house and the reason a half-empty auditorium is not just quieter but genuinely less funny.

Synchrony. This is the one with real numbers behind it. Weakly coupled oscillators entrain — the same physics as pendulum clocks on a shared shelf. A room clapping converges on a common period within a few seconds, with nobody conducting, and moving together measurably raises reported closeness and cooperation between people who did it. The coupling in an auditorium is sound, which puts a system design decision directly underneath a social effect: if the crowd cannot hear itself, it cannot lock, and a room that never locks stays a set of individuals who happen to be in the same building.

rule
If the audience cannot hear itself, it cannot become an audience. Crowd mic returns, delay coverage at the back, and how much the PA masks the room are all decisions about whether communitas is available.
What an audience doesWhat is actually underneath itWho decides it
Laughs louder in a full houseExpression contagion compounding across a dense roomSeating density, and whether the room is acoustically live enough to carry it
Claps in time within secondsEntrainment through an audible common signalPA coverage, crowd returns, delay alignment
Leans in during real strainSimulation of effort the observer’s own body knowsWhether the effort is visible — light, scale, camera
Remembers the interval as the low pointPeak-end weighting of a badly designed return-and-reseparateInterval length, bar and toilet capacity, house state
Will not settle in the first ten minutesThe room has not demonstrated it will hold themTemperature, sightlines, level, front-of-house

The target state

Absorption, not comfort

It is worth being precise about what you are aiming at, because the obvious answer is wrong. The goal is not that an audience is comfortable. Mihály Csíkszentmihályi’s account of flow puts the good state at full absorption — attention entirely occupied, self-consciousness gone, time distorted — and absorption is not a relaxed state. It is a demanding one that people report as deeply satisfying afterwards.

What follows for a technical team is a slightly counter-intuitive rule. Comfort matters up to the point where discomfort would capture attention, and not one step further. A seat you notice is a problem. A seat so comfortable you drift is also a problem. The thing being protected is the audience’s attention budget, and every unresolved irritation spends some of it.

This is also why the failures on this page compound rather than add. Slightly too warm, slightly too loud, slightly obstructed sightline, slightly late caption: none of them individually ruins anything, and together they leave an audience with nothing left to give the show. People describe the result as the show not landing, which is not what happened.

The honest limits

What this does not give you

None of this predicts whether a particular show will move a particular person. It describes conditions and mechanisms, and there is a large gap between a condition being met and something good happening in it. A room can be perfectly held, perfectly synchronised and completely unmoved.

The lens material especially should be held loosely. Turner was describing initiation rituals in Ndembu society, not a Tuesday preview. Winnicott was describing infants and, later, therapy. Both travel well and neither was built for this, and anybody using them should be able to say what they would look like if they were wrong.

What survives the caution is the operational part, and it is enough to act on: the gate is real and it is made of technical decisions; the passage has two ends and one of them is almost never designed; and synchrony has a system-design dependency that is easy to check and easy to get wrong.

Where this goes next

The rest of the person layer

How a feeling is built has the mechanism of emotion itself, including the peak-end structure this page leans on. The person on the other end has attention, which is the resource everything here spends. Who the show is not reaching is the same argument as the precondition section, made concrete and legally specified. And experience architecture is where the whole chain gets treated as one design problem.

Check yourself

3 questions on the mechanism rather than the trivia. Nothing is scored and you can answer again — the explanation under a wrong answer is the point of it.

  1. 1The same comedy plays to a full house and to a half-empty one. The half-empty audience laughs measurably less. What is the main mechanism?

  2. 2A room of people starts clapping and locks to a common beat within a few seconds, with nobody leading. What has to be true for that to happen?

  3. 3An audience will not settle in the first ten minutes of a show. What is the most likely explanation on this page?

Answers are kept in this browser on this device. Nothing is sent anywhere.

You can now answer
  • Why is a full house funnier?
  • Why will they not settle?
  • Who designs the ending?
Kept in this browser on this device. Nothing is sent anywhere.

Where to go next

These pages are a chain rather than a library — the next one along usually answers the question this one raised.