The Booth and the Headset
Start with the apparatus, because the apparatus is where the argument lives. A describer sits in a soundproof room with a clear sightline to the stage, or with a monitor if the sightline is blocked, and speaks into a microphone. The narration travels to wireless receivers and headsets worn by blind and low-vision audience members. It does not reach the rest of the house. The described performance is therefore two performances running in parallel: one onstage, one in the ear, synchronized but not identical.
Arts Access Aotearoa, the New Zealand arts accessibility organization, defines audio description as “a narrated commentary for blind and low vision patrons that provides descriptions of the visual elements.” That definition is accurate and almost useless on its own, because it tells you nothing about what gets selected. The same page supplies the more revealing list. In a theatrical performance, description narrates what happens onstage between dialogue or songs: “new scenes, settings, costume, body language, facial expression, movements across the stage and sight gags.”
Read that list again as a choreographer would. Scenes, settings, costume, facial expression, sight gags — these are the furniture of narrative theater. They are the things a play needs described so that a listener can follow a plot. Body language and movement across the stage appear on the list, but as items among others, subordinate to the story being told. When the work is movement — when there is no dialogue to speak between, no plot to anchor the description, no costume change to mark the passage of time — the list stops being a guide and becomes a problem. What, exactly, is the describer supposed to narrate when the visual element is a body shifting weight from one leg to the other over eight slow counts?
The Touch Tour as Admission
The standard practice around described performances gives the answer obliquely. Arts Access notes that “an audio described performance usually includes a touch tour before the performance.” For dance or circus, patrons can handle “the different props, moves or poses.” This is the detail that should reorganize how the field thinks about description. The touch tour is not a bonus feature appended to the main access service. It is an acknowledgment, built into the format, that words alone cannot carry the spatial and tactile information of a moving body.
Consider what a touch tour actually transmits. A patron’s hand on a dancer’s extended arm learns the angle of the elbow, the rotation of the wrist, the tension in the forearm. That information arrives through pressure and proprioceptive inference, not through syntax. No sentence — “her arm extends diagonally upward” — delivers the same data, because the sentence has already converted continuous spatial fact into discrete verbal categories. The touch tour exists because the conversion loses something the field has decided is worth recovering by other means.
This is not a criticism of description. It is a description of what description is. Every act of verbal narration is a compression, and compression is lossy by definition. The question is not whether information is lost but which information, and who decides.
What and How: The Two Jobs of Movement Description
Research on accessible dance instruction makes the compression explicit. A 2025 co-design study by Das, Kappala, Chen, Huh, and Pavel, published through arXiv, worked with 28 participants — blind and low-vision dancers, dance instructors, sound designers, a haptics engineer, and an audio describer — across three workshops. The study’s framing of the problem is precise: describing movement is hard because audio description “must convey what to do (mechanics, location, orientation) and how to do it (speed, fluidity, timing).”
Those are two different jobs, and they do not fit comfortably into one channel. What is positional and structural: which limb, which direction, which level, which count. How is qualitative and temporal: the acceleration into a turn, the suspension at the top of a jump, the drag of a foot leaving the floor. A describer working live, in the gaps between phrases of music, has perhaps two seconds to say something. Two seconds is enough for “she crosses downstage left.” It is not enough for “she crosses downstage left with a weighted, reluctant quality, as if the floor is pulling at her.” The first sentence fits the gap. The second does not.
So the describer chooses. And in choosing, the describer performs the same operation a choreographer performs when setting a phrase: deciding what counts as the material and what counts as decoration.
The Multimodal Finding and Its Consequence
The co-design study’s most useful result for this argument is not a technology but a distribution. Participants designed eight systems, and across them a common pattern emerged: they “selectively used modalities — narration for movement structure, sound for expression, and haptics for spatial cues.”
That distribution is an implicit critique of the single-channel norm. If narration is best suited to structure, sound to expression, and touch to spatial orientation, then a purely verbal audio description of dance is a system that has assigned all three jobs to the channel best suited to only one of them. The expressive and spatial information does not disappear; it gets demoted. It becomes “atmosphere” — the thing a describer gestures at with an adjective when there is no time or no vocabulary to do better.
The study also cites prior work by Bläsing and Zimmermann finding that “the more modalities through which an observer has experienced a movement, the more enjoyment the observer derives from watching that movement.” This is a claim about reception, and it cuts against any account of access that treats description as a neutral substitute for sight. Description is not a substitute. It is an additional channel, and its value depends on what it carries relative to what the other channels — music, the sound of feet, the felt vibration of the floor, the touch tour’s remembered pressure — are already carrying.
The Vocabulary Problem
Here is where the argument becomes uncomfortable for the field. The words available to a describer are not neutral tools. They come preloaded with aesthetic assumptions. “Flowing” and “sharp” are not opposites in any physical sense; they are conventions inherited from a particular history of Western dance training, and they carry implicit valuations. A phrase described as “fluid” sounds accomplished. A phrase described as “jerky” sounds like a mistake. The describer who reaches for the first word has made an aesthetic judgment, whether or not they intended to.
The co-design study found that participants addressed this by “crafting vocabularies for movements” — inventing names and sound cues specific to a given routine rather than relying on generic descriptors. That is a compositional act. It treats description as a language that must be built for the work at hand, not a standard service applied to it. The implication for concert dance is direct: a describer preparing a work should be asking the same questions a rehearsal director asks. What is the phrase’s organizing principle? Where is its weight? What is the count structure, and which counts are the ones the audience is meant to feel as arrival?
None of this is currently standard practice, and the sources here do not document what concert-level dance description actually does. The Arts Access guidance is theatrical; the co-design study addresses instruction, not spectatorship. The gap between the two is where the real work sits, and it is a gap the field has not yet filled with published scripts, describer training specific to dance, or documented reception from blind spectators of concert work. That absence is itself a finding.
Description as Choreographic Apparatus
The argument, stated plainly: audio description of dance is not a translation service appended to a choreographic work. It is a second choreographic apparatus that runs alongside the first, selecting which body parts, directions, and qualities are legible as “the movement” and which are consigned to atmosphere. The describer’s choices — what to name, what to skip, when to say “she turns” and when to say “she spirals” — are compositional decisions with aesthetic consequences. They shape what a blind spectator understands the work to be about.
This is not a reason to abandon verbal description. It is a reason to stop treating it as compliance labor and start treating it as craft. The touch tour already admits that words are insufficient. The co-design research already shows that structure, expression, and space want different channels. What remains is for the field to take the next step: to train describers in movement analysis, to publish description scripts as documents worth reading, and to ask blind spectators what the description did to their experience of the work — not as a satisfaction metric but as reception data.
The booth and the headset are not a service window. They are a second stage, and someone is dancing in it.
FAQ
What is audio description for dance?
It is a live narrated commentary delivered to blind and low-vision audience members through headsets, describing visual elements of the performance as it happens. Arts Access Aotearoa defines it as narration of visual elements, and for theater it covers scenes, settings, costume, body language, facial expression, and movement across the stage.
Why is a touch tour usually offered alongside description?
Because verbal description cannot fully convey spatial and tactile information. Arts Access notes that described performances usually include a pre-show touch tour, and for dance or circus, patrons can handle props, moves, or poses. The touch tour supplements what words compress away.
What makes describing movement different from describing a play?
A play has dialogue and plot to anchor description. Movement does not. Research on accessible dance instruction identifies two separate jobs: conveying what to do (mechanics, location, orientation) and how to do it (speed, fluidity, timing). Live description has only brief gaps in which to do either.
Can words alone carry a dance performance?
Not fully. The 2025 co-design study found participants distributed movement information across channels — narration for structure, sound for expression, haptics for spatial cues — rather than loading everything onto words. A purely verbal description is a lossy compression that tends to demote expressive and spatial information to atmosphere.
Are a describer’s word choices neutral?
No. Terms like “flowing” and “sharp” carry aesthetic valuations inherited from dance training. The co-design study found participants built vocabularies specific to each routine rather than relying on generic descriptors, which treats description as a compositional act rather than a standard service.